diff --git a/.gitattributes b/.gitattributes index c3c0c194a14b7b8865273f04cc303f29883d5da4..f0af8c4d9cbf6a442f54eaa8dc9114549de39f86 100644 --- a/.gitattributes +++ b/.gitattributes @@ -572,3 +572,10 @@ ood-vifactcheck-mmbert-source-source-07-s44-20260905T091424Z-fe08d0a1/trainer/ch ood-vifactcheck-mmbert-source-source-07-s44-20260905T091424Z-fe08d0a1/trainer/checkpoint-925/tokenizer.json filter=lfs diff=lfs merge=lfs -text ood-vifactcheck-mmbert-source-source-08-s44-20260905T092145Z-fe08d0a1/trainer/checkpoint-1140/tokenizer.json filter=lfs diff=lfs merge=lfs -text ood-vifactcheck-mmbert-source-source-08-s44-20260905T092145Z-fe08d0a1/trainer/checkpoint-570/tokenizer.json filter=lfs diff=lfs merge=lfs -text +ood-viclickbait2025-mmbert-temporal-temporal-00-s42-20260905T103412Z-3e0b1c91/trainer/checkpoint-680/tokenizer.json filter=lfs diff=lfs merge=lfs -text +ood-viclickbait2025-mmbert-temporal-temporal-00-s42-20260905T103412Z-3e0b1c91/trainer/checkpoint-884/tokenizer.json filter=lfs diff=lfs merge=lfs -text +ood-viclickbait2025-mmbert-temporal-temporal-00-s43-20260905T104139Z-520aec17/trainer/checkpoint-204/tokenizer.json filter=lfs diff=lfs merge=lfs -text +ood-viclickbait2025-mmbert-temporal-temporal-00-s43-20260905T104139Z-520aec17/trainer/checkpoint-408/tokenizer.json filter=lfs diff=lfs merge=lfs -text +ood-viclickbait2025-mmbert-temporal-temporal-00-s44-20260905T104521Z-e84c80be/trainer/checkpoint-204/tokenizer.json filter=lfs diff=lfs merge=lfs -text +ood-viclickbait2025-mmbert-temporal-temporal-00-s44-20260905T104521Z-e84c80be/trainer/checkpoint-408/tokenizer.json filter=lfs diff=lfs merge=lfs -text +ood-vigoemotions-mmbert-source-source-00-s42-20260905T105045Z-bd528b37/trainer/checkpoint-1632/tokenizer.json filter=lfs diff=lfs merge=lfs -text diff --git a/ood-viclickbait2025-mmbert-temporal-temporal-00-s42-20260905T103412Z-3e0b1c91/model.safetensors b/ood-viclickbait2025-mmbert-temporal-temporal-00-s42-20260905T103412Z-3e0b1c91/model.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..46685eaffdd1f7ca243037d84784ce6d41267ab4 --- /dev/null +++ b/ood-viclickbait2025-mmbert-temporal-temporal-00-s42-20260905T103412Z-3e0b1c91/model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:0b538ac6e8e894216542ac9751b8c13b3125b3cfbf81d64e24b7fe0a1e6cf672 +size 1227779120 diff --git a/ood-viclickbait2025-mmbert-temporal-temporal-00-s42-20260905T103412Z-3e0b1c91/trainer/checkpoint-680/model.safetensors b/ood-viclickbait2025-mmbert-temporal-temporal-00-s42-20260905T103412Z-3e0b1c91/trainer/checkpoint-680/model.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..17fe95a8f7fb8a0d33e9636a6b32d26b96bee8c8 --- /dev/null +++ b/ood-viclickbait2025-mmbert-temporal-temporal-00-s42-20260905T103412Z-3e0b1c91/trainer/checkpoint-680/model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:25d452b5e61660401dc01801e6b5b6065f87e8042dc8a63f1974c1815daace0d +size 1227779152 diff --git a/ood-viclickbait2025-mmbert-temporal-temporal-00-s42-20260905T103412Z-3e0b1c91/trainer/checkpoint-680/optimizer.pt b/ood-viclickbait2025-mmbert-temporal-temporal-00-s42-20260905T103412Z-3e0b1c91/trainer/checkpoint-680/optimizer.pt new file mode 100644 index 0000000000000000000000000000000000000000..767f55b1a5a445cf9050201f2c9334f6abc445e4 --- /dev/null +++ b/ood-viclickbait2025-mmbert-temporal-temporal-00-s42-20260905T103412Z-3e0b1c91/trainer/checkpoint-680/optimizer.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:33bbc9df16b3184cac1bfb1ab2f3da7ac57adbcbc69000956b6c06e0712bd0bf +size 2455644026 diff --git a/ood-viclickbait2025-mmbert-temporal-temporal-00-s42-20260905T103412Z-3e0b1c91/trainer/checkpoint-680/rng_state.pth b/ood-viclickbait2025-mmbert-temporal-temporal-00-s42-20260905T103412Z-3e0b1c91/trainer/checkpoint-680/rng_state.pth new file mode 100644 index 0000000000000000000000000000000000000000..88ebbab6d0e63e9de0891163dd2a73d812ab6f61 --- /dev/null +++ b/ood-viclickbait2025-mmbert-temporal-temporal-00-s42-20260905T103412Z-3e0b1c91/trainer/checkpoint-680/rng_state.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:a9c4a4a215a41b4a579792744e0affcde28f43046b8556d5b0c5d74e03d8cc4a +size 14244 diff --git a/ood-viclickbait2025-mmbert-temporal-temporal-00-s42-20260905T103412Z-3e0b1c91/trainer/checkpoint-680/scheduler.pt b/ood-viclickbait2025-mmbert-temporal-temporal-00-s42-20260905T103412Z-3e0b1c91/trainer/checkpoint-680/scheduler.pt new file mode 100644 index 0000000000000000000000000000000000000000..26cb8263cba8bf431b6ab772fb2e6fa0e5339884 --- /dev/null +++ b/ood-viclickbait2025-mmbert-temporal-temporal-00-s42-20260905T103412Z-3e0b1c91/trainer/checkpoint-680/scheduler.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:c1c6a5acf7a5d49ba245aa8f417bf9d2e26bb969c0421cfb3867292074387065 +size 1064 diff --git a/ood-viclickbait2025-mmbert-temporal-temporal-00-s42-20260905T103412Z-3e0b1c91/trainer/checkpoint-680/tokenizer.json b/ood-viclickbait2025-mmbert-temporal-temporal-00-s42-20260905T103412Z-3e0b1c91/trainer/checkpoint-680/tokenizer.json new file mode 100644 index 0000000000000000000000000000000000000000..5638b5d73671dac16e626c55ea60d3b45ebd7c9e --- /dev/null +++ b/ood-viclickbait2025-mmbert-temporal-temporal-00-s42-20260905T103412Z-3e0b1c91/trainer/checkpoint-680/tokenizer.json @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:4379fe7f4a7af10b1ac15ffc8ae90f446f92082393699e9dd54b8c3b5ec0abad +size 34363286 diff --git a/ood-viclickbait2025-mmbert-temporal-temporal-00-s42-20260905T103412Z-3e0b1c91/trainer/checkpoint-680/training_args.bin b/ood-viclickbait2025-mmbert-temporal-temporal-00-s42-20260905T103412Z-3e0b1c91/trainer/checkpoint-680/training_args.bin new file mode 100644 index 0000000000000000000000000000000000000000..cf4c8e341aff90c18729f89a7d79cc7cddd63464 --- /dev/null +++ b/ood-viclickbait2025-mmbert-temporal-temporal-00-s42-20260905T103412Z-3e0b1c91/trainer/checkpoint-680/training_args.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:ab158aaa995b84679c1823f86c68e62b03227f48539cba17d072f98abee7d96d +size 5688 diff --git a/ood-viclickbait2025-mmbert-temporal-temporal-00-s42-20260905T103412Z-3e0b1c91/trainer/checkpoint-884/model.safetensors b/ood-viclickbait2025-mmbert-temporal-temporal-00-s42-20260905T103412Z-3e0b1c91/trainer/checkpoint-884/model.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..9535d234a51a72c3575d76d73c100d4a0261b12e --- /dev/null +++ b/ood-viclickbait2025-mmbert-temporal-temporal-00-s42-20260905T103412Z-3e0b1c91/trainer/checkpoint-884/model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:b3ad50111cf5a75e3fc7e73fe6454089fab580855ffbd8305a4f5a7fe93fbf62 +size 1227779152 diff --git a/ood-viclickbait2025-mmbert-temporal-temporal-00-s42-20260905T103412Z-3e0b1c91/trainer/checkpoint-884/optimizer.pt b/ood-viclickbait2025-mmbert-temporal-temporal-00-s42-20260905T103412Z-3e0b1c91/trainer/checkpoint-884/optimizer.pt new file mode 100644 index 0000000000000000000000000000000000000000..5176cc451d150ac292bac9b12f6f681272667f25 --- /dev/null +++ b/ood-viclickbait2025-mmbert-temporal-temporal-00-s42-20260905T103412Z-3e0b1c91/trainer/checkpoint-884/optimizer.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:df6a3036cc0376afdc8bea4694481af8ce6f3fb4892435d8553e340c9d48061c +size 2455644026 diff --git a/ood-viclickbait2025-mmbert-temporal-temporal-00-s42-20260905T103412Z-3e0b1c91/trainer/checkpoint-884/rng_state.pth b/ood-viclickbait2025-mmbert-temporal-temporal-00-s42-20260905T103412Z-3e0b1c91/trainer/checkpoint-884/rng_state.pth new file mode 100644 index 0000000000000000000000000000000000000000..ca51e6ce7679deb659d7a439adc016f7b4410b9e --- /dev/null +++ b/ood-viclickbait2025-mmbert-temporal-temporal-00-s42-20260905T103412Z-3e0b1c91/trainer/checkpoint-884/rng_state.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:3da7f3356a0a19eeb1c899b32e6e246396708205c98c9e6f47bd284178cf59d8 +size 14244 diff --git a/ood-viclickbait2025-mmbert-temporal-temporal-00-s42-20260905T103412Z-3e0b1c91/trainer/checkpoint-884/scheduler.pt b/ood-viclickbait2025-mmbert-temporal-temporal-00-s42-20260905T103412Z-3e0b1c91/trainer/checkpoint-884/scheduler.pt new file mode 100644 index 0000000000000000000000000000000000000000..a2b98f3e04be5066e9677ae2a022d0bd0eea8de3 --- /dev/null +++ b/ood-viclickbait2025-mmbert-temporal-temporal-00-s42-20260905T103412Z-3e0b1c91/trainer/checkpoint-884/scheduler.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:fff8d4ee82801e6d8c25af1ffee6be3de6776e69dc7874a75dd949deae5a9828 +size 1064 diff --git a/ood-viclickbait2025-mmbert-temporal-temporal-00-s42-20260905T103412Z-3e0b1c91/trainer/checkpoint-884/tokenizer.json b/ood-viclickbait2025-mmbert-temporal-temporal-00-s42-20260905T103412Z-3e0b1c91/trainer/checkpoint-884/tokenizer.json new file mode 100644 index 0000000000000000000000000000000000000000..5638b5d73671dac16e626c55ea60d3b45ebd7c9e --- /dev/null +++ b/ood-viclickbait2025-mmbert-temporal-temporal-00-s42-20260905T103412Z-3e0b1c91/trainer/checkpoint-884/tokenizer.json @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:4379fe7f4a7af10b1ac15ffc8ae90f446f92082393699e9dd54b8c3b5ec0abad +size 34363286 diff --git a/ood-viclickbait2025-mmbert-temporal-temporal-00-s42-20260905T103412Z-3e0b1c91/trainer/checkpoint-884/training_args.bin b/ood-viclickbait2025-mmbert-temporal-temporal-00-s42-20260905T103412Z-3e0b1c91/trainer/checkpoint-884/training_args.bin new file mode 100644 index 0000000000000000000000000000000000000000..cf4c8e341aff90c18729f89a7d79cc7cddd63464 --- /dev/null +++ b/ood-viclickbait2025-mmbert-temporal-temporal-00-s42-20260905T103412Z-3e0b1c91/trainer/checkpoint-884/training_args.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:ab158aaa995b84679c1823f86c68e62b03227f48539cba17d072f98abee7d96d +size 5688 diff --git a/ood-viclickbait2025-mmbert-temporal-temporal-00-s43-20260905T104139Z-520aec17/model.safetensors b/ood-viclickbait2025-mmbert-temporal-temporal-00-s43-20260905T104139Z-520aec17/model.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..2b88ceb13ea2be63933a633db89ba982f53fd9c6 --- /dev/null +++ b/ood-viclickbait2025-mmbert-temporal-temporal-00-s43-20260905T104139Z-520aec17/model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:f9e9af4116bbf7171563ef0706292dd21a15789b875dbb5ada9b8f906512485c +size 1227779120 diff --git a/ood-viclickbait2025-mmbert-temporal-temporal-00-s43-20260905T104139Z-520aec17/trainer/checkpoint-204/model.safetensors b/ood-viclickbait2025-mmbert-temporal-temporal-00-s43-20260905T104139Z-520aec17/trainer/checkpoint-204/model.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..def3301b28095d912ce3f1ef30117ae8d557a864 --- /dev/null +++ b/ood-viclickbait2025-mmbert-temporal-temporal-00-s43-20260905T104139Z-520aec17/trainer/checkpoint-204/model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:7135037c6f95cbecc480f8f1589c1e56c4daada820e4df5742576132998ffe90 +size 1227779152 diff --git a/ood-viclickbait2025-mmbert-temporal-temporal-00-s43-20260905T104139Z-520aec17/trainer/checkpoint-204/optimizer.pt b/ood-viclickbait2025-mmbert-temporal-temporal-00-s43-20260905T104139Z-520aec17/trainer/checkpoint-204/optimizer.pt new file mode 100644 index 0000000000000000000000000000000000000000..969a4d37947867dfd30cb6c188374a1f9b3d7066 --- /dev/null +++ b/ood-viclickbait2025-mmbert-temporal-temporal-00-s43-20260905T104139Z-520aec17/trainer/checkpoint-204/optimizer.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:4df71f9288a769ec43f95c38b438743e3a982a86e95de9c07363959678d0f0b1 +size 2455644026 diff --git a/ood-viclickbait2025-mmbert-temporal-temporal-00-s43-20260905T104139Z-520aec17/trainer/checkpoint-204/rng_state.pth b/ood-viclickbait2025-mmbert-temporal-temporal-00-s43-20260905T104139Z-520aec17/trainer/checkpoint-204/rng_state.pth new file mode 100644 index 0000000000000000000000000000000000000000..4a0461bd5d1ac2cc8974d658026e8e238a60cf2f --- /dev/null +++ b/ood-viclickbait2025-mmbert-temporal-temporal-00-s43-20260905T104139Z-520aec17/trainer/checkpoint-204/rng_state.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:43685594baa296b0cc2c60b53d0ec1eb2ba11f750a795f40b7d167152ac8aa83 +size 14244 diff --git a/ood-viclickbait2025-mmbert-temporal-temporal-00-s43-20260905T104139Z-520aec17/trainer/checkpoint-204/scheduler.pt b/ood-viclickbait2025-mmbert-temporal-temporal-00-s43-20260905T104139Z-520aec17/trainer/checkpoint-204/scheduler.pt new file mode 100644 index 0000000000000000000000000000000000000000..96c3ca6a3f617a1e44cd49fc519894bef0156c1f --- /dev/null +++ b/ood-viclickbait2025-mmbert-temporal-temporal-00-s43-20260905T104139Z-520aec17/trainer/checkpoint-204/scheduler.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:4a6dd6fc6eb9755920b44e57454a0b38e345ccdcf2e6ebebbf0d29fcb70b2c73 +size 1064 diff --git a/ood-viclickbait2025-mmbert-temporal-temporal-00-s43-20260905T104139Z-520aec17/trainer/checkpoint-204/tokenizer.json b/ood-viclickbait2025-mmbert-temporal-temporal-00-s43-20260905T104139Z-520aec17/trainer/checkpoint-204/tokenizer.json new file mode 100644 index 0000000000000000000000000000000000000000..5638b5d73671dac16e626c55ea60d3b45ebd7c9e --- /dev/null +++ b/ood-viclickbait2025-mmbert-temporal-temporal-00-s43-20260905T104139Z-520aec17/trainer/checkpoint-204/tokenizer.json @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:4379fe7f4a7af10b1ac15ffc8ae90f446f92082393699e9dd54b8c3b5ec0abad +size 34363286 diff --git a/ood-viclickbait2025-mmbert-temporal-temporal-00-s43-20260905T104139Z-520aec17/trainer/checkpoint-204/training_args.bin b/ood-viclickbait2025-mmbert-temporal-temporal-00-s43-20260905T104139Z-520aec17/trainer/checkpoint-204/training_args.bin new file mode 100644 index 0000000000000000000000000000000000000000..a4a4f626b93566544d871c210c278f87cf647d6a --- /dev/null +++ b/ood-viclickbait2025-mmbert-temporal-temporal-00-s43-20260905T104139Z-520aec17/trainer/checkpoint-204/training_args.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:9768960a5fbd29069cbd2220495485e45c7a9686955009494756013db83efd4f +size 5688 diff --git a/ood-viclickbait2025-mmbert-temporal-temporal-00-s43-20260905T104139Z-520aec17/trainer/checkpoint-408/model.safetensors b/ood-viclickbait2025-mmbert-temporal-temporal-00-s43-20260905T104139Z-520aec17/trainer/checkpoint-408/model.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..bf5dfa6be7747e20ea12cbf78c5a2dc0a405800c --- /dev/null +++ b/ood-viclickbait2025-mmbert-temporal-temporal-00-s43-20260905T104139Z-520aec17/trainer/checkpoint-408/model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:1bb20e686f31577eedbb5fa3caac7e38ab1a7b636774eaa29bc4d52e7e9c95e1 +size 1227779152 diff --git a/ood-viclickbait2025-mmbert-temporal-temporal-00-s43-20260905T104139Z-520aec17/trainer/checkpoint-408/optimizer.pt b/ood-viclickbait2025-mmbert-temporal-temporal-00-s43-20260905T104139Z-520aec17/trainer/checkpoint-408/optimizer.pt new file mode 100644 index 0000000000000000000000000000000000000000..fe7dac39d3c3ea376ffa0da7352aef76afebc053 --- /dev/null +++ b/ood-viclickbait2025-mmbert-temporal-temporal-00-s43-20260905T104139Z-520aec17/trainer/checkpoint-408/optimizer.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:8eb712a34a2ae95a93bc5b7290532a85ec39e002b7101c1235723ad98478c173 +size 2455644026 diff --git a/ood-viclickbait2025-mmbert-temporal-temporal-00-s43-20260905T104139Z-520aec17/trainer/checkpoint-408/rng_state.pth b/ood-viclickbait2025-mmbert-temporal-temporal-00-s43-20260905T104139Z-520aec17/trainer/checkpoint-408/rng_state.pth new file mode 100644 index 0000000000000000000000000000000000000000..bfeca3fa39abc8286c5f7dca05e0823dd3c6c685 --- /dev/null +++ b/ood-viclickbait2025-mmbert-temporal-temporal-00-s43-20260905T104139Z-520aec17/trainer/checkpoint-408/rng_state.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:18f91dd2edb8584ef19c9f5b68fe93293483cc9f457802a2dbe62b89104c8945 +size 14244 diff --git a/ood-viclickbait2025-mmbert-temporal-temporal-00-s43-20260905T104139Z-520aec17/trainer/checkpoint-408/scheduler.pt b/ood-viclickbait2025-mmbert-temporal-temporal-00-s43-20260905T104139Z-520aec17/trainer/checkpoint-408/scheduler.pt new file mode 100644 index 0000000000000000000000000000000000000000..73409ba75bac64c0353ddaa1eb238dd3310d4d97 --- /dev/null +++ b/ood-viclickbait2025-mmbert-temporal-temporal-00-s43-20260905T104139Z-520aec17/trainer/checkpoint-408/scheduler.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:2c58b5fc1e3a79ccdc291dcb7a1d491894e72d0014b1978419dc82fac7adf274 +size 1064 diff --git a/ood-viclickbait2025-mmbert-temporal-temporal-00-s43-20260905T104139Z-520aec17/trainer/checkpoint-408/tokenizer.json b/ood-viclickbait2025-mmbert-temporal-temporal-00-s43-20260905T104139Z-520aec17/trainer/checkpoint-408/tokenizer.json new file mode 100644 index 0000000000000000000000000000000000000000..5638b5d73671dac16e626c55ea60d3b45ebd7c9e --- /dev/null +++ b/ood-viclickbait2025-mmbert-temporal-temporal-00-s43-20260905T104139Z-520aec17/trainer/checkpoint-408/tokenizer.json @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:4379fe7f4a7af10b1ac15ffc8ae90f446f92082393699e9dd54b8c3b5ec0abad +size 34363286 diff --git a/ood-viclickbait2025-mmbert-temporal-temporal-00-s43-20260905T104139Z-520aec17/trainer/checkpoint-408/training_args.bin b/ood-viclickbait2025-mmbert-temporal-temporal-00-s43-20260905T104139Z-520aec17/trainer/checkpoint-408/training_args.bin new file mode 100644 index 0000000000000000000000000000000000000000..a4a4f626b93566544d871c210c278f87cf647d6a --- /dev/null +++ b/ood-viclickbait2025-mmbert-temporal-temporal-00-s43-20260905T104139Z-520aec17/trainer/checkpoint-408/training_args.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:9768960a5fbd29069cbd2220495485e45c7a9686955009494756013db83efd4f +size 5688 diff --git a/ood-viclickbait2025-mmbert-temporal-temporal-00-s44-20260905T104521Z-e84c80be/model.safetensors b/ood-viclickbait2025-mmbert-temporal-temporal-00-s44-20260905T104521Z-e84c80be/model.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..e302b6cda16deb1e22206a583100c03a20f11339 --- /dev/null +++ b/ood-viclickbait2025-mmbert-temporal-temporal-00-s44-20260905T104521Z-e84c80be/model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:6a033bc7d272bab3d48061754d4e5e6e6bee72d561e4b3723677fbb97a4411ce +size 1227779120 diff --git a/ood-viclickbait2025-mmbert-temporal-temporal-00-s44-20260905T104521Z-e84c80be/trainer/checkpoint-204/model.safetensors b/ood-viclickbait2025-mmbert-temporal-temporal-00-s44-20260905T104521Z-e84c80be/trainer/checkpoint-204/model.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..39a0780e50ce622472ccc7e670402b80a192be46 --- /dev/null +++ b/ood-viclickbait2025-mmbert-temporal-temporal-00-s44-20260905T104521Z-e84c80be/trainer/checkpoint-204/model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:8d2a6766a032248b42e89de886ad994864db5efd32e7d64f6d64544c10196067 +size 1227779152 diff --git a/ood-viclickbait2025-mmbert-temporal-temporal-00-s44-20260905T104521Z-e84c80be/trainer/checkpoint-204/rng_state.pth b/ood-viclickbait2025-mmbert-temporal-temporal-00-s44-20260905T104521Z-e84c80be/trainer/checkpoint-204/rng_state.pth new file mode 100644 index 0000000000000000000000000000000000000000..78406579816e60e5677aa2272fec015854028108 --- /dev/null +++ b/ood-viclickbait2025-mmbert-temporal-temporal-00-s44-20260905T104521Z-e84c80be/trainer/checkpoint-204/rng_state.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:d36cd7a8c7f825b7f770739eed9d1e43be498f1d860a6b683a5971b05087dca2 +size 14244 diff --git a/ood-viclickbait2025-mmbert-temporal-temporal-00-s44-20260905T104521Z-e84c80be/trainer/checkpoint-204/scheduler.pt b/ood-viclickbait2025-mmbert-temporal-temporal-00-s44-20260905T104521Z-e84c80be/trainer/checkpoint-204/scheduler.pt new file mode 100644 index 0000000000000000000000000000000000000000..96c3ca6a3f617a1e44cd49fc519894bef0156c1f --- /dev/null +++ b/ood-viclickbait2025-mmbert-temporal-temporal-00-s44-20260905T104521Z-e84c80be/trainer/checkpoint-204/scheduler.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:4a6dd6fc6eb9755920b44e57454a0b38e345ccdcf2e6ebebbf0d29fcb70b2c73 +size 1064 diff --git a/ood-viclickbait2025-mmbert-temporal-temporal-00-s44-20260905T104521Z-e84c80be/trainer/checkpoint-204/tokenizer.json b/ood-viclickbait2025-mmbert-temporal-temporal-00-s44-20260905T104521Z-e84c80be/trainer/checkpoint-204/tokenizer.json new file mode 100644 index 0000000000000000000000000000000000000000..5638b5d73671dac16e626c55ea60d3b45ebd7c9e --- /dev/null +++ b/ood-viclickbait2025-mmbert-temporal-temporal-00-s44-20260905T104521Z-e84c80be/trainer/checkpoint-204/tokenizer.json @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:4379fe7f4a7af10b1ac15ffc8ae90f446f92082393699e9dd54b8c3b5ec0abad +size 34363286 diff --git a/ood-viclickbait2025-mmbert-temporal-temporal-00-s44-20260905T104521Z-e84c80be/trainer/checkpoint-204/training_args.bin b/ood-viclickbait2025-mmbert-temporal-temporal-00-s44-20260905T104521Z-e84c80be/trainer/checkpoint-204/training_args.bin new file mode 100644 index 0000000000000000000000000000000000000000..0c13c341dbc25e38d18fc0f631ef9dcd17a3279c --- /dev/null +++ b/ood-viclickbait2025-mmbert-temporal-temporal-00-s44-20260905T104521Z-e84c80be/trainer/checkpoint-204/training_args.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:a43d759a208cb18a6c7efd7360f7b7e23e49e0294894f02577030114684ea79a +size 5688 diff --git a/ood-viclickbait2025-mmbert-temporal-temporal-00-s44-20260905T104521Z-e84c80be/trainer/checkpoint-408/model.safetensors b/ood-viclickbait2025-mmbert-temporal-temporal-00-s44-20260905T104521Z-e84c80be/trainer/checkpoint-408/model.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..d6a15190e120df32befe93eaef4b421f32c60dfa --- /dev/null +++ b/ood-viclickbait2025-mmbert-temporal-temporal-00-s44-20260905T104521Z-e84c80be/trainer/checkpoint-408/model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:d0184d983c542d1adc61d9cbe120f7f38cc775056bea75c16d78359a7752a8b4 +size 1227779152 diff --git a/ood-viclickbait2025-mmbert-temporal-temporal-00-s44-20260905T104521Z-e84c80be/trainer/checkpoint-408/rng_state.pth b/ood-viclickbait2025-mmbert-temporal-temporal-00-s44-20260905T104521Z-e84c80be/trainer/checkpoint-408/rng_state.pth new file mode 100644 index 0000000000000000000000000000000000000000..6702ca1013bbefc8bb30409d0fddd5bbf27e1ec3 --- /dev/null +++ b/ood-viclickbait2025-mmbert-temporal-temporal-00-s44-20260905T104521Z-e84c80be/trainer/checkpoint-408/rng_state.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:9c89eea2d7e97771c58f15760ab58746aabb99ce351d3bf437d55beb0f66cc01 +size 14244 diff --git a/ood-viclickbait2025-mmbert-temporal-temporal-00-s44-20260905T104521Z-e84c80be/trainer/checkpoint-408/scheduler.pt b/ood-viclickbait2025-mmbert-temporal-temporal-00-s44-20260905T104521Z-e84c80be/trainer/checkpoint-408/scheduler.pt new file mode 100644 index 0000000000000000000000000000000000000000..73409ba75bac64c0353ddaa1eb238dd3310d4d97 --- /dev/null +++ b/ood-viclickbait2025-mmbert-temporal-temporal-00-s44-20260905T104521Z-e84c80be/trainer/checkpoint-408/scheduler.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:2c58b5fc1e3a79ccdc291dcb7a1d491894e72d0014b1978419dc82fac7adf274 +size 1064 diff --git a/ood-viclickbait2025-mmbert-temporal-temporal-00-s44-20260905T104521Z-e84c80be/trainer/checkpoint-408/tokenizer.json b/ood-viclickbait2025-mmbert-temporal-temporal-00-s44-20260905T104521Z-e84c80be/trainer/checkpoint-408/tokenizer.json new file mode 100644 index 0000000000000000000000000000000000000000..5638b5d73671dac16e626c55ea60d3b45ebd7c9e --- /dev/null +++ b/ood-viclickbait2025-mmbert-temporal-temporal-00-s44-20260905T104521Z-e84c80be/trainer/checkpoint-408/tokenizer.json @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:4379fe7f4a7af10b1ac15ffc8ae90f446f92082393699e9dd54b8c3b5ec0abad +size 34363286 diff --git a/ood-viclickbait2025-mmbert-temporal-temporal-00-s44-20260905T104521Z-e84c80be/trainer/checkpoint-408/training_args.bin b/ood-viclickbait2025-mmbert-temporal-temporal-00-s44-20260905T104521Z-e84c80be/trainer/checkpoint-408/training_args.bin new file mode 100644 index 0000000000000000000000000000000000000000..0c13c341dbc25e38d18fc0f631ef9dcd17a3279c --- /dev/null +++ b/ood-viclickbait2025-mmbert-temporal-temporal-00-s44-20260905T104521Z-e84c80be/trainer/checkpoint-408/training_args.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:a43d759a208cb18a6c7efd7360f7b7e23e49e0294894f02577030114684ea79a +size 5688 diff --git a/ood-vigoemotions-mmbert-source-source-00-s42-20260905T105045Z-bd528b37/model.safetensors b/ood-vigoemotions-mmbert-source-source-00-s42-20260905T105045Z-bd528b37/model.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..3eb34183f56d593ed93ddfe616ab60a17e16ea00 --- /dev/null +++ b/ood-vigoemotions-mmbert-source-source-00-s42-20260905T105045Z-bd528b37/model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:a65b9a854946f28004a6eab5db5d70a48e94fd9f78358d167df5ab07961a3b36 +size 1227859104 diff --git a/ood-vigoemotions-mmbert-source-source-00-s42-20260905T105045Z-bd528b37/trainer/checkpoint-1632/tokenizer.json b/ood-vigoemotions-mmbert-source-source-00-s42-20260905T105045Z-bd528b37/trainer/checkpoint-1632/tokenizer.json new file mode 100644 index 0000000000000000000000000000000000000000..5638b5d73671dac16e626c55ea60d3b45ebd7c9e --- /dev/null +++ b/ood-vigoemotions-mmbert-source-source-00-s42-20260905T105045Z-bd528b37/trainer/checkpoint-1632/tokenizer.json @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:4379fe7f4a7af10b1ac15ffc8ae90f446f92082393699e9dd54b8c3b5ec0abad +size 34363286 diff --git a/ood-vigoemotions-mmbert-source-source-00-s42-20260905T105045Z-bd528b37/trainer/checkpoint-1632/training_args.bin b/ood-vigoemotions-mmbert-source-source-00-s42-20260905T105045Z-bd528b37/trainer/checkpoint-1632/training_args.bin new file mode 100644 index 0000000000000000000000000000000000000000..18c251ed2f5e2cdb11172012e703e86a3794e1b6 --- /dev/null +++ b/ood-vigoemotions-mmbert-source-source-00-s42-20260905T105045Z-bd528b37/trainer/checkpoint-1632/training_args.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:453ef74cf23df784d0b9a9b48cb79af790d4858ad2a43581082a01c5a647ad01 +size 5624 diff --git a/ood-vigoemotions-mmbert-source-source-00-s43-20260905T123747Z-60628050/model.safetensors b/ood-vigoemotions-mmbert-source-source-00-s43-20260905T123747Z-60628050/model.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..b5af9735d590f5c28c6307093cb3be529c8f8ede --- /dev/null +++ b/ood-vigoemotions-mmbert-source-source-00-s43-20260905T123747Z-60628050/model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:59b6def2b87a6ae069067db51f4fe6e7fd16dace26f1ff2cdb4b1812c26d4f44 +size 1227859104 diff --git a/ood-vigoemotions-mmbert-source-source-00-s44-20260905T142419Z-37344fee/model.safetensors b/ood-vigoemotions-mmbert-source-source-00-s44-20260905T142419Z-37344fee/model.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..a7d3874111c9a1ee3efe01b3d0cdf3c444482690 --- /dev/null +++ b/ood-vigoemotions-mmbert-source-source-00-s44-20260905T142419Z-37344fee/model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:a74f20e4ce35f0d39c76804d95b0d288ac1eea8a9444d4919b07ce64cf47c4dd +size 1227859104 diff --git a/ood-vigoemotions-mmbert-source-source-00-s44-20260905T142419Z-37344fee/trainer/checkpoint-2244/special_tokens_map.json b/ood-vigoemotions-mmbert-source-source-00-s44-20260905T142419Z-37344fee/trainer/checkpoint-2244/special_tokens_map.json new file mode 100644 index 0000000000000000000000000000000000000000..be4ad795f49075ac5e56d8108ef63be314a1e967 --- /dev/null +++ b/ood-vigoemotions-mmbert-source-source-00-s44-20260905T142419Z-37344fee/trainer/checkpoint-2244/special_tokens_map.json @@ -0,0 +1,55 @@ +{ + "additional_special_tokens": [ + "", + "" + ], + "bos_token": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false + }, + "cls_token": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false + }, + "eos_token": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false + }, + "mask_token": { + "content": "", + "lstrip": true, + "normalized": false, + "rstrip": false, + "single_word": false + }, + "pad_token": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false + }, + "sep_token": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false + }, + "unk_token": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false + } +} diff --git a/ood-vigoemotions-mmbert-source-source-00-s44-20260905T142419Z-37344fee/trainer/checkpoint-2244/tokenizer_config.json b/ood-vigoemotions-mmbert-source-source-00-s44-20260905T142419Z-37344fee/trainer/checkpoint-2244/tokenizer_config.json new file mode 100644 index 0000000000000000000000000000000000000000..b002fb8c2960be8ebf9c373d6e5784b4565d7edc --- /dev/null +++ b/ood-vigoemotions-mmbert-source-source-00-s44-20260905T142419Z-37344fee/trainer/checkpoint-2244/tokenizer_config.json @@ -0,0 +1,2018 @@ +{ + "add_bos_token": true, + "added_tokens_decoder": { + "0": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "1": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "2": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "3": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "4": { + "content": "", + "lstrip": true, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "5": { + "content": "<2mass>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "6": { + "content": "[@BOS@]", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "7": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "8": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "9": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "10": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "11": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "12": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "13": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "14": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "15": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "16": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "17": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "18": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "19": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "20": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "21": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "22": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "23": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "24": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "25": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "26": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "27": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "28": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "29": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "30": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "31": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "32": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "33": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "34": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "35": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "36": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "37": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "38": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "39": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "40": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "41": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "42": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "43": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "44": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "45": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "46": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "47": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "48": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "49": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "50": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "51": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "52": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "53": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "54": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "55": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "56": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "57": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "58": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "59": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "60": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "61": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "62": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "63": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "64": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "65": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "66": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "67": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "68": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "69": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "70": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "71": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "72": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "73": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "74": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "75": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "76": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "77": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "78": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "79": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "80": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "81": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "82": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "83": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "84": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "85": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "86": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "87": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "88": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "89": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "90": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "91": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "92": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "93": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "94": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "95": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "96": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "97": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "98": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "99": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "100": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "101": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "102": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "103": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "104": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "105": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "106": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "107": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "108": { + "content": "\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "109": { + "content": "\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "110": { + "content": "\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "111": { + "content": "\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "112": { + "content": "\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "113": { + "content": "\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "114": { + "content": "\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "115": { + "content": "\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "116": { + "content": "\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "117": { + "content": "\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "118": { + "content": "\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "119": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "120": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "121": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "122": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "123": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "124": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "125": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "126": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "127": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "128": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "129": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "130": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "131": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "132": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "133": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "134": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "135": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "136": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "137": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "138": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "139": { + "content": "▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "140": { + "content": "▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "141": { + "content": "▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "142": { + "content": "▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "143": { + "content": "▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "144": { + "content": "▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "145": { + "content": "▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "146": { + "content": "▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "147": { + "content": "▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "148": { + "content": "▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "149": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "150": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "151": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "152": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "153": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "154": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "155": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "156": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "157": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "158": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "159": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "160": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "161": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "162": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "163": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "164": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "165": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "166": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "167": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "168": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "169": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "170": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "172": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "173": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "174": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "175": { + "content": "
", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "171": { + "content": "
", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "176": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "177": { + "content": "
", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "178": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "179": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "180": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "181": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "182": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "183": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "184": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "185": { + "content": "

", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "186": { + "content": "

", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "187": { + "content": "

", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "188": { + "content": "

", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "189": { + "content": "

", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "190": { + "content": "
", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "191": { + "content": "
", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "192": { + "content": "
", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "193": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "194": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "195": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "196": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "197": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "198": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "199": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "200": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "201": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "202": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "203": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "204": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "205": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "206": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "207": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "208": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "209": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "210": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "211": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "212": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "213": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "214": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "215": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "216": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255968": { + "content": "[toxicity=0]", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255969": { + "content": "\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255970": { + "content": "\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255971": { + "content": "\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255972": { + "content": "\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255973": { + "content": "\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255974": { + "content": "\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255975": { + "content": "\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255976": { + "content": "\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255977": { + "content": "\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255978": { + "content": "\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255979": { + "content": "\t\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255980": { + "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255981": { + "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255982": { + "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255983": { + "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255984": { + "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255985": { + "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255986": { + "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255987": { + "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255988": { + "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255989": { + "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255990": { + "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255991": { + "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255992": { + "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255993": { + "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255994": { + "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255995": { + "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255996": { + "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255997": { + "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255998": { + "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255999": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + } + }, + "additional_special_tokens": [ + "", + "" + ], + "bos_token": "", + "clean_up_tokenization_spaces": false, + "cls_token": "", + "eos_token": "", + "extra_special_tokens": {}, + "mask_token": "", + "model_input_names": [ + "input_ids", + "attention_mask" + ], + "model_max_length": 8192, + "pad_token": "", + "padding_side": "right", + "sep_token": "", + "spaces_between_special_tokens": false, + "tokenizer_class": "PreTrainedTokenizerFast", + "unk_token": "" +} diff --git a/ood-vigoemotions-mmbert-source-source-00-s44-20260905T142419Z-37344fee/trainer/checkpoint-2244/trainer_state.json b/ood-vigoemotions-mmbert-source-source-00-s44-20260905T142419Z-37344fee/trainer/checkpoint-2244/trainer_state.json new file mode 100644 index 0000000000000000000000000000000000000000..a7dfc8c602ac56d731524d142a832a8a8a9a1f96 --- /dev/null +++ b/ood-vigoemotions-mmbert-source-source-00-s44-20260905T142419Z-37344fee/trainer/checkpoint-2244/trainer_state.json @@ -0,0 +1,798 @@ +{ + "best_global_step": 2244, + "best_metric": 0.5854704605866574, + "best_model_checkpoint": "/root/VNEncoder/results/checkpoints/ood-vigoemotions-mmbert-source-source-00-s44-20260905T142419Z-37344fee/trainer/checkpoint-2244", + "epoch": 11.0, + "eval_steps": 500, + "global_step": 2244, + "is_hyper_param_search": false, + "is_local_process_zero": true, + "is_world_process_zero": true, + "log_history": [ + { + "epoch": 0.12300123001230012, + "grad_norm": 22.820940017700195, + "learning_rate": 1.1764705882352942e-06, + "loss": 6.3498, + "step": 25 + }, + { + "epoch": 0.24600246002460024, + "grad_norm": 25.63102149963379, + "learning_rate": 2.401960784313726e-06, + "loss": 6.1346, + "step": 50 + }, + { + "epoch": 0.36900369003690037, + "grad_norm": 26.335094451904297, + "learning_rate": 3.6274509803921573e-06, + "loss": 6.4895, + "step": 75 + }, + { + "epoch": 0.4920049200492005, + "grad_norm": 132.60525512695312, + "learning_rate": 4.852941176470589e-06, + "loss": 6.225, + "step": 100 + }, + { + "epoch": 0.6150061500615006, + "grad_norm": 54.02048873901367, + "learning_rate": 6.07843137254902e-06, + "loss": 5.797, + "step": 125 + }, + { + "epoch": 0.7380073800738007, + "grad_norm": 38.846336364746094, + "learning_rate": 7.3039215686274515e-06, + "loss": 5.1479, + "step": 150 + }, + { + "epoch": 0.8610086100861009, + "grad_norm": 43.72068786621094, + "learning_rate": 8.529411764705883e-06, + "loss": 4.8019, + "step": 175 + }, + { + "epoch": 0.984009840098401, + "grad_norm": 66.03960418701172, + "learning_rate": 9.754901960784315e-06, + "loss": 4.348, + "step": 200 + }, + { + "epoch": 1.0, + "eval_accuracy": 0.0013831258644536654, + "eval_loss": 1.0146747827529907, + "eval_macro_f1": 0.27515518782346493, + "eval_micro_f1": 0.28834819404564, + "eval_runtime": 3.2821, + "eval_samples_per_second": 220.284, + "eval_steps_per_second": 14.015, + "eval_weighted_f1": 0.3539820483990392, + "step": 204 + }, + { + "epoch": 1.103321033210332, + "grad_norm": 44.26076126098633, + "learning_rate": 1.0980392156862747e-05, + "loss": 3.7842, + "step": 225 + }, + { + "epoch": 1.2263222632226323, + "grad_norm": 66.00853729248047, + "learning_rate": 1.2205882352941179e-05, + "loss": 3.6033, + "step": 250 + }, + { + "epoch": 1.3493234932349323, + "grad_norm": 92.09275817871094, + "learning_rate": 1.3431372549019607e-05, + "loss": 3.5391, + "step": 275 + }, + { + "epoch": 1.4723247232472325, + "grad_norm": 43.25274658203125, + "learning_rate": 1.4656862745098039e-05, + "loss": 3.1562, + "step": 300 + }, + { + "epoch": 1.5953259532595325, + "grad_norm": 65.29901885986328, + "learning_rate": 1.5882352941176473e-05, + "loss": 3.3981, + "step": 325 + }, + { + "epoch": 1.7183271832718328, + "grad_norm": 54.06962203979492, + "learning_rate": 1.7107843137254903e-05, + "loss": 3.1319, + "step": 350 + }, + { + "epoch": 1.841328413284133, + "grad_norm": 53.133201599121094, + "learning_rate": 1.8333333333333333e-05, + "loss": 3.0991, + "step": 375 + }, + { + "epoch": 1.964329643296433, + "grad_norm": 51.17352294921875, + "learning_rate": 1.9558823529411766e-05, + "loss": 2.8816, + "step": 400 + }, + { + "epoch": 2.0, + "eval_accuracy": 0.027662517289073305, + "eval_loss": 0.7307998538017273, + "eval_macro_f1": 0.3687256673408171, + "eval_micro_f1": 0.37910085054678005, + "eval_runtime": 2.0045, + "eval_samples_per_second": 360.689, + "eval_steps_per_second": 22.948, + "eval_weighted_f1": 0.465781434005995, + "step": 408 + }, + { + "epoch": 2.0836408364083643, + "grad_norm": 32.426170349121094, + "learning_rate": 1.9912854030501092e-05, + "loss": 2.4614, + "step": 425 + }, + { + "epoch": 2.206642066420664, + "grad_norm": 28.6024112701416, + "learning_rate": 1.9776688453159043e-05, + "loss": 2.2084, + "step": 450 + }, + { + "epoch": 2.3296432964329643, + "grad_norm": 40.920738220214844, + "learning_rate": 1.9640522875816993e-05, + "loss": 2.1315, + "step": 475 + }, + { + "epoch": 2.4526445264452645, + "grad_norm": 52.49609375, + "learning_rate": 1.9504357298474947e-05, + "loss": 2.2481, + "step": 500 + }, + { + "epoch": 2.5756457564575648, + "grad_norm": 50.98602294921875, + "learning_rate": 1.9368191721132898e-05, + "loss": 2.1713, + "step": 525 + }, + { + "epoch": 2.6986469864698646, + "grad_norm": 23.695022583007812, + "learning_rate": 1.9232026143790852e-05, + "loss": 2.2749, + "step": 550 + }, + { + "epoch": 2.821648216482165, + "grad_norm": 163.39114379882812, + "learning_rate": 1.9095860566448803e-05, + "loss": 2.1554, + "step": 575 + }, + { + "epoch": 2.944649446494465, + "grad_norm": 37.01408767700195, + "learning_rate": 1.8959694989106757e-05, + "loss": 2.1779, + "step": 600 + }, + { + "epoch": 3.0, + "eval_accuracy": 0.11756569847856155, + "eval_loss": 0.6780247688293457, + "eval_macro_f1": 0.44742742763662813, + "eval_micro_f1": 0.4847926267281106, + "eval_runtime": 2.0675, + "eval_samples_per_second": 349.693, + "eval_steps_per_second": 22.249, + "eval_weighted_f1": 0.5449678663096842, + "step": 612 + }, + { + "epoch": 3.063960639606396, + "grad_norm": 27.182884216308594, + "learning_rate": 1.8823529411764708e-05, + "loss": 1.6404, + "step": 625 + }, + { + "epoch": 3.1869618696186963, + "grad_norm": 28.162418365478516, + "learning_rate": 1.8687363834422662e-05, + "loss": 1.4181, + "step": 650 + }, + { + "epoch": 3.3099630996309966, + "grad_norm": 22.72654151916504, + "learning_rate": 1.8551198257080612e-05, + "loss": 1.4334, + "step": 675 + }, + { + "epoch": 3.4329643296432963, + "grad_norm": 23.24102210998535, + "learning_rate": 1.8415032679738563e-05, + "loss": 1.3488, + "step": 700 + }, + { + "epoch": 3.5559655596555966, + "grad_norm": 47.752838134765625, + "learning_rate": 1.8278867102396517e-05, + "loss": 1.2765, + "step": 725 + }, + { + "epoch": 3.678966789667897, + "grad_norm": 41.577880859375, + "learning_rate": 1.8142701525054468e-05, + "loss": 1.4195, + "step": 750 + }, + { + "epoch": 3.8019680196801966, + "grad_norm": 29.362136840820312, + "learning_rate": 1.800653594771242e-05, + "loss": 1.4393, + "step": 775 + }, + { + "epoch": 3.924969249692497, + "grad_norm": 41.29766845703125, + "learning_rate": 1.7870370370370373e-05, + "loss": 1.4676, + "step": 800 + }, + { + "epoch": 4.0, + "eval_accuracy": 0.1396957123098202, + "eval_loss": 0.7599406242370605, + "eval_macro_f1": 0.49790743621877676, + "eval_micro_f1": 0.5432230522945571, + "eval_runtime": 2.0136, + "eval_samples_per_second": 359.053, + "eval_steps_per_second": 22.844, + "eval_weighted_f1": 0.5845070792455435, + "step": 816 + }, + { + "epoch": 4.044280442804428, + "grad_norm": 24.567283630371094, + "learning_rate": 1.7734204793028323e-05, + "loss": 1.1482, + "step": 825 + }, + { + "epoch": 4.167281672816729, + "grad_norm": 18.17946434020996, + "learning_rate": 1.7598039215686274e-05, + "loss": 0.9048, + "step": 850 + }, + { + "epoch": 4.290282902829028, + "grad_norm": 30.113676071166992, + "learning_rate": 1.7461873638344228e-05, + "loss": 0.8644, + "step": 875 + }, + { + "epoch": 4.413284132841328, + "grad_norm": 19.796863555908203, + "learning_rate": 1.732570806100218e-05, + "loss": 0.9021, + "step": 900 + }, + { + "epoch": 4.536285362853628, + "grad_norm": 38.6467170715332, + "learning_rate": 1.7189542483660133e-05, + "loss": 0.8805, + "step": 925 + }, + { + "epoch": 4.659286592865929, + "grad_norm": 19.9637451171875, + "learning_rate": 1.7053376906318084e-05, + "loss": 0.8908, + "step": 950 + }, + { + "epoch": 4.782287822878229, + "grad_norm": 19.54431915283203, + "learning_rate": 1.6917211328976034e-05, + "loss": 0.8628, + "step": 975 + }, + { + "epoch": 4.905289052890529, + "grad_norm": 42.1319694519043, + "learning_rate": 1.678104575163399e-05, + "loss": 0.913, + "step": 1000 + }, + { + "epoch": 5.0, + "eval_accuracy": 0.21853388658367912, + "eval_loss": 0.8532860279083252, + "eval_macro_f1": 0.5446277224859946, + "eval_micro_f1": 0.6013943619278569, + "eval_runtime": 2.1705, + "eval_samples_per_second": 333.107, + "eval_steps_per_second": 21.194, + "eval_weighted_f1": 0.6197483086177735, + "step": 1020 + }, + { + "epoch": 5.02460024600246, + "grad_norm": 14.424633026123047, + "learning_rate": 1.664488017429194e-05, + "loss": 0.7874, + "step": 1025 + }, + { + "epoch": 5.14760147601476, + "grad_norm": 22.660926818847656, + "learning_rate": 1.6508714596949893e-05, + "loss": 0.5939, + "step": 1050 + }, + { + "epoch": 5.27060270602706, + "grad_norm": 10.341612815856934, + "learning_rate": 1.6372549019607844e-05, + "loss": 0.6016, + "step": 1075 + }, + { + "epoch": 5.39360393603936, + "grad_norm": 13.016510963439941, + "learning_rate": 1.6236383442265798e-05, + "loss": 0.5896, + "step": 1100 + }, + { + "epoch": 5.516605166051661, + "grad_norm": 12.98374080657959, + "learning_rate": 1.610021786492375e-05, + "loss": 0.5628, + "step": 1125 + }, + { + "epoch": 5.639606396063961, + "grad_norm": 18.153182983398438, + "learning_rate": 1.5964052287581703e-05, + "loss": 0.5384, + "step": 1150 + }, + { + "epoch": 5.762607626076261, + "grad_norm": 31.739131927490234, + "learning_rate": 1.5827886710239654e-05, + "loss": 0.588, + "step": 1175 + }, + { + "epoch": 5.885608856088561, + "grad_norm": 17.49053192138672, + "learning_rate": 1.5691721132897604e-05, + "loss": 0.5548, + "step": 1200 + }, + { + "epoch": 6.0, + "eval_accuracy": 0.24896265560165975, + "eval_loss": 0.9286503195762634, + "eval_macro_f1": 0.5594913430296943, + "eval_micro_f1": 0.6213844653883653, + "eval_runtime": 2.1194, + "eval_samples_per_second": 341.126, + "eval_steps_per_second": 21.704, + "eval_weighted_f1": 0.6305536481148704, + "step": 1224 + }, + { + "epoch": 6.004920049200492, + "grad_norm": 11.133593559265137, + "learning_rate": 1.555555555555556e-05, + "loss": 0.5617, + "step": 1225 + }, + { + "epoch": 6.127921279212792, + "grad_norm": 19.947908401489258, + "learning_rate": 1.541938997821351e-05, + "loss": 0.3871, + "step": 1250 + }, + { + "epoch": 6.250922509225092, + "grad_norm": 10.981277465820312, + "learning_rate": 1.528322440087146e-05, + "loss": 0.3695, + "step": 1275 + }, + { + "epoch": 6.373923739237393, + "grad_norm": 10.755515098571777, + "learning_rate": 1.5147058823529412e-05, + "loss": 0.3878, + "step": 1300 + }, + { + "epoch": 6.496924969249693, + "grad_norm": 15.75013256072998, + "learning_rate": 1.5010893246187366e-05, + "loss": 0.3848, + "step": 1325 + }, + { + "epoch": 6.619926199261993, + "grad_norm": 16.140092849731445, + "learning_rate": 1.4874727668845317e-05, + "loss": 0.3697, + "step": 1350 + }, + { + "epoch": 6.7429274292742925, + "grad_norm": 7.486774444580078, + "learning_rate": 1.473856209150327e-05, + "loss": 0.3593, + "step": 1375 + }, + { + "epoch": 6.865928659286593, + "grad_norm": 14.175186157226562, + "learning_rate": 1.4602396514161222e-05, + "loss": 0.359, + "step": 1400 + }, + { + "epoch": 6.988929889298893, + "grad_norm": 10.35640811920166, + "learning_rate": 1.4466230936819172e-05, + "loss": 0.3633, + "step": 1425 + }, + { + "epoch": 7.0, + "eval_accuracy": 0.2987551867219917, + "eval_loss": 1.0920287370681763, + "eval_macro_f1": 0.5689211077657381, + "eval_micro_f1": 0.6445804825351098, + "eval_runtime": 2.0034, + "eval_samples_per_second": 360.879, + "eval_steps_per_second": 22.961, + "eval_weighted_f1": 0.6444084807872194, + "step": 1428 + }, + { + "epoch": 7.108241082410824, + "grad_norm": 8.464420318603516, + "learning_rate": 1.4330065359477125e-05, + "loss": 0.2676, + "step": 1450 + }, + { + "epoch": 7.231242312423125, + "grad_norm": 14.645303726196289, + "learning_rate": 1.4193899782135077e-05, + "loss": 0.2554, + "step": 1475 + }, + { + "epoch": 7.354243542435424, + "grad_norm": 11.637530326843262, + "learning_rate": 1.405773420479303e-05, + "loss": 0.2556, + "step": 1500 + }, + { + "epoch": 7.477244772447724, + "grad_norm": 6.641851425170898, + "learning_rate": 1.392156862745098e-05, + "loss": 0.2307, + "step": 1525 + }, + { + "epoch": 7.6002460024600245, + "grad_norm": 10.870826721191406, + "learning_rate": 1.3785403050108934e-05, + "loss": 0.2557, + "step": 1550 + }, + { + "epoch": 7.723247232472325, + "grad_norm": 5.410327434539795, + "learning_rate": 1.3649237472766885e-05, + "loss": 0.2265, + "step": 1575 + }, + { + "epoch": 7.846248462484625, + "grad_norm": 8.480887413024902, + "learning_rate": 1.351307189542484e-05, + "loss": 0.2693, + "step": 1600 + }, + { + "epoch": 7.969249692496925, + "grad_norm": 18.576095581054688, + "learning_rate": 1.337690631808279e-05, + "loss": 0.255, + "step": 1625 + }, + { + "epoch": 8.0, + "eval_accuracy": 0.313969571230982, + "eval_loss": 1.2188948392868042, + "eval_macro_f1": 0.5811899137156786, + "eval_micro_f1": 0.6468622354251764, + "eval_runtime": 2.497, + "eval_samples_per_second": 289.543, + "eval_steps_per_second": 18.422, + "eval_weighted_f1": 0.6483152693096889, + "step": 1632 + }, + { + "epoch": 8.088560885608857, + "grad_norm": 5.418167591094971, + "learning_rate": 1.3240740740740742e-05, + "loss": 0.1742, + "step": 1650 + }, + { + "epoch": 8.211562115621156, + "grad_norm": 6.825408935546875, + "learning_rate": 1.3104575163398695e-05, + "loss": 0.1686, + "step": 1675 + }, + { + "epoch": 8.334563345633457, + "grad_norm": 6.550326824188232, + "learning_rate": 1.2968409586056645e-05, + "loss": 0.1649, + "step": 1700 + }, + { + "epoch": 8.457564575645756, + "grad_norm": 4.261825084686279, + "learning_rate": 1.2832244008714598e-05, + "loss": 0.1437, + "step": 1725 + }, + { + "epoch": 8.580565805658056, + "grad_norm": 5.172179222106934, + "learning_rate": 1.269607843137255e-05, + "loss": 0.1642, + "step": 1750 + }, + { + "epoch": 8.703567035670357, + "grad_norm": 23.130123138427734, + "learning_rate": 1.2559912854030503e-05, + "loss": 0.1641, + "step": 1775 + }, + { + "epoch": 8.826568265682656, + "grad_norm": 30.504459381103516, + "learning_rate": 1.2423747276688453e-05, + "loss": 0.1669, + "step": 1800 + }, + { + "epoch": 8.949569495694957, + "grad_norm": 8.059348106384277, + "learning_rate": 1.2287581699346407e-05, + "loss": 0.1654, + "step": 1825 + }, + { + "epoch": 9.0, + "eval_accuracy": 0.3278008298755187, + "eval_loss": 1.3381789922714233, + "eval_macro_f1": 0.5762311884316192, + "eval_micro_f1": 0.648936170212766, + "eval_runtime": 2.0185, + "eval_samples_per_second": 358.184, + "eval_steps_per_second": 22.789, + "eval_weighted_f1": 0.6484288983365116, + "step": 1836 + }, + { + "epoch": 9.068880688806889, + "grad_norm": 6.1497297286987305, + "learning_rate": 1.2151416122004358e-05, + "loss": 0.135, + "step": 1850 + }, + { + "epoch": 9.191881918819188, + "grad_norm": 5.762045860290527, + "learning_rate": 1.2015250544662312e-05, + "loss": 0.1209, + "step": 1875 + }, + { + "epoch": 9.314883148831488, + "grad_norm": 4.636457443237305, + "learning_rate": 1.1879084967320263e-05, + "loss": 0.1211, + "step": 1900 + }, + { + "epoch": 9.437884378843789, + "grad_norm": 6.796161651611328, + "learning_rate": 1.1742919389978214e-05, + "loss": 0.1124, + "step": 1925 + }, + { + "epoch": 9.560885608856088, + "grad_norm": 4.42055606842041, + "learning_rate": 1.1606753812636166e-05, + "loss": 0.1129, + "step": 1950 + }, + { + "epoch": 9.68388683886839, + "grad_norm": 17.374441146850586, + "learning_rate": 1.1470588235294118e-05, + "loss": 0.1064, + "step": 1975 + }, + { + "epoch": 9.806888068880689, + "grad_norm": 2.4789116382598877, + "learning_rate": 1.133442265795207e-05, + "loss": 0.113, + "step": 2000 + }, + { + "epoch": 9.92988929889299, + "grad_norm": 11.736351013183594, + "learning_rate": 1.1198257080610021e-05, + "loss": 0.1101, + "step": 2025 + }, + { + "epoch": 10.0, + "eval_accuracy": 0.3333333333333333, + "eval_loss": 1.4571764469146729, + "eval_macro_f1": 0.571516196920113, + "eval_micro_f1": 0.6534730306558013, + "eval_runtime": 2.043, + "eval_samples_per_second": 353.897, + "eval_steps_per_second": 22.516, + "eval_weighted_f1": 0.6500766986172553, + "step": 2040 + }, + { + "epoch": 10.04920049200492, + "grad_norm": 4.998640537261963, + "learning_rate": 1.1062091503267976e-05, + "loss": 0.0896, + "step": 2050 + }, + { + "epoch": 10.17220172201722, + "grad_norm": 2.433173179626465, + "learning_rate": 1.0925925925925926e-05, + "loss": 0.0786, + "step": 2075 + }, + { + "epoch": 10.29520295202952, + "grad_norm": 2.942702531814575, + "learning_rate": 1.078976034858388e-05, + "loss": 0.0839, + "step": 2100 + }, + { + "epoch": 10.418204182041821, + "grad_norm": 4.1860785484313965, + "learning_rate": 1.0653594771241831e-05, + "loss": 0.0809, + "step": 2125 + }, + { + "epoch": 10.54120541205412, + "grad_norm": 4.45076847076416, + "learning_rate": 1.0517429193899782e-05, + "loss": 0.0713, + "step": 2150 + }, + { + "epoch": 10.664206642066421, + "grad_norm": 5.46282958984375, + "learning_rate": 1.0381263616557736e-05, + "loss": 0.0877, + "step": 2175 + }, + { + "epoch": 10.78720787207872, + "grad_norm": 6.310950756072998, + "learning_rate": 1.0245098039215687e-05, + "loss": 0.081, + "step": 2200 + }, + { + "epoch": 10.91020910209102, + "grad_norm": 3.4199132919311523, + "learning_rate": 1.0108932461873639e-05, + "loss": 0.0725, + "step": 2225 + }, + { + "epoch": 11.0, + "eval_accuracy": 0.34716459197787, + "eval_loss": 1.595555067062378, + "eval_macro_f1": 0.5854704605866574, + "eval_micro_f1": 0.6532744073925271, + "eval_runtime": 2.0731, + "eval_samples_per_second": 348.749, + "eval_steps_per_second": 22.189, + "eval_weighted_f1": 0.650225101846109, + "step": 2244 + } + ], + "logging_steps": 25, + "max_steps": 4080, + "num_input_tokens_seen": 0, + "num_train_epochs": 20, + "save_steps": 500, + "stateful_callbacks": { + "EarlyStoppingCallback": { + "args": { + "early_stopping_patience": 3, + "early_stopping_threshold": 0.0 + }, + "attributes": { + "early_stopping_patience_counter": 0 + } + }, + "TrainerControl": { + "args": { + "should_epoch_stop": false, + "should_evaluate": false, + "should_log": false, + "should_save": true, + "should_training_stop": false + }, + "attributes": {} + } + }, + "total_flos": 0.0, + "train_batch_size": 8, + "trial_name": null, + "trial_params": null +} diff --git a/ood-vigoemotions-mmbert-source-source-00-s44-20260905T142419Z-37344fee/trainer/checkpoint-2856/special_tokens_map.json b/ood-vigoemotions-mmbert-source-source-00-s44-20260905T142419Z-37344fee/trainer/checkpoint-2856/special_tokens_map.json new file mode 100644 index 0000000000000000000000000000000000000000..be4ad795f49075ac5e56d8108ef63be314a1e967 --- /dev/null +++ b/ood-vigoemotions-mmbert-source-source-00-s44-20260905T142419Z-37344fee/trainer/checkpoint-2856/special_tokens_map.json @@ -0,0 +1,55 @@ +{ + "additional_special_tokens": [ + "", + "" + ], + "bos_token": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false + }, + "cls_token": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false + }, + "eos_token": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false + }, + "mask_token": { + "content": "", + "lstrip": true, + "normalized": false, + "rstrip": false, + "single_word": false + }, + "pad_token": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false + }, + "sep_token": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false + }, + "unk_token": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false + } +} diff --git a/ood-vigoemotions-mmbert-source-source-00-s44-20260905T142419Z-37344fee/trainer/checkpoint-2856/tokenizer_config.json b/ood-vigoemotions-mmbert-source-source-00-s44-20260905T142419Z-37344fee/trainer/checkpoint-2856/tokenizer_config.json new file mode 100644 index 0000000000000000000000000000000000000000..b002fb8c2960be8ebf9c373d6e5784b4565d7edc --- /dev/null +++ b/ood-vigoemotions-mmbert-source-source-00-s44-20260905T142419Z-37344fee/trainer/checkpoint-2856/tokenizer_config.json @@ -0,0 +1,2018 @@ +{ + "add_bos_token": true, + "added_tokens_decoder": { + "0": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "1": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "2": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "3": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "4": { + "content": "", + "lstrip": true, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "5": { + "content": "<2mass>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "6": { + "content": "[@BOS@]", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "7": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "8": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "9": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "10": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "11": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "12": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "13": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "14": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "15": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "16": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "17": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "18": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "19": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "20": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "21": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "22": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "23": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "24": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "25": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "26": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "27": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "28": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "29": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "30": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "31": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "32": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "33": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "34": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "35": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "36": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "37": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "38": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "39": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "40": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "41": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "42": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "43": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "44": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "45": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "46": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "47": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "48": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "49": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "50": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "51": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "52": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "53": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "54": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "55": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "56": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "57": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "58": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "59": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "60": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "61": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "62": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "63": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "64": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "65": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "66": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "67": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "68": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "69": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "70": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "71": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "72": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "73": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "74": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "75": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "76": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "77": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "78": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "79": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "80": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "81": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "82": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "83": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "84": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "85": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "86": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "87": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "88": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "89": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "90": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "91": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "92": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "93": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "94": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "95": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "96": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "97": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "98": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "99": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "100": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "101": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "102": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "103": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "104": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "105": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "106": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "107": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "108": { + "content": "\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "109": { + "content": "\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "110": { + "content": "\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "111": { + "content": "\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "112": { + "content": "\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "113": { + "content": "\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "114": { + "content": "\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "115": { + "content": "\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "116": { + "content": "\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "117": { + "content": "\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "118": { + "content": "\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "119": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "120": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "121": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "122": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "123": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "124": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "125": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "126": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "127": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "128": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "129": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "130": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "131": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "132": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "133": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "134": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "135": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "136": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "137": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "138": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "139": { + "content": "▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "140": { + "content": "▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "141": { + "content": "▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "142": { + "content": "▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "143": { + "content": "▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "144": { + "content": "▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "145": { + "content": "▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "146": { + "content": "▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "147": { + "content": "▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "148": { + "content": "▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "149": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "150": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "151": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "152": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "153": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "154": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "155": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "156": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "157": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "158": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "159": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "160": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "161": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "162": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "163": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "164": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "165": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "166": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "167": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "168": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "169": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "170": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "172": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "173": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "174": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "175": { + "content": "
", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "171": { + "content": "
", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "176": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "177": { + "content": "
", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "178": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "179": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "180": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "181": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "182": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "183": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "184": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "185": { + "content": "

", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "186": { + "content": "

", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "187": { + "content": "

", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "188": { + "content": "

", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "189": { + "content": "

", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "190": { + "content": "
", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "191": { + "content": "
", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "192": { + "content": "
", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "193": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "194": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "195": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "196": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "197": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "198": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "199": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "200": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "201": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "202": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "203": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "204": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "205": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "206": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "207": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "208": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "209": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "210": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "211": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "212": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "213": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "214": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "215": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "216": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255968": { + "content": "[toxicity=0]", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255969": { + "content": "\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255970": { + "content": "\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255971": { + "content": "\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255972": { + "content": "\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255973": { + "content": "\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255974": { + "content": "\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255975": { + "content": "\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255976": { + "content": "\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255977": { + "content": "\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255978": { + "content": "\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255979": { + "content": "\t\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255980": { + "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255981": { + "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255982": { + "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255983": { + "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255984": { + "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255985": { + "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255986": { + "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255987": { + "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255988": { + "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255989": { + "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255990": { + "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255991": { + "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255992": { + "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255993": { + "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255994": { + "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255995": { + "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255996": { + "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255997": { + "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255998": { + "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255999": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + } + }, + "additional_special_tokens": [ + "", + "" + ], + "bos_token": "", + "clean_up_tokenization_spaces": false, + "cls_token": "", + "eos_token": "", + "extra_special_tokens": {}, + "mask_token": "", + "model_input_names": [ + "input_ids", + "attention_mask" + ], + "model_max_length": 8192, + "pad_token": "", + "padding_side": "right", + "sep_token": "", + "spaces_between_special_tokens": false, + "tokenizer_class": "PreTrainedTokenizerFast", + "unk_token": "" +} diff --git a/ood-vigoemotions-mmbert-source-source-00-s44-20260905T142419Z-37344fee/trainer/checkpoint-2856/trainer_state.json b/ood-vigoemotions-mmbert-source-source-00-s44-20260905T142419Z-37344fee/trainer/checkpoint-2856/trainer_state.json new file mode 100644 index 0000000000000000000000000000000000000000..413752def42ad55f7a8c30e7f0f901b99522902a --- /dev/null +++ b/ood-vigoemotions-mmbert-source-source-00-s44-20260905T142419Z-37344fee/trainer/checkpoint-2856/trainer_state.json @@ -0,0 +1,1009 @@ +{ + "best_global_step": 2244, + "best_metric": 0.5854704605866574, + "best_model_checkpoint": "/root/VNEncoder/results/checkpoints/ood-vigoemotions-mmbert-source-source-00-s44-20260905T142419Z-37344fee/trainer/checkpoint-2244", + "epoch": 14.0, + "eval_steps": 500, + "global_step": 2856, + "is_hyper_param_search": false, + "is_local_process_zero": true, + "is_world_process_zero": true, + "log_history": [ + { + "epoch": 0.12300123001230012, + "grad_norm": 22.820940017700195, + "learning_rate": 1.1764705882352942e-06, + "loss": 6.3498, + "step": 25 + }, + { + "epoch": 0.24600246002460024, + "grad_norm": 25.63102149963379, + "learning_rate": 2.401960784313726e-06, + "loss": 6.1346, + "step": 50 + }, + { + "epoch": 0.36900369003690037, + "grad_norm": 26.335094451904297, + "learning_rate": 3.6274509803921573e-06, + "loss": 6.4895, + "step": 75 + }, + { + "epoch": 0.4920049200492005, + "grad_norm": 132.60525512695312, + "learning_rate": 4.852941176470589e-06, + "loss": 6.225, + "step": 100 + }, + { + "epoch": 0.6150061500615006, + "grad_norm": 54.02048873901367, + "learning_rate": 6.07843137254902e-06, + "loss": 5.797, + "step": 125 + }, + { + "epoch": 0.7380073800738007, + "grad_norm": 38.846336364746094, + "learning_rate": 7.3039215686274515e-06, + "loss": 5.1479, + "step": 150 + }, + { + "epoch": 0.8610086100861009, + "grad_norm": 43.72068786621094, + "learning_rate": 8.529411764705883e-06, + "loss": 4.8019, + "step": 175 + }, + { + "epoch": 0.984009840098401, + "grad_norm": 66.03960418701172, + "learning_rate": 9.754901960784315e-06, + "loss": 4.348, + "step": 200 + }, + { + "epoch": 1.0, + "eval_accuracy": 0.0013831258644536654, + "eval_loss": 1.0146747827529907, + "eval_macro_f1": 0.27515518782346493, + "eval_micro_f1": 0.28834819404564, + "eval_runtime": 3.2821, + "eval_samples_per_second": 220.284, + "eval_steps_per_second": 14.015, + "eval_weighted_f1": 0.3539820483990392, + "step": 204 + }, + { + "epoch": 1.103321033210332, + "grad_norm": 44.26076126098633, + "learning_rate": 1.0980392156862747e-05, + "loss": 3.7842, + "step": 225 + }, + { + "epoch": 1.2263222632226323, + "grad_norm": 66.00853729248047, + "learning_rate": 1.2205882352941179e-05, + "loss": 3.6033, + "step": 250 + }, + { + "epoch": 1.3493234932349323, + "grad_norm": 92.09275817871094, + "learning_rate": 1.3431372549019607e-05, + "loss": 3.5391, + "step": 275 + }, + { + "epoch": 1.4723247232472325, + "grad_norm": 43.25274658203125, + "learning_rate": 1.4656862745098039e-05, + "loss": 3.1562, + "step": 300 + }, + { + "epoch": 1.5953259532595325, + "grad_norm": 65.29901885986328, + "learning_rate": 1.5882352941176473e-05, + "loss": 3.3981, + "step": 325 + }, + { + "epoch": 1.7183271832718328, + "grad_norm": 54.06962203979492, + "learning_rate": 1.7107843137254903e-05, + "loss": 3.1319, + "step": 350 + }, + { + "epoch": 1.841328413284133, + "grad_norm": 53.133201599121094, + "learning_rate": 1.8333333333333333e-05, + "loss": 3.0991, + "step": 375 + }, + { + "epoch": 1.964329643296433, + "grad_norm": 51.17352294921875, + "learning_rate": 1.9558823529411766e-05, + "loss": 2.8816, + "step": 400 + }, + { + "epoch": 2.0, + "eval_accuracy": 0.027662517289073305, + "eval_loss": 0.7307998538017273, + "eval_macro_f1": 0.3687256673408171, + "eval_micro_f1": 0.37910085054678005, + "eval_runtime": 2.0045, + "eval_samples_per_second": 360.689, + "eval_steps_per_second": 22.948, + "eval_weighted_f1": 0.465781434005995, + "step": 408 + }, + { + "epoch": 2.0836408364083643, + "grad_norm": 32.426170349121094, + "learning_rate": 1.9912854030501092e-05, + "loss": 2.4614, + "step": 425 + }, + { + "epoch": 2.206642066420664, + "grad_norm": 28.6024112701416, + "learning_rate": 1.9776688453159043e-05, + "loss": 2.2084, + "step": 450 + }, + { + "epoch": 2.3296432964329643, + "grad_norm": 40.920738220214844, + "learning_rate": 1.9640522875816993e-05, + "loss": 2.1315, + "step": 475 + }, + { + "epoch": 2.4526445264452645, + "grad_norm": 52.49609375, + "learning_rate": 1.9504357298474947e-05, + "loss": 2.2481, + "step": 500 + }, + { + "epoch": 2.5756457564575648, + "grad_norm": 50.98602294921875, + "learning_rate": 1.9368191721132898e-05, + "loss": 2.1713, + "step": 525 + }, + { + "epoch": 2.6986469864698646, + "grad_norm": 23.695022583007812, + "learning_rate": 1.9232026143790852e-05, + "loss": 2.2749, + "step": 550 + }, + { + "epoch": 2.821648216482165, + "grad_norm": 163.39114379882812, + "learning_rate": 1.9095860566448803e-05, + "loss": 2.1554, + "step": 575 + }, + { + "epoch": 2.944649446494465, + "grad_norm": 37.01408767700195, + "learning_rate": 1.8959694989106757e-05, + "loss": 2.1779, + "step": 600 + }, + { + "epoch": 3.0, + "eval_accuracy": 0.11756569847856155, + "eval_loss": 0.6780247688293457, + "eval_macro_f1": 0.44742742763662813, + "eval_micro_f1": 0.4847926267281106, + "eval_runtime": 2.0675, + "eval_samples_per_second": 349.693, + "eval_steps_per_second": 22.249, + "eval_weighted_f1": 0.5449678663096842, + "step": 612 + }, + { + "epoch": 3.063960639606396, + "grad_norm": 27.182884216308594, + "learning_rate": 1.8823529411764708e-05, + "loss": 1.6404, + "step": 625 + }, + { + "epoch": 3.1869618696186963, + "grad_norm": 28.162418365478516, + "learning_rate": 1.8687363834422662e-05, + "loss": 1.4181, + "step": 650 + }, + { + "epoch": 3.3099630996309966, + "grad_norm": 22.72654151916504, + "learning_rate": 1.8551198257080612e-05, + "loss": 1.4334, + "step": 675 + }, + { + "epoch": 3.4329643296432963, + "grad_norm": 23.24102210998535, + "learning_rate": 1.8415032679738563e-05, + "loss": 1.3488, + "step": 700 + }, + { + "epoch": 3.5559655596555966, + "grad_norm": 47.752838134765625, + "learning_rate": 1.8278867102396517e-05, + "loss": 1.2765, + "step": 725 + }, + { + "epoch": 3.678966789667897, + "grad_norm": 41.577880859375, + "learning_rate": 1.8142701525054468e-05, + "loss": 1.4195, + "step": 750 + }, + { + "epoch": 3.8019680196801966, + "grad_norm": 29.362136840820312, + "learning_rate": 1.800653594771242e-05, + "loss": 1.4393, + "step": 775 + }, + { + "epoch": 3.924969249692497, + "grad_norm": 41.29766845703125, + "learning_rate": 1.7870370370370373e-05, + "loss": 1.4676, + "step": 800 + }, + { + "epoch": 4.0, + "eval_accuracy": 0.1396957123098202, + "eval_loss": 0.7599406242370605, + "eval_macro_f1": 0.49790743621877676, + "eval_micro_f1": 0.5432230522945571, + "eval_runtime": 2.0136, + "eval_samples_per_second": 359.053, + "eval_steps_per_second": 22.844, + "eval_weighted_f1": 0.5845070792455435, + "step": 816 + }, + { + "epoch": 4.044280442804428, + "grad_norm": 24.567283630371094, + "learning_rate": 1.7734204793028323e-05, + "loss": 1.1482, + "step": 825 + }, + { + "epoch": 4.167281672816729, + "grad_norm": 18.17946434020996, + "learning_rate": 1.7598039215686274e-05, + "loss": 0.9048, + "step": 850 + }, + { + "epoch": 4.290282902829028, + "grad_norm": 30.113676071166992, + "learning_rate": 1.7461873638344228e-05, + "loss": 0.8644, + "step": 875 + }, + { + "epoch": 4.413284132841328, + "grad_norm": 19.796863555908203, + "learning_rate": 1.732570806100218e-05, + "loss": 0.9021, + "step": 900 + }, + { + "epoch": 4.536285362853628, + "grad_norm": 38.6467170715332, + "learning_rate": 1.7189542483660133e-05, + "loss": 0.8805, + "step": 925 + }, + { + "epoch": 4.659286592865929, + "grad_norm": 19.9637451171875, + "learning_rate": 1.7053376906318084e-05, + "loss": 0.8908, + "step": 950 + }, + { + "epoch": 4.782287822878229, + "grad_norm": 19.54431915283203, + "learning_rate": 1.6917211328976034e-05, + "loss": 0.8628, + "step": 975 + }, + { + "epoch": 4.905289052890529, + "grad_norm": 42.1319694519043, + "learning_rate": 1.678104575163399e-05, + "loss": 0.913, + "step": 1000 + }, + { + "epoch": 5.0, + "eval_accuracy": 0.21853388658367912, + "eval_loss": 0.8532860279083252, + "eval_macro_f1": 0.5446277224859946, + "eval_micro_f1": 0.6013943619278569, + "eval_runtime": 2.1705, + "eval_samples_per_second": 333.107, + "eval_steps_per_second": 21.194, + "eval_weighted_f1": 0.6197483086177735, + "step": 1020 + }, + { + "epoch": 5.02460024600246, + "grad_norm": 14.424633026123047, + "learning_rate": 1.664488017429194e-05, + "loss": 0.7874, + "step": 1025 + }, + { + "epoch": 5.14760147601476, + "grad_norm": 22.660926818847656, + "learning_rate": 1.6508714596949893e-05, + "loss": 0.5939, + "step": 1050 + }, + { + "epoch": 5.27060270602706, + "grad_norm": 10.341612815856934, + "learning_rate": 1.6372549019607844e-05, + "loss": 0.6016, + "step": 1075 + }, + { + "epoch": 5.39360393603936, + "grad_norm": 13.016510963439941, + "learning_rate": 1.6236383442265798e-05, + "loss": 0.5896, + "step": 1100 + }, + { + "epoch": 5.516605166051661, + "grad_norm": 12.98374080657959, + "learning_rate": 1.610021786492375e-05, + "loss": 0.5628, + "step": 1125 + }, + { + "epoch": 5.639606396063961, + "grad_norm": 18.153182983398438, + "learning_rate": 1.5964052287581703e-05, + "loss": 0.5384, + "step": 1150 + }, + { + "epoch": 5.762607626076261, + "grad_norm": 31.739131927490234, + "learning_rate": 1.5827886710239654e-05, + "loss": 0.588, + "step": 1175 + }, + { + "epoch": 5.885608856088561, + "grad_norm": 17.49053192138672, + "learning_rate": 1.5691721132897604e-05, + "loss": 0.5548, + "step": 1200 + }, + { + "epoch": 6.0, + "eval_accuracy": 0.24896265560165975, + "eval_loss": 0.9286503195762634, + "eval_macro_f1": 0.5594913430296943, + "eval_micro_f1": 0.6213844653883653, + "eval_runtime": 2.1194, + "eval_samples_per_second": 341.126, + "eval_steps_per_second": 21.704, + "eval_weighted_f1": 0.6305536481148704, + "step": 1224 + }, + { + "epoch": 6.004920049200492, + "grad_norm": 11.133593559265137, + "learning_rate": 1.555555555555556e-05, + "loss": 0.5617, + "step": 1225 + }, + { + "epoch": 6.127921279212792, + "grad_norm": 19.947908401489258, + "learning_rate": 1.541938997821351e-05, + "loss": 0.3871, + "step": 1250 + }, + { + "epoch": 6.250922509225092, + "grad_norm": 10.981277465820312, + "learning_rate": 1.528322440087146e-05, + "loss": 0.3695, + "step": 1275 + }, + { + "epoch": 6.373923739237393, + "grad_norm": 10.755515098571777, + "learning_rate": 1.5147058823529412e-05, + "loss": 0.3878, + "step": 1300 + }, + { + "epoch": 6.496924969249693, + "grad_norm": 15.75013256072998, + "learning_rate": 1.5010893246187366e-05, + "loss": 0.3848, + "step": 1325 + }, + { + "epoch": 6.619926199261993, + "grad_norm": 16.140092849731445, + "learning_rate": 1.4874727668845317e-05, + "loss": 0.3697, + "step": 1350 + }, + { + "epoch": 6.7429274292742925, + "grad_norm": 7.486774444580078, + "learning_rate": 1.473856209150327e-05, + "loss": 0.3593, + "step": 1375 + }, + { + "epoch": 6.865928659286593, + "grad_norm": 14.175186157226562, + "learning_rate": 1.4602396514161222e-05, + "loss": 0.359, + "step": 1400 + }, + { + "epoch": 6.988929889298893, + "grad_norm": 10.35640811920166, + "learning_rate": 1.4466230936819172e-05, + "loss": 0.3633, + "step": 1425 + }, + { + "epoch": 7.0, + "eval_accuracy": 0.2987551867219917, + "eval_loss": 1.0920287370681763, + "eval_macro_f1": 0.5689211077657381, + "eval_micro_f1": 0.6445804825351098, + "eval_runtime": 2.0034, + "eval_samples_per_second": 360.879, + "eval_steps_per_second": 22.961, + "eval_weighted_f1": 0.6444084807872194, + "step": 1428 + }, + { + "epoch": 7.108241082410824, + "grad_norm": 8.464420318603516, + "learning_rate": 1.4330065359477125e-05, + "loss": 0.2676, + "step": 1450 + }, + { + "epoch": 7.231242312423125, + "grad_norm": 14.645303726196289, + "learning_rate": 1.4193899782135077e-05, + "loss": 0.2554, + "step": 1475 + }, + { + "epoch": 7.354243542435424, + "grad_norm": 11.637530326843262, + "learning_rate": 1.405773420479303e-05, + "loss": 0.2556, + "step": 1500 + }, + { + "epoch": 7.477244772447724, + "grad_norm": 6.641851425170898, + "learning_rate": 1.392156862745098e-05, + "loss": 0.2307, + "step": 1525 + }, + { + "epoch": 7.6002460024600245, + "grad_norm": 10.870826721191406, + "learning_rate": 1.3785403050108934e-05, + "loss": 0.2557, + "step": 1550 + }, + { + "epoch": 7.723247232472325, + "grad_norm": 5.410327434539795, + "learning_rate": 1.3649237472766885e-05, + "loss": 0.2265, + "step": 1575 + }, + { + "epoch": 7.846248462484625, + "grad_norm": 8.480887413024902, + "learning_rate": 1.351307189542484e-05, + "loss": 0.2693, + "step": 1600 + }, + { + "epoch": 7.969249692496925, + "grad_norm": 18.576095581054688, + "learning_rate": 1.337690631808279e-05, + "loss": 0.255, + "step": 1625 + }, + { + "epoch": 8.0, + "eval_accuracy": 0.313969571230982, + "eval_loss": 1.2188948392868042, + "eval_macro_f1": 0.5811899137156786, + "eval_micro_f1": 0.6468622354251764, + "eval_runtime": 2.497, + "eval_samples_per_second": 289.543, + "eval_steps_per_second": 18.422, + "eval_weighted_f1": 0.6483152693096889, + "step": 1632 + }, + { + "epoch": 8.088560885608857, + "grad_norm": 5.418167591094971, + "learning_rate": 1.3240740740740742e-05, + "loss": 0.1742, + "step": 1650 + }, + { + "epoch": 8.211562115621156, + "grad_norm": 6.825408935546875, + "learning_rate": 1.3104575163398695e-05, + "loss": 0.1686, + "step": 1675 + }, + { + "epoch": 8.334563345633457, + "grad_norm": 6.550326824188232, + "learning_rate": 1.2968409586056645e-05, + "loss": 0.1649, + "step": 1700 + }, + { + "epoch": 8.457564575645756, + "grad_norm": 4.261825084686279, + "learning_rate": 1.2832244008714598e-05, + "loss": 0.1437, + "step": 1725 + }, + { + "epoch": 8.580565805658056, + "grad_norm": 5.172179222106934, + "learning_rate": 1.269607843137255e-05, + "loss": 0.1642, + "step": 1750 + }, + { + "epoch": 8.703567035670357, + "grad_norm": 23.130123138427734, + "learning_rate": 1.2559912854030503e-05, + "loss": 0.1641, + "step": 1775 + }, + { + "epoch": 8.826568265682656, + "grad_norm": 30.504459381103516, + "learning_rate": 1.2423747276688453e-05, + "loss": 0.1669, + "step": 1800 + }, + { + "epoch": 8.949569495694957, + "grad_norm": 8.059348106384277, + "learning_rate": 1.2287581699346407e-05, + "loss": 0.1654, + "step": 1825 + }, + { + "epoch": 9.0, + "eval_accuracy": 0.3278008298755187, + "eval_loss": 1.3381789922714233, + "eval_macro_f1": 0.5762311884316192, + "eval_micro_f1": 0.648936170212766, + "eval_runtime": 2.0185, + "eval_samples_per_second": 358.184, + "eval_steps_per_second": 22.789, + "eval_weighted_f1": 0.6484288983365116, + "step": 1836 + }, + { + "epoch": 9.068880688806889, + "grad_norm": 6.1497297286987305, + "learning_rate": 1.2151416122004358e-05, + "loss": 0.135, + "step": 1850 + }, + { + "epoch": 9.191881918819188, + "grad_norm": 5.762045860290527, + "learning_rate": 1.2015250544662312e-05, + "loss": 0.1209, + "step": 1875 + }, + { + "epoch": 9.314883148831488, + "grad_norm": 4.636457443237305, + "learning_rate": 1.1879084967320263e-05, + "loss": 0.1211, + "step": 1900 + }, + { + "epoch": 9.437884378843789, + "grad_norm": 6.796161651611328, + "learning_rate": 1.1742919389978214e-05, + "loss": 0.1124, + "step": 1925 + }, + { + "epoch": 9.560885608856088, + "grad_norm": 4.42055606842041, + "learning_rate": 1.1606753812636166e-05, + "loss": 0.1129, + "step": 1950 + }, + { + "epoch": 9.68388683886839, + "grad_norm": 17.374441146850586, + "learning_rate": 1.1470588235294118e-05, + "loss": 0.1064, + "step": 1975 + }, + { + "epoch": 9.806888068880689, + "grad_norm": 2.4789116382598877, + "learning_rate": 1.133442265795207e-05, + "loss": 0.113, + "step": 2000 + }, + { + "epoch": 9.92988929889299, + "grad_norm": 11.736351013183594, + "learning_rate": 1.1198257080610021e-05, + "loss": 0.1101, + "step": 2025 + }, + { + "epoch": 10.0, + "eval_accuracy": 0.3333333333333333, + "eval_loss": 1.4571764469146729, + "eval_macro_f1": 0.571516196920113, + "eval_micro_f1": 0.6534730306558013, + "eval_runtime": 2.043, + "eval_samples_per_second": 353.897, + "eval_steps_per_second": 22.516, + "eval_weighted_f1": 0.6500766986172553, + "step": 2040 + }, + { + "epoch": 10.04920049200492, + "grad_norm": 4.998640537261963, + "learning_rate": 1.1062091503267976e-05, + "loss": 0.0896, + "step": 2050 + }, + { + "epoch": 10.17220172201722, + "grad_norm": 2.433173179626465, + "learning_rate": 1.0925925925925926e-05, + "loss": 0.0786, + "step": 2075 + }, + { + "epoch": 10.29520295202952, + "grad_norm": 2.942702531814575, + "learning_rate": 1.078976034858388e-05, + "loss": 0.0839, + "step": 2100 + }, + { + "epoch": 10.418204182041821, + "grad_norm": 4.1860785484313965, + "learning_rate": 1.0653594771241831e-05, + "loss": 0.0809, + "step": 2125 + }, + { + "epoch": 10.54120541205412, + "grad_norm": 4.45076847076416, + "learning_rate": 1.0517429193899782e-05, + "loss": 0.0713, + "step": 2150 + }, + { + "epoch": 10.664206642066421, + "grad_norm": 5.46282958984375, + "learning_rate": 1.0381263616557736e-05, + "loss": 0.0877, + "step": 2175 + }, + { + "epoch": 10.78720787207872, + "grad_norm": 6.310950756072998, + "learning_rate": 1.0245098039215687e-05, + "loss": 0.081, + "step": 2200 + }, + { + "epoch": 10.91020910209102, + "grad_norm": 3.4199132919311523, + "learning_rate": 1.0108932461873639e-05, + "loss": 0.0725, + "step": 2225 + }, + { + "epoch": 11.0, + "eval_accuracy": 0.34716459197787, + "eval_loss": 1.595555067062378, + "eval_macro_f1": 0.5854704605866574, + "eval_micro_f1": 0.6532744073925271, + "eval_runtime": 2.0731, + "eval_samples_per_second": 348.749, + "eval_steps_per_second": 22.189, + "eval_weighted_f1": 0.650225101846109, + "step": 2244 + }, + { + "epoch": 11.029520295202952, + "grad_norm": 1.6294690370559692, + "learning_rate": 9.972766884531591e-06, + "loss": 0.0702, + "step": 2250 + }, + { + "epoch": 11.152521525215253, + "grad_norm": 2.9956741333007812, + "learning_rate": 9.836601307189542e-06, + "loss": 0.0583, + "step": 2275 + }, + { + "epoch": 11.275522755227552, + "grad_norm": 2.8591299057006836, + "learning_rate": 9.700435729847494e-06, + "loss": 0.0609, + "step": 2300 + }, + { + "epoch": 11.398523985239853, + "grad_norm": 20.983976364135742, + "learning_rate": 9.564270152505447e-06, + "loss": 0.0647, + "step": 2325 + }, + { + "epoch": 11.521525215252153, + "grad_norm": 2.2769386768341064, + "learning_rate": 9.4281045751634e-06, + "loss": 0.0549, + "step": 2350 + }, + { + "epoch": 11.644526445264452, + "grad_norm": 2.289628028869629, + "learning_rate": 9.291938997821352e-06, + "loss": 0.0509, + "step": 2375 + }, + { + "epoch": 11.767527675276753, + "grad_norm": 3.487907648086548, + "learning_rate": 9.155773420479304e-06, + "loss": 0.0489, + "step": 2400 + }, + { + "epoch": 11.890528905289052, + "grad_norm": 41.98941421508789, + "learning_rate": 9.019607843137256e-06, + "loss": 0.0551, + "step": 2425 + }, + { + "epoch": 12.0, + "eval_accuracy": 0.38589211618257263, + "eval_loss": 1.6902517080307007, + "eval_macro_f1": 0.5753988585917662, + "eval_micro_f1": 0.6636661211129297, + "eval_runtime": 2.0295, + "eval_samples_per_second": 356.241, + "eval_steps_per_second": 22.665, + "eval_weighted_f1": 0.6572887032240203, + "step": 2448 + }, + { + "epoch": 12.009840098400984, + "grad_norm": 2.556194305419922, + "learning_rate": 8.883442265795209e-06, + "loss": 0.0532, + "step": 2450 + }, + { + "epoch": 12.132841328413285, + "grad_norm": 3.410416841506958, + "learning_rate": 8.74727668845316e-06, + "loss": 0.043, + "step": 2475 + }, + { + "epoch": 12.255842558425584, + "grad_norm": 1.7650806903839111, + "learning_rate": 8.611111111111112e-06, + "loss": 0.041, + "step": 2500 + }, + { + "epoch": 12.378843788437884, + "grad_norm": 2.7926206588745117, + "learning_rate": 8.474945533769063e-06, + "loss": 0.0357, + "step": 2525 + }, + { + "epoch": 12.501845018450185, + "grad_norm": 1.835835576057434, + "learning_rate": 8.338779956427015e-06, + "loss": 0.0361, + "step": 2550 + }, + { + "epoch": 12.624846248462484, + "grad_norm": 2.1783931255340576, + "learning_rate": 8.202614379084967e-06, + "loss": 0.0374, + "step": 2575 + }, + { + "epoch": 12.747847478474785, + "grad_norm": 1.9006906747817993, + "learning_rate": 8.06644880174292e-06, + "loss": 0.0397, + "step": 2600 + }, + { + "epoch": 12.870848708487085, + "grad_norm": 1.8903752565383911, + "learning_rate": 7.930283224400872e-06, + "loss": 0.0456, + "step": 2625 + }, + { + "epoch": 12.993849938499386, + "grad_norm": 1.8099350929260254, + "learning_rate": 7.794117647058825e-06, + "loss": 0.041, + "step": 2650 + }, + { + "epoch": 13.0, + "eval_accuracy": 0.3789764868603043, + "eval_loss": 1.8257192373275757, + "eval_macro_f1": 0.581498457268201, + "eval_micro_f1": 0.6696320258794985, + "eval_runtime": 2.0273, + "eval_samples_per_second": 356.635, + "eval_steps_per_second": 22.69, + "eval_weighted_f1": 0.6628037115578727, + "step": 2652 + }, + { + "epoch": 13.113161131611315, + "grad_norm": 0.8951275944709778, + "learning_rate": 7.657952069716777e-06, + "loss": 0.0311, + "step": 2675 + }, + { + "epoch": 13.236162361623617, + "grad_norm": 0.5448098182678223, + "learning_rate": 7.5217864923747285e-06, + "loss": 0.0289, + "step": 2700 + }, + { + "epoch": 13.359163591635916, + "grad_norm": 3.7454726696014404, + "learning_rate": 7.385620915032681e-06, + "loss": 0.0311, + "step": 2725 + }, + { + "epoch": 13.482164821648217, + "grad_norm": 3.0179667472839355, + "learning_rate": 7.249455337690633e-06, + "loss": 0.0249, + "step": 2750 + }, + { + "epoch": 13.605166051660516, + "grad_norm": 2.5436480045318604, + "learning_rate": 7.113289760348584e-06, + "loss": 0.028, + "step": 2775 + }, + { + "epoch": 13.728167281672818, + "grad_norm": 1.0340428352355957, + "learning_rate": 6.977124183006536e-06, + "loss": 0.029, + "step": 2800 + }, + { + "epoch": 13.851168511685117, + "grad_norm": 1.1028094291687012, + "learning_rate": 6.840958605664488e-06, + "loss": 0.0582, + "step": 2825 + }, + { + "epoch": 13.974169741697416, + "grad_norm": 1.1252424716949463, + "learning_rate": 6.70479302832244e-06, + "loss": 0.0272, + "step": 2850 + }, + { + "epoch": 14.0, + "eval_accuracy": 0.36099585062240663, + "eval_loss": 1.857070803642273, + "eval_macro_f1": 0.5820218138308284, + "eval_micro_f1": 0.6585858585858586, + "eval_runtime": 2.021, + "eval_samples_per_second": 357.748, + "eval_steps_per_second": 22.761, + "eval_weighted_f1": 0.6538994713907645, + "step": 2856 + } + ], + "logging_steps": 25, + "max_steps": 4080, + "num_input_tokens_seen": 0, + "num_train_epochs": 20, + "save_steps": 500, + "stateful_callbacks": { + "EarlyStoppingCallback": { + "args": { + "early_stopping_patience": 3, + "early_stopping_threshold": 0.0 + }, + "attributes": { + "early_stopping_patience_counter": 3 + } + }, + "TrainerControl": { + "args": { + "should_epoch_stop": false, + "should_evaluate": false, + "should_log": false, + "should_save": true, + "should_training_stop": true + }, + "attributes": {} + } + }, + "total_flos": 0.0, + "train_batch_size": 8, + "trial_name": null, + "trial_params": null +} diff --git a/ood-vigoemotions-mmbert-source-source-01-s42-20260905T110803Z-bd528b37/model.safetensors b/ood-vigoemotions-mmbert-source-source-01-s42-20260905T110803Z-bd528b37/model.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..ae1a19d68933dc81a185bb9c60f4717813cb6a3f --- /dev/null +++ b/ood-vigoemotions-mmbert-source-source-01-s42-20260905T110803Z-bd528b37/model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:59fdbf190b9f08117a207ea3a2c18b0606f13ac7a675598abb45b45678451f9f +size 1227859104 diff --git a/ood-vigoemotions-mmbert-source-source-01-s43-20260905T125625Z-60628050/model.safetensors b/ood-vigoemotions-mmbert-source-source-01-s43-20260905T125625Z-60628050/model.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..ec6498d4d4e5cb3296757db86c1919761ffe4567 --- /dev/null +++ b/ood-vigoemotions-mmbert-source-source-01-s43-20260905T125625Z-60628050/model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:63dbe82f28a8db4e3193641b20dc1aac0b66e19e0bba8a06cb6d2683c0c74351 +size 1227859104 diff --git a/ood-vigoemotions-mmbert-source-source-01-s44-20260905T144605Z-37344fee/model.safetensors b/ood-vigoemotions-mmbert-source-source-01-s44-20260905T144605Z-37344fee/model.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..f8f90234a3080e4629fb6ec6fe89fd431e65b3b9 --- /dev/null +++ b/ood-vigoemotions-mmbert-source-source-01-s44-20260905T144605Z-37344fee/model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:3213afb176ab9b364d655b256de7ecd1f2e5fc8e0cb4abdefa64f6319e350400 +size 1227859104 diff --git a/ood-vigoemotions-mmbert-source-source-01-s44-20260905T144605Z-37344fee/trainer/checkpoint-3632/optimizer.pt b/ood-vigoemotions-mmbert-source-source-01-s44-20260905T144605Z-37344fee/trainer/checkpoint-3632/optimizer.pt new file mode 100644 index 0000000000000000000000000000000000000000..e6c949039648167fcb78623e3e4843dd61c9d85f --- /dev/null +++ b/ood-vigoemotions-mmbert-source-source-01-s44-20260905T144605Z-37344fee/trainer/checkpoint-3632/optimizer.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:d767dd0c22fa36aa59cec96a35f4254d91bc87e482ee3a2dfbaf5f7d5d2abf4c +size 2455803898 diff --git a/ood-vigoemotions-mmbert-source-source-01-s44-20260905T144605Z-37344fee/trainer/checkpoint-3632/special_tokens_map.json b/ood-vigoemotions-mmbert-source-source-01-s44-20260905T144605Z-37344fee/trainer/checkpoint-3632/special_tokens_map.json new file mode 100644 index 0000000000000000000000000000000000000000..be4ad795f49075ac5e56d8108ef63be314a1e967 --- /dev/null +++ b/ood-vigoemotions-mmbert-source-source-01-s44-20260905T144605Z-37344fee/trainer/checkpoint-3632/special_tokens_map.json @@ -0,0 +1,55 @@ +{ + "additional_special_tokens": [ + "", + "" + ], + "bos_token": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false + }, + "cls_token": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false + }, + "eos_token": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false + }, + "mask_token": { + "content": "", + "lstrip": true, + "normalized": false, + "rstrip": false, + "single_word": false + }, + "pad_token": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false + }, + "sep_token": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false + }, + "unk_token": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false + } +} diff --git a/ood-vigoemotions-mmbert-source-source-01-s44-20260905T144605Z-37344fee/trainer/checkpoint-3632/tokenizer_config.json b/ood-vigoemotions-mmbert-source-source-01-s44-20260905T144605Z-37344fee/trainer/checkpoint-3632/tokenizer_config.json new file mode 100644 index 0000000000000000000000000000000000000000..b002fb8c2960be8ebf9c373d6e5784b4565d7edc --- /dev/null +++ b/ood-vigoemotions-mmbert-source-source-01-s44-20260905T144605Z-37344fee/trainer/checkpoint-3632/tokenizer_config.json @@ -0,0 +1,2018 @@ +{ + "add_bos_token": true, + "added_tokens_decoder": { + "0": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "1": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "2": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "3": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "4": { + "content": "", + "lstrip": true, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "5": { + "content": "<2mass>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "6": { + "content": "[@BOS@]", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "7": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "8": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "9": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "10": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "11": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "12": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "13": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "14": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "15": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "16": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "17": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "18": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "19": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "20": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "21": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "22": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "23": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "24": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "25": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "26": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "27": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "28": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "29": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "30": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "31": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "32": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "33": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "34": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "35": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "36": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "37": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "38": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "39": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "40": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "41": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "42": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "43": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "44": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "45": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "46": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "47": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "48": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "49": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "50": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "51": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "52": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "53": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "54": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "55": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "56": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "57": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "58": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "59": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "60": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "61": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "62": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "63": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "64": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "65": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "66": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "67": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "68": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "69": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "70": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "71": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "72": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "73": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "74": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "75": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "76": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "77": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "78": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "79": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "80": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "81": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "82": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "83": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "84": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "85": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "86": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "87": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "88": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "89": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "90": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "91": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "92": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "93": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "94": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "95": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "96": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "97": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "98": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "99": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "100": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "101": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "102": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "103": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "104": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "105": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "106": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "107": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "108": { + "content": "\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "109": { + "content": "\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "110": { + "content": "\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "111": { + "content": "\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "112": { + "content": "\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "113": { + "content": "\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "114": { + "content": "\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "115": { + "content": "\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "116": { + "content": "\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "117": { + "content": "\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "118": { + "content": "\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "119": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "120": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "121": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "122": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "123": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "124": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "125": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "126": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "127": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "128": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "129": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "130": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "131": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "132": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "133": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "134": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "135": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "136": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "137": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "138": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "139": { + "content": "▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "140": { + "content": "▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "141": { + "content": "▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "142": { + "content": "▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "143": { + "content": "▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "144": { + "content": "▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "145": { + "content": "▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "146": { + "content": "▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "147": { + "content": "▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "148": { + "content": "▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "149": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "150": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "151": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "152": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "153": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "154": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "155": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "156": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "157": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "158": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "159": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "160": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "161": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "162": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "163": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "164": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "165": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "166": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "167": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "168": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "169": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "170": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "172": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "173": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "174": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "175": { + "content": "
", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "171": { + "content": "
", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "176": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "177": { + "content": "
", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "178": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "179": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "180": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "181": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "182": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "183": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "184": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "185": { + "content": "

", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "186": { + "content": "

", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "187": { + "content": "

", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "188": { + "content": "

", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "189": { + "content": "

", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "190": { + "content": "
", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "191": { + "content": "
", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "192": { + "content": "
", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "193": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "194": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "195": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "196": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "197": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "198": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "199": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "200": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "201": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "202": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "203": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "204": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "205": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "206": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "207": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "208": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "209": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "210": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "211": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "212": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "213": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "214": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "215": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "216": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255968": { + "content": "[toxicity=0]", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255969": { + "content": "\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255970": { + "content": "\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255971": { + "content": "\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255972": { + "content": "\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255973": { + "content": "\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255974": { + "content": "\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255975": { + "content": "\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255976": { + "content": "\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255977": { + "content": "\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255978": { + "content": "\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255979": { + "content": "\t\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255980": { + "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255981": { + "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255982": { + "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255983": { + "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255984": { + "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255985": { + "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255986": { + "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255987": { + "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255988": { + "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255989": { + "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255990": { + "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255991": { + "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255992": { + "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255993": { + "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255994": { + "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255995": { + "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255996": { + "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255997": { + "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255998": { + "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255999": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + } + }, + "additional_special_tokens": [ + "", + "" + ], + "bos_token": "", + "clean_up_tokenization_spaces": false, + "cls_token": "", + "eos_token": "", + "extra_special_tokens": {}, + "mask_token": "", + "model_input_names": [ + "input_ids", + "attention_mask" + ], + "model_max_length": 8192, + "pad_token": "", + "padding_side": "right", + "sep_token": "", + "spaces_between_special_tokens": false, + "tokenizer_class": "PreTrainedTokenizerFast", + "unk_token": "" +} diff --git a/ood-vigoemotions-mmbert-source-source-01-s44-20260905T144605Z-37344fee/trainer/checkpoint-3632/trainer_state.json b/ood-vigoemotions-mmbert-source-source-01-s44-20260905T144605Z-37344fee/trainer/checkpoint-3632/trainer_state.json new file mode 100644 index 0000000000000000000000000000000000000000..206fdbdcad41a36c469a8712e1774c27cb7419e7 --- /dev/null +++ b/ood-vigoemotions-mmbert-source-source-01-s44-20260905T144605Z-37344fee/trainer/checkpoint-3632/trainer_state.json @@ -0,0 +1,1154 @@ +{ + "best_global_step": 3632, + "best_metric": 0.550203038718693, + "best_model_checkpoint": "/root/VNEncoder/results/checkpoints/ood-vigoemotions-mmbert-source-source-01-s44-20260905T144605Z-37344fee/trainer/checkpoint-3632", + "epoch": 8.0, + "eval_steps": 500, + "global_step": 3632, + "is_hyper_param_search": false, + "is_local_process_zero": true, + "is_world_process_zero": true, + "log_history": [ + { + "epoch": 0.05509641873278237, + "grad_norm": 42.07517623901367, + "learning_rate": 5.286343612334802e-07, + "loss": 6.2699, + "step": 25 + }, + { + "epoch": 0.11019283746556474, + "grad_norm": 34.53789520263672, + "learning_rate": 1.0792951541850223e-06, + "loss": 6.0286, + "step": 50 + }, + { + "epoch": 0.1652892561983471, + "grad_norm": 56.00322341918945, + "learning_rate": 1.629955947136564e-06, + "loss": 6.3079, + "step": 75 + }, + { + "epoch": 0.22038567493112948, + "grad_norm": 34.499778747558594, + "learning_rate": 2.180616740088106e-06, + "loss": 6.3055, + "step": 100 + }, + { + "epoch": 0.27548209366391185, + "grad_norm": 49.277896881103516, + "learning_rate": 2.731277533039648e-06, + "loss": 6.0817, + "step": 125 + }, + { + "epoch": 0.3305785123966942, + "grad_norm": 293.85943603515625, + "learning_rate": 3.2819383259911898e-06, + "loss": 5.8855, + "step": 150 + }, + { + "epoch": 0.3856749311294766, + "grad_norm": 170.3839874267578, + "learning_rate": 3.8325991189427315e-06, + "loss": 5.4578, + "step": 175 + }, + { + "epoch": 0.44077134986225897, + "grad_norm": 54.24516296386719, + "learning_rate": 4.383259911894274e-06, + "loss": 5.1302, + "step": 200 + }, + { + "epoch": 0.49586776859504134, + "grad_norm": 76.05721282958984, + "learning_rate": 4.933920704845816e-06, + "loss": 5.1066, + "step": 225 + }, + { + "epoch": 0.5509641873278237, + "grad_norm": 61.08894348144531, + "learning_rate": 5.484581497797358e-06, + "loss": 4.9914, + "step": 250 + }, + { + "epoch": 0.6060606060606061, + "grad_norm": 47.69670486450195, + "learning_rate": 6.035242290748899e-06, + "loss": 4.8303, + "step": 275 + }, + { + "epoch": 0.6611570247933884, + "grad_norm": 149.77024841308594, + "learning_rate": 6.585903083700441e-06, + "loss": 4.502, + "step": 300 + }, + { + "epoch": 0.7162534435261708, + "grad_norm": 50.70989990234375, + "learning_rate": 7.136563876651983e-06, + "loss": 4.383, + "step": 325 + }, + { + "epoch": 0.7713498622589532, + "grad_norm": 43.84785079956055, + "learning_rate": 7.687224669603525e-06, + "loss": 4.3155, + "step": 350 + }, + { + "epoch": 0.8264462809917356, + "grad_norm": 50.88126754760742, + "learning_rate": 8.237885462555067e-06, + "loss": 4.0407, + "step": 375 + }, + { + "epoch": 0.8815426997245179, + "grad_norm": 47.921146392822266, + "learning_rate": 8.788546255506607e-06, + "loss": 3.6664, + "step": 400 + }, + { + "epoch": 0.9366391184573003, + "grad_norm": 36.562015533447266, + "learning_rate": 9.339207048458151e-06, + "loss": 3.8568, + "step": 425 + }, + { + "epoch": 0.9917355371900827, + "grad_norm": 96.27021026611328, + "learning_rate": 9.889867841409693e-06, + "loss": 3.7885, + "step": 450 + }, + { + "epoch": 1.0, + "eval_accuracy": 0.0, + "eval_loss": 0.8759939670562744, + "eval_macro_f1": 0.30856328770558283, + "eval_micro_f1": 0.31929322466247745, + "eval_runtime": 5.2481, + "eval_samples_per_second": 313.064, + "eval_steps_per_second": 19.626, + "eval_weighted_f1": 0.38774477435725324, + "step": 454 + }, + { + "epoch": 1.0462809917355371, + "grad_norm": 45.211910247802734, + "learning_rate": 1.0440528634361234e-05, + "loss": 3.2569, + "step": 475 + }, + { + "epoch": 1.1013774104683196, + "grad_norm": 39.812557220458984, + "learning_rate": 1.0991189427312776e-05, + "loss": 3.2178, + "step": 500 + }, + { + "epoch": 1.1564738292011019, + "grad_norm": 93.90754699707031, + "learning_rate": 1.154185022026432e-05, + "loss": 3.3733, + "step": 525 + }, + { + "epoch": 1.2115702479338843, + "grad_norm": 50.11622619628906, + "learning_rate": 1.209251101321586e-05, + "loss": 3.2525, + "step": 550 + }, + { + "epoch": 1.2666666666666666, + "grad_norm": 54.41451644897461, + "learning_rate": 1.2643171806167402e-05, + "loss": 3.0884, + "step": 575 + }, + { + "epoch": 1.321763085399449, + "grad_norm": 48.95606231689453, + "learning_rate": 1.3193832599118943e-05, + "loss": 3.1357, + "step": 600 + }, + { + "epoch": 1.3768595041322313, + "grad_norm": 62.82511520385742, + "learning_rate": 1.3744493392070487e-05, + "loss": 3.3523, + "step": 625 + }, + { + "epoch": 1.4319559228650138, + "grad_norm": 49.2446403503418, + "learning_rate": 1.4295154185022027e-05, + "loss": 2.9562, + "step": 650 + }, + { + "epoch": 1.487052341597796, + "grad_norm": 34.752410888671875, + "learning_rate": 1.484581497797357e-05, + "loss": 3.0026, + "step": 675 + }, + { + "epoch": 1.5421487603305786, + "grad_norm": 50.21882247924805, + "learning_rate": 1.539647577092511e-05, + "loss": 3.2388, + "step": 700 + }, + { + "epoch": 1.5972451790633608, + "grad_norm": 72.29084014892578, + "learning_rate": 1.5947136563876654e-05, + "loss": 3.0325, + "step": 725 + }, + { + "epoch": 1.652341597796143, + "grad_norm": 32.43894577026367, + "learning_rate": 1.6497797356828196e-05, + "loss": 3.0251, + "step": 750 + }, + { + "epoch": 1.7074380165289256, + "grad_norm": 40.4652099609375, + "learning_rate": 1.7048458149779738e-05, + "loss": 2.9036, + "step": 775 + }, + { + "epoch": 1.762534435261708, + "grad_norm": 52.69956588745117, + "learning_rate": 1.759911894273128e-05, + "loss": 2.9234, + "step": 800 + }, + { + "epoch": 1.8176308539944903, + "grad_norm": 30.133697509765625, + "learning_rate": 1.814977973568282e-05, + "loss": 2.9855, + "step": 825 + }, + { + "epoch": 1.8727272727272726, + "grad_norm": 61.72572708129883, + "learning_rate": 1.8700440528634364e-05, + "loss": 3.0321, + "step": 850 + }, + { + "epoch": 1.927823691460055, + "grad_norm": 28.934057235717773, + "learning_rate": 1.9251101321585906e-05, + "loss": 2.985, + "step": 875 + }, + { + "epoch": 1.9829201101928375, + "grad_norm": 39.00049591064453, + "learning_rate": 1.9801762114537445e-05, + "loss": 2.906, + "step": 900 + }, + { + "epoch": 2.0, + "eval_accuracy": 0.013390139987827145, + "eval_loss": 0.7132589221000671, + "eval_macro_f1": 0.3972028926547339, + "eval_micro_f1": 0.4105655588165305, + "eval_runtime": 4.347, + "eval_samples_per_second": 377.962, + "eval_steps_per_second": 23.695, + "eval_weighted_f1": 0.4630428399620474, + "step": 908 + }, + { + "epoch": 2.037465564738292, + "grad_norm": 24.804983139038086, + "learning_rate": 1.9960841899167892e-05, + "loss": 2.4633, + "step": 925 + }, + { + "epoch": 2.0925619834710742, + "grad_norm": 64.12801361083984, + "learning_rate": 1.989965736661772e-05, + "loss": 2.306, + "step": 950 + }, + { + "epoch": 2.1476584022038567, + "grad_norm": 27.915681838989258, + "learning_rate": 1.983847283406755e-05, + "loss": 2.3691, + "step": 975 + }, + { + "epoch": 2.202754820936639, + "grad_norm": 24.95664405822754, + "learning_rate": 1.9777288301517378e-05, + "loss": 2.293, + "step": 1000 + }, + { + "epoch": 2.2578512396694217, + "grad_norm": 25.19192886352539, + "learning_rate": 1.9716103768967207e-05, + "loss": 2.234, + "step": 1025 + }, + { + "epoch": 2.3129476584022037, + "grad_norm": 28.52288246154785, + "learning_rate": 1.9654919236417035e-05, + "loss": 2.1174, + "step": 1050 + }, + { + "epoch": 2.368044077134986, + "grad_norm": 48.03308868408203, + "learning_rate": 1.9593734703866864e-05, + "loss": 2.2298, + "step": 1075 + }, + { + "epoch": 2.4231404958677687, + "grad_norm": 43.81739807128906, + "learning_rate": 1.9532550171316693e-05, + "loss": 2.2613, + "step": 1100 + }, + { + "epoch": 2.4782369146005507, + "grad_norm": 29.52232551574707, + "learning_rate": 1.947136563876652e-05, + "loss": 2.257, + "step": 1125 + }, + { + "epoch": 2.533333333333333, + "grad_norm": 30.75737190246582, + "learning_rate": 1.941018110621635e-05, + "loss": 2.2481, + "step": 1150 + }, + { + "epoch": 2.5884297520661157, + "grad_norm": 35.87025451660156, + "learning_rate": 1.934899657366618e-05, + "loss": 2.2418, + "step": 1175 + }, + { + "epoch": 2.643526170798898, + "grad_norm": 45.66483688354492, + "learning_rate": 1.9287812041116007e-05, + "loss": 2.2211, + "step": 1200 + }, + { + "epoch": 2.6986225895316807, + "grad_norm": 46.8340950012207, + "learning_rate": 1.9226627508565836e-05, + "loss": 2.2759, + "step": 1225 + }, + { + "epoch": 2.7537190082644627, + "grad_norm": 33.303741455078125, + "learning_rate": 1.9165442976015665e-05, + "loss": 2.1401, + "step": 1250 + }, + { + "epoch": 2.808815426997245, + "grad_norm": 44.45813751220703, + "learning_rate": 1.9104258443465493e-05, + "loss": 2.2384, + "step": 1275 + }, + { + "epoch": 2.8639118457300277, + "grad_norm": 28.29966163635254, + "learning_rate": 1.9043073910915322e-05, + "loss": 2.161, + "step": 1300 + }, + { + "epoch": 2.9190082644628097, + "grad_norm": 35.3993034362793, + "learning_rate": 1.898188937836515e-05, + "loss": 2.1901, + "step": 1325 + }, + { + "epoch": 2.974104683195592, + "grad_norm": 47.99441146850586, + "learning_rate": 1.892070484581498e-05, + "loss": 2.0656, + "step": 1350 + }, + { + "epoch": 3.0, + "eval_accuracy": 0.03469263542300669, + "eval_loss": 0.6868035197257996, + "eval_macro_f1": 0.4371516400689209, + "eval_micro_f1": 0.47007058392153267, + "eval_runtime": 4.3696, + "eval_samples_per_second": 376.006, + "eval_steps_per_second": 23.572, + "eval_weighted_f1": 0.49734188960489123, + "step": 1362 + }, + { + "epoch": 3.028650137741047, + "grad_norm": 28.140533447265625, + "learning_rate": 1.8859520313264808e-05, + "loss": 1.7875, + "step": 1375 + }, + { + "epoch": 3.0837465564738293, + "grad_norm": 30.750205993652344, + "learning_rate": 1.8798335780714636e-05, + "loss": 1.4453, + "step": 1400 + }, + { + "epoch": 3.1388429752066114, + "grad_norm": 21.363359451293945, + "learning_rate": 1.8737151248164465e-05, + "loss": 1.4556, + "step": 1425 + }, + { + "epoch": 3.193939393939394, + "grad_norm": 27.651782989501953, + "learning_rate": 1.8675966715614294e-05, + "loss": 1.5046, + "step": 1450 + }, + { + "epoch": 3.2490358126721763, + "grad_norm": 83.09233093261719, + "learning_rate": 1.8614782183064122e-05, + "loss": 1.455, + "step": 1475 + }, + { + "epoch": 3.304132231404959, + "grad_norm": 24.389867782592773, + "learning_rate": 1.855359765051395e-05, + "loss": 1.5406, + "step": 1500 + }, + { + "epoch": 3.359228650137741, + "grad_norm": 38.59128952026367, + "learning_rate": 1.849241311796378e-05, + "loss": 1.5871, + "step": 1525 + }, + { + "epoch": 3.4143250688705233, + "grad_norm": 28.31092643737793, + "learning_rate": 1.843122858541361e-05, + "loss": 1.4711, + "step": 1550 + }, + { + "epoch": 3.469421487603306, + "grad_norm": 44.37133026123047, + "learning_rate": 1.8370044052863437e-05, + "loss": 1.4974, + "step": 1575 + }, + { + "epoch": 3.5245179063360883, + "grad_norm": 29.924423217773438, + "learning_rate": 1.8308859520313266e-05, + "loss": 1.4409, + "step": 1600 + }, + { + "epoch": 3.579614325068871, + "grad_norm": 30.498910903930664, + "learning_rate": 1.8247674987763094e-05, + "loss": 1.5332, + "step": 1625 + }, + { + "epoch": 3.634710743801653, + "grad_norm": 31.547550201416016, + "learning_rate": 1.8186490455212923e-05, + "loss": 1.5329, + "step": 1650 + }, + { + "epoch": 3.6898071625344353, + "grad_norm": 58.15964126586914, + "learning_rate": 1.812530592266275e-05, + "loss": 1.5934, + "step": 1675 + }, + { + "epoch": 3.744903581267218, + "grad_norm": 31.391183853149414, + "learning_rate": 1.806412139011258e-05, + "loss": 1.4584, + "step": 1700 + }, + { + "epoch": 3.8, + "grad_norm": 30.878021240234375, + "learning_rate": 1.800293685756241e-05, + "loss": 1.5079, + "step": 1725 + }, + { + "epoch": 3.8550964187327823, + "grad_norm": 67.45068359375, + "learning_rate": 1.7941752325012238e-05, + "loss": 1.4805, + "step": 1750 + }, + { + "epoch": 3.910192837465565, + "grad_norm": 39.02509307861328, + "learning_rate": 1.7880567792462066e-05, + "loss": 1.4594, + "step": 1775 + }, + { + "epoch": 3.9652892561983473, + "grad_norm": 49.32225799560547, + "learning_rate": 1.7819383259911895e-05, + "loss": 1.4746, + "step": 1800 + }, + { + "epoch": 4.0, + "eval_accuracy": 0.05842970176506391, + "eval_loss": 0.7517656087875366, + "eval_macro_f1": 0.4665027010747376, + "eval_micro_f1": 0.4996953077391834, + "eval_runtime": 4.3027, + "eval_samples_per_second": 381.849, + "eval_steps_per_second": 23.938, + "eval_weighted_f1": 0.5233151116104942, + "step": 1816 + }, + { + "epoch": 4.019834710743802, + "grad_norm": 21.571191787719727, + "learning_rate": 1.7758198727361724e-05, + "loss": 1.337, + "step": 1825 + }, + { + "epoch": 4.074931129476584, + "grad_norm": 30.74432373046875, + "learning_rate": 1.7697014194811552e-05, + "loss": 1.0965, + "step": 1850 + }, + { + "epoch": 4.130027548209366, + "grad_norm": 23.688343048095703, + "learning_rate": 1.763582966226138e-05, + "loss": 1.0734, + "step": 1875 + }, + { + "epoch": 4.1851239669421485, + "grad_norm": 19.626678466796875, + "learning_rate": 1.757464512971121e-05, + "loss": 1.0143, + "step": 1900 + }, + { + "epoch": 4.240220385674931, + "grad_norm": 23.00777816772461, + "learning_rate": 1.7513460597161038e-05, + "loss": 0.9913, + "step": 1925 + }, + { + "epoch": 4.2953168044077135, + "grad_norm": 18.951629638671875, + "learning_rate": 1.7452276064610867e-05, + "loss": 1.006, + "step": 1950 + }, + { + "epoch": 4.350413223140496, + "grad_norm": 23.306678771972656, + "learning_rate": 1.7391091532060695e-05, + "loss": 1.0146, + "step": 1975 + }, + { + "epoch": 4.405509641873278, + "grad_norm": 28.98760986328125, + "learning_rate": 1.7329906999510527e-05, + "loss": 1.0294, + "step": 2000 + }, + { + "epoch": 4.460606060606061, + "grad_norm": 33.49600601196289, + "learning_rate": 1.7268722466960356e-05, + "loss": 1.0505, + "step": 2025 + }, + { + "epoch": 4.515702479338843, + "grad_norm": 38.13397979736328, + "learning_rate": 1.7207537934410185e-05, + "loss": 1.0155, + "step": 2050 + }, + { + "epoch": 4.570798898071625, + "grad_norm": 27.773197174072266, + "learning_rate": 1.7146353401860013e-05, + "loss": 1.0241, + "step": 2075 + }, + { + "epoch": 4.6258953168044075, + "grad_norm": 20.84379005432129, + "learning_rate": 1.7085168869309842e-05, + "loss": 1.0118, + "step": 2100 + }, + { + "epoch": 4.68099173553719, + "grad_norm": 17.566801071166992, + "learning_rate": 1.702398433675967e-05, + "loss": 0.9935, + "step": 2125 + }, + { + "epoch": 4.736088154269972, + "grad_norm": 103.26502227783203, + "learning_rate": 1.69627998042095e-05, + "loss": 1.0447, + "step": 2150 + }, + { + "epoch": 4.791184573002755, + "grad_norm": 25.271377563476562, + "learning_rate": 1.6901615271659325e-05, + "loss": 0.9944, + "step": 2175 + }, + { + "epoch": 4.846280991735537, + "grad_norm": 24.52176856994629, + "learning_rate": 1.6840430739109153e-05, + "loss": 0.9887, + "step": 2200 + }, + { + "epoch": 4.90137741046832, + "grad_norm": 63.398162841796875, + "learning_rate": 1.6779246206558982e-05, + "loss": 1.0061, + "step": 2225 + }, + { + "epoch": 4.9564738292011015, + "grad_norm": 18.910703659057617, + "learning_rate": 1.671806167400881e-05, + "loss": 0.9774, + "step": 2250 + }, + { + "epoch": 5.0, + "eval_accuracy": 0.11259890444309191, + "eval_loss": 0.8310469388961792, + "eval_macro_f1": 0.5111319853497628, + "eval_micro_f1": 0.5437756497948016, + "eval_runtime": 4.2823, + "eval_samples_per_second": 383.67, + "eval_steps_per_second": 24.052, + "eval_weighted_f1": 0.557824350849695, + "step": 2270 + }, + { + "epoch": 5.011019283746556, + "grad_norm": 26.505420684814453, + "learning_rate": 1.665687714145864e-05, + "loss": 0.9553, + "step": 2275 + }, + { + "epoch": 5.066115702479339, + "grad_norm": 15.603583335876465, + "learning_rate": 1.6595692608908468e-05, + "loss": 0.6928, + "step": 2300 + }, + { + "epoch": 5.121212121212121, + "grad_norm": 18.5767879486084, + "learning_rate": 1.6534508076358297e-05, + "loss": 0.7256, + "step": 2325 + }, + { + "epoch": 5.176308539944904, + "grad_norm": 26.94700813293457, + "learning_rate": 1.6473323543808125e-05, + "loss": 0.6982, + "step": 2350 + }, + { + "epoch": 5.231404958677686, + "grad_norm": 31.01991844177246, + "learning_rate": 1.6412139011257954e-05, + "loss": 0.7116, + "step": 2375 + }, + { + "epoch": 5.2865013774104685, + "grad_norm": 23.558834075927734, + "learning_rate": 1.6350954478707782e-05, + "loss": 0.7001, + "step": 2400 + }, + { + "epoch": 5.341597796143251, + "grad_norm": 12.35789966583252, + "learning_rate": 1.628976994615761e-05, + "loss": 0.6926, + "step": 2425 + }, + { + "epoch": 5.3966942148760335, + "grad_norm": 42.21820831298828, + "learning_rate": 1.622858541360744e-05, + "loss": 0.6865, + "step": 2450 + }, + { + "epoch": 5.451790633608815, + "grad_norm": 16.15004539489746, + "learning_rate": 1.616740088105727e-05, + "loss": 0.7263, + "step": 2475 + }, + { + "epoch": 5.506887052341598, + "grad_norm": 16.69156837463379, + "learning_rate": 1.6106216348507097e-05, + "loss": 0.6735, + "step": 2500 + }, + { + "epoch": 5.56198347107438, + "grad_norm": 22.371732711791992, + "learning_rate": 1.604503181595693e-05, + "loss": 0.6621, + "step": 2525 + }, + { + "epoch": 5.6170798898071626, + "grad_norm": 32.58012771606445, + "learning_rate": 1.5983847283406758e-05, + "loss": 0.7244, + "step": 2550 + }, + { + "epoch": 5.672176308539945, + "grad_norm": 15.710920333862305, + "learning_rate": 1.5922662750856586e-05, + "loss": 0.6896, + "step": 2575 + }, + { + "epoch": 5.7272727272727275, + "grad_norm": 17.526052474975586, + "learning_rate": 1.5861478218306415e-05, + "loss": 0.7016, + "step": 2600 + }, + { + "epoch": 5.78236914600551, + "grad_norm": 29.30988121032715, + "learning_rate": 1.5800293685756244e-05, + "loss": 0.6489, + "step": 2625 + }, + { + "epoch": 5.837465564738292, + "grad_norm": 15.684271812438965, + "learning_rate": 1.5739109153206072e-05, + "loss": 0.6965, + "step": 2650 + }, + { + "epoch": 5.892561983471074, + "grad_norm": 21.394363403320312, + "learning_rate": 1.56779246206559e-05, + "loss": 0.6685, + "step": 2675 + }, + { + "epoch": 5.9476584022038566, + "grad_norm": 24.16263198852539, + "learning_rate": 1.561674008810573e-05, + "loss": 0.6993, + "step": 2700 + }, + { + "epoch": 6.0, + "eval_accuracy": 0.1503347534996957, + "eval_loss": 0.9988670945167542, + "eval_macro_f1": 0.532413830333447, + "eval_micro_f1": 0.5667087011349307, + "eval_runtime": 4.3593, + "eval_samples_per_second": 376.895, + "eval_steps_per_second": 23.628, + "eval_weighted_f1": 0.5745348982790813, + "step": 2724 + }, + { + "epoch": 6.002203856749311, + "grad_norm": 12.535059928894043, + "learning_rate": 1.555555555555556e-05, + "loss": 0.684, + "step": 2725 + }, + { + "epoch": 6.057300275482094, + "grad_norm": 10.186878204345703, + "learning_rate": 1.5494371023005387e-05, + "loss": 0.5155, + "step": 2750 + }, + { + "epoch": 6.112396694214876, + "grad_norm": 16.518949508666992, + "learning_rate": 1.5433186490455216e-05, + "loss": 0.475, + "step": 2775 + }, + { + "epoch": 6.167493112947659, + "grad_norm": 16.47127914428711, + "learning_rate": 1.5372001957905044e-05, + "loss": 0.5018, + "step": 2800 + }, + { + "epoch": 6.222589531680441, + "grad_norm": 13.560384750366211, + "learning_rate": 1.5310817425354873e-05, + "loss": 0.4664, + "step": 2825 + }, + { + "epoch": 6.277685950413223, + "grad_norm": 13.8343505859375, + "learning_rate": 1.52496328928047e-05, + "loss": 0.4305, + "step": 2850 + }, + { + "epoch": 6.332782369146005, + "grad_norm": 29.02962303161621, + "learning_rate": 1.5188448360254529e-05, + "loss": 0.48, + "step": 2875 + }, + { + "epoch": 6.387878787878788, + "grad_norm": 26.647354125976562, + "learning_rate": 1.5127263827704357e-05, + "loss": 0.4708, + "step": 2900 + }, + { + "epoch": 6.44297520661157, + "grad_norm": 24.077594757080078, + "learning_rate": 1.5066079295154186e-05, + "loss": 0.4963, + "step": 2925 + }, + { + "epoch": 6.498071625344353, + "grad_norm": 33.84148406982422, + "learning_rate": 1.5004894762604014e-05, + "loss": 0.4715, + "step": 2950 + }, + { + "epoch": 6.553168044077135, + "grad_norm": 22.279882431030273, + "learning_rate": 1.4943710230053843e-05, + "loss": 0.5047, + "step": 2975 + }, + { + "epoch": 6.608264462809918, + "grad_norm": 17.463790893554688, + "learning_rate": 1.4882525697503672e-05, + "loss": 0.4452, + "step": 3000 + }, + { + "epoch": 6.663360881542699, + "grad_norm": 16.56401252746582, + "learning_rate": 1.48213411649535e-05, + "loss": 0.4577, + "step": 3025 + }, + { + "epoch": 6.718457300275482, + "grad_norm": 17.448455810546875, + "learning_rate": 1.476015663240333e-05, + "loss": 0.4442, + "step": 3050 + }, + { + "epoch": 6.773553719008264, + "grad_norm": 21.724430084228516, + "learning_rate": 1.469897209985316e-05, + "loss": 0.481, + "step": 3075 + }, + { + "epoch": 6.828650137741047, + "grad_norm": 40.37847900390625, + "learning_rate": 1.4637787567302988e-05, + "loss": 0.4398, + "step": 3100 + }, + { + "epoch": 6.883746556473829, + "grad_norm": 22.059410095214844, + "learning_rate": 1.4576603034752817e-05, + "loss": 0.4912, + "step": 3125 + }, + { + "epoch": 6.938842975206612, + "grad_norm": 15.54787540435791, + "learning_rate": 1.4515418502202645e-05, + "loss": 0.4943, + "step": 3150 + }, + { + "epoch": 6.993939393939394, + "grad_norm": 22.88062858581543, + "learning_rate": 1.4454233969652474e-05, + "loss": 0.4821, + "step": 3175 + }, + { + "epoch": 7.0, + "eval_accuracy": 0.18137553256238587, + "eval_loss": 1.1318719387054443, + "eval_macro_f1": 0.547079669309891, + "eval_micro_f1": 0.5825033647375505, + "eval_runtime": 4.34, + "eval_samples_per_second": 378.569, + "eval_steps_per_second": 23.733, + "eval_weighted_f1": 0.5844643231764589, + "step": 3178 + }, + { + "epoch": 7.048484848484849, + "grad_norm": 17.657073974609375, + "learning_rate": 1.4393049437102303e-05, + "loss": 0.3667, + "step": 3200 + }, + { + "epoch": 7.103581267217631, + "grad_norm": 11.658130645751953, + "learning_rate": 1.4331864904552131e-05, + "loss": 0.339, + "step": 3225 + }, + { + "epoch": 7.158677685950413, + "grad_norm": 13.482918739318848, + "learning_rate": 1.427068037200196e-05, + "loss": 0.3379, + "step": 3250 + }, + { + "epoch": 7.213774104683195, + "grad_norm": 7.972480773925781, + "learning_rate": 1.4209495839451789e-05, + "loss": 0.3067, + "step": 3275 + }, + { + "epoch": 7.268870523415978, + "grad_norm": 13.777979850769043, + "learning_rate": 1.4148311306901617e-05, + "loss": 0.3032, + "step": 3300 + }, + { + "epoch": 7.32396694214876, + "grad_norm": 13.239235877990723, + "learning_rate": 1.4087126774351444e-05, + "loss": 0.2966, + "step": 3325 + }, + { + "epoch": 7.379063360881543, + "grad_norm": 13.902510643005371, + "learning_rate": 1.4025942241801273e-05, + "loss": 0.3165, + "step": 3350 + }, + { + "epoch": 7.434159779614325, + "grad_norm": 8.007161140441895, + "learning_rate": 1.3964757709251102e-05, + "loss": 0.3151, + "step": 3375 + }, + { + "epoch": 7.489256198347108, + "grad_norm": 22.011545181274414, + "learning_rate": 1.390357317670093e-05, + "loss": 0.3528, + "step": 3400 + }, + { + "epoch": 7.544352617079889, + "grad_norm": 21.754802703857422, + "learning_rate": 1.3842388644150759e-05, + "loss": 0.3255, + "step": 3425 + }, + { + "epoch": 7.599449035812672, + "grad_norm": 11.811413764953613, + "learning_rate": 1.3781204111600588e-05, + "loss": 0.2947, + "step": 3450 + }, + { + "epoch": 7.654545454545454, + "grad_norm": 14.519998550415039, + "learning_rate": 1.3720019579050416e-05, + "loss": 0.3041, + "step": 3475 + }, + { + "epoch": 7.709641873278237, + "grad_norm": 14.036474227905273, + "learning_rate": 1.3658835046500245e-05, + "loss": 0.2913, + "step": 3500 + }, + { + "epoch": 7.764738292011019, + "grad_norm": 10.115962982177734, + "learning_rate": 1.3597650513950073e-05, + "loss": 0.312, + "step": 3525 + }, + { + "epoch": 7.819834710743802, + "grad_norm": 19.058610916137695, + "learning_rate": 1.3536465981399902e-05, + "loss": 0.2998, + "step": 3550 + }, + { + "epoch": 7.874931129476584, + "grad_norm": 13.24122428894043, + "learning_rate": 1.3475281448849732e-05, + "loss": 0.3188, + "step": 3575 + }, + { + "epoch": 7.930027548209367, + "grad_norm": 10.293213844299316, + "learning_rate": 1.3414096916299561e-05, + "loss": 0.2867, + "step": 3600 + }, + { + "epoch": 7.985123966942149, + "grad_norm": 14.120196342468262, + "learning_rate": 1.335291238374939e-05, + "loss": 0.3082, + "step": 3625 + }, + { + "epoch": 8.0, + "eval_accuracy": 0.20998174071819842, + "eval_loss": 1.3308709859848022, + "eval_macro_f1": 0.550203038718693, + "eval_micro_f1": 0.5870346413684059, + "eval_runtime": 4.4629, + "eval_samples_per_second": 368.143, + "eval_steps_per_second": 23.079, + "eval_weighted_f1": 0.5874242663055962, + "step": 3632 + } + ], + "logging_steps": 25, + "max_steps": 9080, + "num_input_tokens_seen": 0, + "num_train_epochs": 20, + "save_steps": 500, + "stateful_callbacks": { + "EarlyStoppingCallback": { + "args": { + "early_stopping_patience": 3, + "early_stopping_threshold": 0.0 + }, + "attributes": { + "early_stopping_patience_counter": 0 + } + }, + "TrainerControl": { + "args": { + "should_epoch_stop": false, + "should_evaluate": false, + "should_log": false, + "should_save": true, + "should_training_stop": false + }, + "attributes": {} + } + }, + "total_flos": 0.0, + "train_batch_size": 8, + "trial_name": null, + "trial_params": null +} diff --git a/ood-vigoemotions-mmbert-source-source-01-s44-20260905T144605Z-37344fee/trainer/checkpoint-4994/optimizer.pt b/ood-vigoemotions-mmbert-source-source-01-s44-20260905T144605Z-37344fee/trainer/checkpoint-4994/optimizer.pt new file mode 100644 index 0000000000000000000000000000000000000000..95b926f20be8ed8e58f3ba1a9029c15037d10203 --- /dev/null +++ b/ood-vigoemotions-mmbert-source-source-01-s44-20260905T144605Z-37344fee/trainer/checkpoint-4994/optimizer.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:20ffd0247aeb38c6f073afb9930dab73e50312dc4355ccc8b1124649ebda4de5 +size 2455803898 diff --git a/ood-vigoemotions-mmbert-source-source-01-s44-20260905T144605Z-37344fee/trainer/checkpoint-4994/special_tokens_map.json b/ood-vigoemotions-mmbert-source-source-01-s44-20260905T144605Z-37344fee/trainer/checkpoint-4994/special_tokens_map.json new file mode 100644 index 0000000000000000000000000000000000000000..be4ad795f49075ac5e56d8108ef63be314a1e967 --- /dev/null +++ b/ood-vigoemotions-mmbert-source-source-01-s44-20260905T144605Z-37344fee/trainer/checkpoint-4994/special_tokens_map.json @@ -0,0 +1,55 @@ +{ + "additional_special_tokens": [ + "", + "" + ], + "bos_token": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false + }, + "cls_token": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false + }, + "eos_token": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false + }, + "mask_token": { + "content": "", + "lstrip": true, + "normalized": false, + "rstrip": false, + "single_word": false + }, + "pad_token": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false + }, + "sep_token": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false + }, + "unk_token": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false + } +} diff --git a/ood-vigoemotions-mmbert-source-source-01-s44-20260905T144605Z-37344fee/trainer/checkpoint-4994/tokenizer_config.json b/ood-vigoemotions-mmbert-source-source-01-s44-20260905T144605Z-37344fee/trainer/checkpoint-4994/tokenizer_config.json new file mode 100644 index 0000000000000000000000000000000000000000..b002fb8c2960be8ebf9c373d6e5784b4565d7edc --- /dev/null +++ b/ood-vigoemotions-mmbert-source-source-01-s44-20260905T144605Z-37344fee/trainer/checkpoint-4994/tokenizer_config.json @@ -0,0 +1,2018 @@ +{ + "add_bos_token": true, + "added_tokens_decoder": { + "0": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "1": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "2": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "3": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "4": { + "content": "", + "lstrip": true, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "5": { + "content": "<2mass>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "6": { + "content": "[@BOS@]", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "7": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "8": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "9": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "10": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "11": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "12": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "13": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "14": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "15": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "16": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "17": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "18": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "19": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "20": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "21": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "22": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "23": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "24": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "25": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "26": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "27": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "28": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "29": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "30": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "31": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "32": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "33": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "34": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "35": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "36": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "37": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "38": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "39": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "40": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "41": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "42": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "43": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "44": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "45": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "46": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "47": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "48": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "49": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "50": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "51": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "52": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "53": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "54": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "55": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "56": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "57": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "58": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "59": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "60": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "61": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "62": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "63": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "64": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "65": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "66": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "67": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "68": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "69": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "70": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "71": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "72": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "73": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "74": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "75": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "76": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "77": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "78": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "79": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "80": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "81": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "82": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "83": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "84": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "85": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "86": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "87": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "88": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "89": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "90": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "91": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "92": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "93": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "94": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "95": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "96": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "97": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "98": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "99": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "100": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "101": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "102": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "103": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "104": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "105": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "106": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "107": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "108": { + "content": "\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "109": { + "content": "\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "110": { + "content": "\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "111": { + "content": "\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "112": { + "content": "\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "113": { + "content": "\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "114": { + "content": "\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "115": { + "content": "\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "116": { + "content": "\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "117": { + "content": "\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "118": { + "content": "\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "119": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "120": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "121": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "122": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "123": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "124": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "125": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "126": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "127": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "128": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "129": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "130": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "131": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "132": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "133": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "134": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "135": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "136": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "137": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "138": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "139": { + "content": "▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "140": { + "content": "▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "141": { + "content": "▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "142": { + "content": "▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "143": { + "content": "▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "144": { + "content": "▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "145": { + "content": "▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "146": { + "content": "▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "147": { + "content": "▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "148": { + "content": "▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "149": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "150": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "151": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "152": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "153": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "154": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "155": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "156": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "157": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "158": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "159": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "160": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "161": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "162": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "163": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "164": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "165": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "166": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "167": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "168": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "169": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "170": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "172": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "173": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "174": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "175": { + "content": "
", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "171": { + "content": "
", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "176": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "177": { + "content": "
", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "178": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "179": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "180": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "181": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "182": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "183": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "184": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "185": { + "content": "

", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "186": { + "content": "

", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "187": { + "content": "

", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "188": { + "content": "

", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "189": { + "content": "

", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "190": { + "content": "
", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "191": { + "content": "
", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "192": { + "content": "
", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "193": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "194": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "195": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "196": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "197": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "198": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "199": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "200": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "201": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "202": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "203": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "204": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "205": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "206": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "207": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "208": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "209": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "210": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "211": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "212": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "213": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "214": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "215": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "216": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255968": { + "content": "[toxicity=0]", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255969": { + "content": "\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255970": { + "content": "\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255971": { + "content": "\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255972": { + "content": "\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255973": { + "content": "\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255974": { + "content": "\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255975": { + "content": "\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255976": { + "content": "\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255977": { + "content": "\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255978": { + "content": "\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255979": { + "content": "\t\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255980": { + "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255981": { + "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255982": { + "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255983": { + "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255984": { + "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255985": { + "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255986": { + "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255987": { + "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255988": { + "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255989": { + "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255990": { + "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255991": { + "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255992": { + "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255993": { + "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255994": { + "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255995": { + "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255996": { + "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255997": { + "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255998": { + "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255999": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + } + }, + "additional_special_tokens": [ + "", + "" + ], + "bos_token": "", + "clean_up_tokenization_spaces": false, + "cls_token": "", + "eos_token": "", + "extra_special_tokens": {}, + "mask_token": "", + "model_input_names": [ + "input_ids", + "attention_mask" + ], + "model_max_length": 8192, + "pad_token": "", + "padding_side": "right", + "sep_token": "", + "spaces_between_special_tokens": false, + "tokenizer_class": "PreTrainedTokenizerFast", + "unk_token": "" +} diff --git a/ood-vigoemotions-mmbert-source-source-01-s44-20260905T144605Z-37344fee/trainer/checkpoint-4994/trainer_state.json b/ood-vigoemotions-mmbert-source-source-01-s44-20260905T144605Z-37344fee/trainer/checkpoint-4994/trainer_state.json new file mode 100644 index 0000000000000000000000000000000000000000..fbf2ce74fd30b0928276afe64264ff1d353f8002 --- /dev/null +++ b/ood-vigoemotions-mmbert-source-source-01-s44-20260905T144605Z-37344fee/trainer/checkpoint-4994/trainer_state.json @@ -0,0 +1,1568 @@ +{ + "best_global_step": 3632, + "best_metric": 0.550203038718693, + "best_model_checkpoint": "/root/VNEncoder/results/checkpoints/ood-vigoemotions-mmbert-source-source-01-s44-20260905T144605Z-37344fee/trainer/checkpoint-3632", + "epoch": 11.0, + "eval_steps": 500, + "global_step": 4994, + "is_hyper_param_search": false, + "is_local_process_zero": true, + "is_world_process_zero": true, + "log_history": [ + { + "epoch": 0.05509641873278237, + "grad_norm": 42.07517623901367, + "learning_rate": 5.286343612334802e-07, + "loss": 6.2699, + "step": 25 + }, + { + "epoch": 0.11019283746556474, + "grad_norm": 34.53789520263672, + "learning_rate": 1.0792951541850223e-06, + "loss": 6.0286, + "step": 50 + }, + { + "epoch": 0.1652892561983471, + "grad_norm": 56.00322341918945, + "learning_rate": 1.629955947136564e-06, + "loss": 6.3079, + "step": 75 + }, + { + "epoch": 0.22038567493112948, + "grad_norm": 34.499778747558594, + "learning_rate": 2.180616740088106e-06, + "loss": 6.3055, + "step": 100 + }, + { + "epoch": 0.27548209366391185, + "grad_norm": 49.277896881103516, + "learning_rate": 2.731277533039648e-06, + "loss": 6.0817, + "step": 125 + }, + { + "epoch": 0.3305785123966942, + "grad_norm": 293.85943603515625, + "learning_rate": 3.2819383259911898e-06, + "loss": 5.8855, + "step": 150 + }, + { + "epoch": 0.3856749311294766, + "grad_norm": 170.3839874267578, + "learning_rate": 3.8325991189427315e-06, + "loss": 5.4578, + "step": 175 + }, + { + "epoch": 0.44077134986225897, + "grad_norm": 54.24516296386719, + "learning_rate": 4.383259911894274e-06, + "loss": 5.1302, + "step": 200 + }, + { + "epoch": 0.49586776859504134, + "grad_norm": 76.05721282958984, + "learning_rate": 4.933920704845816e-06, + "loss": 5.1066, + "step": 225 + }, + { + "epoch": 0.5509641873278237, + "grad_norm": 61.08894348144531, + "learning_rate": 5.484581497797358e-06, + "loss": 4.9914, + "step": 250 + }, + { + "epoch": 0.6060606060606061, + "grad_norm": 47.69670486450195, + "learning_rate": 6.035242290748899e-06, + "loss": 4.8303, + "step": 275 + }, + { + "epoch": 0.6611570247933884, + "grad_norm": 149.77024841308594, + "learning_rate": 6.585903083700441e-06, + "loss": 4.502, + "step": 300 + }, + { + "epoch": 0.7162534435261708, + "grad_norm": 50.70989990234375, + "learning_rate": 7.136563876651983e-06, + "loss": 4.383, + "step": 325 + }, + { + "epoch": 0.7713498622589532, + "grad_norm": 43.84785079956055, + "learning_rate": 7.687224669603525e-06, + "loss": 4.3155, + "step": 350 + }, + { + "epoch": 0.8264462809917356, + "grad_norm": 50.88126754760742, + "learning_rate": 8.237885462555067e-06, + "loss": 4.0407, + "step": 375 + }, + { + "epoch": 0.8815426997245179, + "grad_norm": 47.921146392822266, + "learning_rate": 8.788546255506607e-06, + "loss": 3.6664, + "step": 400 + }, + { + "epoch": 0.9366391184573003, + "grad_norm": 36.562015533447266, + "learning_rate": 9.339207048458151e-06, + "loss": 3.8568, + "step": 425 + }, + { + "epoch": 0.9917355371900827, + "grad_norm": 96.27021026611328, + "learning_rate": 9.889867841409693e-06, + "loss": 3.7885, + "step": 450 + }, + { + "epoch": 1.0, + "eval_accuracy": 0.0, + "eval_loss": 0.8759939670562744, + "eval_macro_f1": 0.30856328770558283, + "eval_micro_f1": 0.31929322466247745, + "eval_runtime": 5.2481, + "eval_samples_per_second": 313.064, + "eval_steps_per_second": 19.626, + "eval_weighted_f1": 0.38774477435725324, + "step": 454 + }, + { + "epoch": 1.0462809917355371, + "grad_norm": 45.211910247802734, + "learning_rate": 1.0440528634361234e-05, + "loss": 3.2569, + "step": 475 + }, + { + "epoch": 1.1013774104683196, + "grad_norm": 39.812557220458984, + "learning_rate": 1.0991189427312776e-05, + "loss": 3.2178, + "step": 500 + }, + { + "epoch": 1.1564738292011019, + "grad_norm": 93.90754699707031, + "learning_rate": 1.154185022026432e-05, + "loss": 3.3733, + "step": 525 + }, + { + "epoch": 1.2115702479338843, + "grad_norm": 50.11622619628906, + "learning_rate": 1.209251101321586e-05, + "loss": 3.2525, + "step": 550 + }, + { + "epoch": 1.2666666666666666, + "grad_norm": 54.41451644897461, + "learning_rate": 1.2643171806167402e-05, + "loss": 3.0884, + "step": 575 + }, + { + "epoch": 1.321763085399449, + "grad_norm": 48.95606231689453, + "learning_rate": 1.3193832599118943e-05, + "loss": 3.1357, + "step": 600 + }, + { + "epoch": 1.3768595041322313, + "grad_norm": 62.82511520385742, + "learning_rate": 1.3744493392070487e-05, + "loss": 3.3523, + "step": 625 + }, + { + "epoch": 1.4319559228650138, + "grad_norm": 49.2446403503418, + "learning_rate": 1.4295154185022027e-05, + "loss": 2.9562, + "step": 650 + }, + { + "epoch": 1.487052341597796, + "grad_norm": 34.752410888671875, + "learning_rate": 1.484581497797357e-05, + "loss": 3.0026, + "step": 675 + }, + { + "epoch": 1.5421487603305786, + "grad_norm": 50.21882247924805, + "learning_rate": 1.539647577092511e-05, + "loss": 3.2388, + "step": 700 + }, + { + "epoch": 1.5972451790633608, + "grad_norm": 72.29084014892578, + "learning_rate": 1.5947136563876654e-05, + "loss": 3.0325, + "step": 725 + }, + { + "epoch": 1.652341597796143, + "grad_norm": 32.43894577026367, + "learning_rate": 1.6497797356828196e-05, + "loss": 3.0251, + "step": 750 + }, + { + "epoch": 1.7074380165289256, + "grad_norm": 40.4652099609375, + "learning_rate": 1.7048458149779738e-05, + "loss": 2.9036, + "step": 775 + }, + { + "epoch": 1.762534435261708, + "grad_norm": 52.69956588745117, + "learning_rate": 1.759911894273128e-05, + "loss": 2.9234, + "step": 800 + }, + { + "epoch": 1.8176308539944903, + "grad_norm": 30.133697509765625, + "learning_rate": 1.814977973568282e-05, + "loss": 2.9855, + "step": 825 + }, + { + "epoch": 1.8727272727272726, + "grad_norm": 61.72572708129883, + "learning_rate": 1.8700440528634364e-05, + "loss": 3.0321, + "step": 850 + }, + { + "epoch": 1.927823691460055, + "grad_norm": 28.934057235717773, + "learning_rate": 1.9251101321585906e-05, + "loss": 2.985, + "step": 875 + }, + { + "epoch": 1.9829201101928375, + "grad_norm": 39.00049591064453, + "learning_rate": 1.9801762114537445e-05, + "loss": 2.906, + "step": 900 + }, + { + "epoch": 2.0, + "eval_accuracy": 0.013390139987827145, + "eval_loss": 0.7132589221000671, + "eval_macro_f1": 0.3972028926547339, + "eval_micro_f1": 0.4105655588165305, + "eval_runtime": 4.347, + "eval_samples_per_second": 377.962, + "eval_steps_per_second": 23.695, + "eval_weighted_f1": 0.4630428399620474, + "step": 908 + }, + { + "epoch": 2.037465564738292, + "grad_norm": 24.804983139038086, + "learning_rate": 1.9960841899167892e-05, + "loss": 2.4633, + "step": 925 + }, + { + "epoch": 2.0925619834710742, + "grad_norm": 64.12801361083984, + "learning_rate": 1.989965736661772e-05, + "loss": 2.306, + "step": 950 + }, + { + "epoch": 2.1476584022038567, + "grad_norm": 27.915681838989258, + "learning_rate": 1.983847283406755e-05, + "loss": 2.3691, + "step": 975 + }, + { + "epoch": 2.202754820936639, + "grad_norm": 24.95664405822754, + "learning_rate": 1.9777288301517378e-05, + "loss": 2.293, + "step": 1000 + }, + { + "epoch": 2.2578512396694217, + "grad_norm": 25.19192886352539, + "learning_rate": 1.9716103768967207e-05, + "loss": 2.234, + "step": 1025 + }, + { + "epoch": 2.3129476584022037, + "grad_norm": 28.52288246154785, + "learning_rate": 1.9654919236417035e-05, + "loss": 2.1174, + "step": 1050 + }, + { + "epoch": 2.368044077134986, + "grad_norm": 48.03308868408203, + "learning_rate": 1.9593734703866864e-05, + "loss": 2.2298, + "step": 1075 + }, + { + "epoch": 2.4231404958677687, + "grad_norm": 43.81739807128906, + "learning_rate": 1.9532550171316693e-05, + "loss": 2.2613, + "step": 1100 + }, + { + "epoch": 2.4782369146005507, + "grad_norm": 29.52232551574707, + "learning_rate": 1.947136563876652e-05, + "loss": 2.257, + "step": 1125 + }, + { + "epoch": 2.533333333333333, + "grad_norm": 30.75737190246582, + "learning_rate": 1.941018110621635e-05, + "loss": 2.2481, + "step": 1150 + }, + { + "epoch": 2.5884297520661157, + "grad_norm": 35.87025451660156, + "learning_rate": 1.934899657366618e-05, + "loss": 2.2418, + "step": 1175 + }, + { + "epoch": 2.643526170798898, + "grad_norm": 45.66483688354492, + "learning_rate": 1.9287812041116007e-05, + "loss": 2.2211, + "step": 1200 + }, + { + "epoch": 2.6986225895316807, + "grad_norm": 46.8340950012207, + "learning_rate": 1.9226627508565836e-05, + "loss": 2.2759, + "step": 1225 + }, + { + "epoch": 2.7537190082644627, + "grad_norm": 33.303741455078125, + "learning_rate": 1.9165442976015665e-05, + "loss": 2.1401, + "step": 1250 + }, + { + "epoch": 2.808815426997245, + "grad_norm": 44.45813751220703, + "learning_rate": 1.9104258443465493e-05, + "loss": 2.2384, + "step": 1275 + }, + { + "epoch": 2.8639118457300277, + "grad_norm": 28.29966163635254, + "learning_rate": 1.9043073910915322e-05, + "loss": 2.161, + "step": 1300 + }, + { + "epoch": 2.9190082644628097, + "grad_norm": 35.3993034362793, + "learning_rate": 1.898188937836515e-05, + "loss": 2.1901, + "step": 1325 + }, + { + "epoch": 2.974104683195592, + "grad_norm": 47.99441146850586, + "learning_rate": 1.892070484581498e-05, + "loss": 2.0656, + "step": 1350 + }, + { + "epoch": 3.0, + "eval_accuracy": 0.03469263542300669, + "eval_loss": 0.6868035197257996, + "eval_macro_f1": 0.4371516400689209, + "eval_micro_f1": 0.47007058392153267, + "eval_runtime": 4.3696, + "eval_samples_per_second": 376.006, + "eval_steps_per_second": 23.572, + "eval_weighted_f1": 0.49734188960489123, + "step": 1362 + }, + { + "epoch": 3.028650137741047, + "grad_norm": 28.140533447265625, + "learning_rate": 1.8859520313264808e-05, + "loss": 1.7875, + "step": 1375 + }, + { + "epoch": 3.0837465564738293, + "grad_norm": 30.750205993652344, + "learning_rate": 1.8798335780714636e-05, + "loss": 1.4453, + "step": 1400 + }, + { + "epoch": 3.1388429752066114, + "grad_norm": 21.363359451293945, + "learning_rate": 1.8737151248164465e-05, + "loss": 1.4556, + "step": 1425 + }, + { + "epoch": 3.193939393939394, + "grad_norm": 27.651782989501953, + "learning_rate": 1.8675966715614294e-05, + "loss": 1.5046, + "step": 1450 + }, + { + "epoch": 3.2490358126721763, + "grad_norm": 83.09233093261719, + "learning_rate": 1.8614782183064122e-05, + "loss": 1.455, + "step": 1475 + }, + { + "epoch": 3.304132231404959, + "grad_norm": 24.389867782592773, + "learning_rate": 1.855359765051395e-05, + "loss": 1.5406, + "step": 1500 + }, + { + "epoch": 3.359228650137741, + "grad_norm": 38.59128952026367, + "learning_rate": 1.849241311796378e-05, + "loss": 1.5871, + "step": 1525 + }, + { + "epoch": 3.4143250688705233, + "grad_norm": 28.31092643737793, + "learning_rate": 1.843122858541361e-05, + "loss": 1.4711, + "step": 1550 + }, + { + "epoch": 3.469421487603306, + "grad_norm": 44.37133026123047, + "learning_rate": 1.8370044052863437e-05, + "loss": 1.4974, + "step": 1575 + }, + { + "epoch": 3.5245179063360883, + "grad_norm": 29.924423217773438, + "learning_rate": 1.8308859520313266e-05, + "loss": 1.4409, + "step": 1600 + }, + { + "epoch": 3.579614325068871, + "grad_norm": 30.498910903930664, + "learning_rate": 1.8247674987763094e-05, + "loss": 1.5332, + "step": 1625 + }, + { + "epoch": 3.634710743801653, + "grad_norm": 31.547550201416016, + "learning_rate": 1.8186490455212923e-05, + "loss": 1.5329, + "step": 1650 + }, + { + "epoch": 3.6898071625344353, + "grad_norm": 58.15964126586914, + "learning_rate": 1.812530592266275e-05, + "loss": 1.5934, + "step": 1675 + }, + { + "epoch": 3.744903581267218, + "grad_norm": 31.391183853149414, + "learning_rate": 1.806412139011258e-05, + "loss": 1.4584, + "step": 1700 + }, + { + "epoch": 3.8, + "grad_norm": 30.878021240234375, + "learning_rate": 1.800293685756241e-05, + "loss": 1.5079, + "step": 1725 + }, + { + "epoch": 3.8550964187327823, + "grad_norm": 67.45068359375, + "learning_rate": 1.7941752325012238e-05, + "loss": 1.4805, + "step": 1750 + }, + { + "epoch": 3.910192837465565, + "grad_norm": 39.02509307861328, + "learning_rate": 1.7880567792462066e-05, + "loss": 1.4594, + "step": 1775 + }, + { + "epoch": 3.9652892561983473, + "grad_norm": 49.32225799560547, + "learning_rate": 1.7819383259911895e-05, + "loss": 1.4746, + "step": 1800 + }, + { + "epoch": 4.0, + "eval_accuracy": 0.05842970176506391, + "eval_loss": 0.7517656087875366, + "eval_macro_f1": 0.4665027010747376, + "eval_micro_f1": 0.4996953077391834, + "eval_runtime": 4.3027, + "eval_samples_per_second": 381.849, + "eval_steps_per_second": 23.938, + "eval_weighted_f1": 0.5233151116104942, + "step": 1816 + }, + { + "epoch": 4.019834710743802, + "grad_norm": 21.571191787719727, + "learning_rate": 1.7758198727361724e-05, + "loss": 1.337, + "step": 1825 + }, + { + "epoch": 4.074931129476584, + "grad_norm": 30.74432373046875, + "learning_rate": 1.7697014194811552e-05, + "loss": 1.0965, + "step": 1850 + }, + { + "epoch": 4.130027548209366, + "grad_norm": 23.688343048095703, + "learning_rate": 1.763582966226138e-05, + "loss": 1.0734, + "step": 1875 + }, + { + "epoch": 4.1851239669421485, + "grad_norm": 19.626678466796875, + "learning_rate": 1.757464512971121e-05, + "loss": 1.0143, + "step": 1900 + }, + { + "epoch": 4.240220385674931, + "grad_norm": 23.00777816772461, + "learning_rate": 1.7513460597161038e-05, + "loss": 0.9913, + "step": 1925 + }, + { + "epoch": 4.2953168044077135, + "grad_norm": 18.951629638671875, + "learning_rate": 1.7452276064610867e-05, + "loss": 1.006, + "step": 1950 + }, + { + "epoch": 4.350413223140496, + "grad_norm": 23.306678771972656, + "learning_rate": 1.7391091532060695e-05, + "loss": 1.0146, + "step": 1975 + }, + { + "epoch": 4.405509641873278, + "grad_norm": 28.98760986328125, + "learning_rate": 1.7329906999510527e-05, + "loss": 1.0294, + "step": 2000 + }, + { + "epoch": 4.460606060606061, + "grad_norm": 33.49600601196289, + "learning_rate": 1.7268722466960356e-05, + "loss": 1.0505, + "step": 2025 + }, + { + "epoch": 4.515702479338843, + "grad_norm": 38.13397979736328, + "learning_rate": 1.7207537934410185e-05, + "loss": 1.0155, + "step": 2050 + }, + { + "epoch": 4.570798898071625, + "grad_norm": 27.773197174072266, + "learning_rate": 1.7146353401860013e-05, + "loss": 1.0241, + "step": 2075 + }, + { + "epoch": 4.6258953168044075, + "grad_norm": 20.84379005432129, + "learning_rate": 1.7085168869309842e-05, + "loss": 1.0118, + "step": 2100 + }, + { + "epoch": 4.68099173553719, + "grad_norm": 17.566801071166992, + "learning_rate": 1.702398433675967e-05, + "loss": 0.9935, + "step": 2125 + }, + { + "epoch": 4.736088154269972, + "grad_norm": 103.26502227783203, + "learning_rate": 1.69627998042095e-05, + "loss": 1.0447, + "step": 2150 + }, + { + "epoch": 4.791184573002755, + "grad_norm": 25.271377563476562, + "learning_rate": 1.6901615271659325e-05, + "loss": 0.9944, + "step": 2175 + }, + { + "epoch": 4.846280991735537, + "grad_norm": 24.52176856994629, + "learning_rate": 1.6840430739109153e-05, + "loss": 0.9887, + "step": 2200 + }, + { + "epoch": 4.90137741046832, + "grad_norm": 63.398162841796875, + "learning_rate": 1.6779246206558982e-05, + "loss": 1.0061, + "step": 2225 + }, + { + "epoch": 4.9564738292011015, + "grad_norm": 18.910703659057617, + "learning_rate": 1.671806167400881e-05, + "loss": 0.9774, + "step": 2250 + }, + { + "epoch": 5.0, + "eval_accuracy": 0.11259890444309191, + "eval_loss": 0.8310469388961792, + "eval_macro_f1": 0.5111319853497628, + "eval_micro_f1": 0.5437756497948016, + "eval_runtime": 4.2823, + "eval_samples_per_second": 383.67, + "eval_steps_per_second": 24.052, + "eval_weighted_f1": 0.557824350849695, + "step": 2270 + }, + { + "epoch": 5.011019283746556, + "grad_norm": 26.505420684814453, + "learning_rate": 1.665687714145864e-05, + "loss": 0.9553, + "step": 2275 + }, + { + "epoch": 5.066115702479339, + "grad_norm": 15.603583335876465, + "learning_rate": 1.6595692608908468e-05, + "loss": 0.6928, + "step": 2300 + }, + { + "epoch": 5.121212121212121, + "grad_norm": 18.5767879486084, + "learning_rate": 1.6534508076358297e-05, + "loss": 0.7256, + "step": 2325 + }, + { + "epoch": 5.176308539944904, + "grad_norm": 26.94700813293457, + "learning_rate": 1.6473323543808125e-05, + "loss": 0.6982, + "step": 2350 + }, + { + "epoch": 5.231404958677686, + "grad_norm": 31.01991844177246, + "learning_rate": 1.6412139011257954e-05, + "loss": 0.7116, + "step": 2375 + }, + { + "epoch": 5.2865013774104685, + "grad_norm": 23.558834075927734, + "learning_rate": 1.6350954478707782e-05, + "loss": 0.7001, + "step": 2400 + }, + { + "epoch": 5.341597796143251, + "grad_norm": 12.35789966583252, + "learning_rate": 1.628976994615761e-05, + "loss": 0.6926, + "step": 2425 + }, + { + "epoch": 5.3966942148760335, + "grad_norm": 42.21820831298828, + "learning_rate": 1.622858541360744e-05, + "loss": 0.6865, + "step": 2450 + }, + { + "epoch": 5.451790633608815, + "grad_norm": 16.15004539489746, + "learning_rate": 1.616740088105727e-05, + "loss": 0.7263, + "step": 2475 + }, + { + "epoch": 5.506887052341598, + "grad_norm": 16.69156837463379, + "learning_rate": 1.6106216348507097e-05, + "loss": 0.6735, + "step": 2500 + }, + { + "epoch": 5.56198347107438, + "grad_norm": 22.371732711791992, + "learning_rate": 1.604503181595693e-05, + "loss": 0.6621, + "step": 2525 + }, + { + "epoch": 5.6170798898071626, + "grad_norm": 32.58012771606445, + "learning_rate": 1.5983847283406758e-05, + "loss": 0.7244, + "step": 2550 + }, + { + "epoch": 5.672176308539945, + "grad_norm": 15.710920333862305, + "learning_rate": 1.5922662750856586e-05, + "loss": 0.6896, + "step": 2575 + }, + { + "epoch": 5.7272727272727275, + "grad_norm": 17.526052474975586, + "learning_rate": 1.5861478218306415e-05, + "loss": 0.7016, + "step": 2600 + }, + { + "epoch": 5.78236914600551, + "grad_norm": 29.30988121032715, + "learning_rate": 1.5800293685756244e-05, + "loss": 0.6489, + "step": 2625 + }, + { + "epoch": 5.837465564738292, + "grad_norm": 15.684271812438965, + "learning_rate": 1.5739109153206072e-05, + "loss": 0.6965, + "step": 2650 + }, + { + "epoch": 5.892561983471074, + "grad_norm": 21.394363403320312, + "learning_rate": 1.56779246206559e-05, + "loss": 0.6685, + "step": 2675 + }, + { + "epoch": 5.9476584022038566, + "grad_norm": 24.16263198852539, + "learning_rate": 1.561674008810573e-05, + "loss": 0.6993, + "step": 2700 + }, + { + "epoch": 6.0, + "eval_accuracy": 0.1503347534996957, + "eval_loss": 0.9988670945167542, + "eval_macro_f1": 0.532413830333447, + "eval_micro_f1": 0.5667087011349307, + "eval_runtime": 4.3593, + "eval_samples_per_second": 376.895, + "eval_steps_per_second": 23.628, + "eval_weighted_f1": 0.5745348982790813, + "step": 2724 + }, + { + "epoch": 6.002203856749311, + "grad_norm": 12.535059928894043, + "learning_rate": 1.555555555555556e-05, + "loss": 0.684, + "step": 2725 + }, + { + "epoch": 6.057300275482094, + "grad_norm": 10.186878204345703, + "learning_rate": 1.5494371023005387e-05, + "loss": 0.5155, + "step": 2750 + }, + { + "epoch": 6.112396694214876, + "grad_norm": 16.518949508666992, + "learning_rate": 1.5433186490455216e-05, + "loss": 0.475, + "step": 2775 + }, + { + "epoch": 6.167493112947659, + "grad_norm": 16.47127914428711, + "learning_rate": 1.5372001957905044e-05, + "loss": 0.5018, + "step": 2800 + }, + { + "epoch": 6.222589531680441, + "grad_norm": 13.560384750366211, + "learning_rate": 1.5310817425354873e-05, + "loss": 0.4664, + "step": 2825 + }, + { + "epoch": 6.277685950413223, + "grad_norm": 13.8343505859375, + "learning_rate": 1.52496328928047e-05, + "loss": 0.4305, + "step": 2850 + }, + { + "epoch": 6.332782369146005, + "grad_norm": 29.02962303161621, + "learning_rate": 1.5188448360254529e-05, + "loss": 0.48, + "step": 2875 + }, + { + "epoch": 6.387878787878788, + "grad_norm": 26.647354125976562, + "learning_rate": 1.5127263827704357e-05, + "loss": 0.4708, + "step": 2900 + }, + { + "epoch": 6.44297520661157, + "grad_norm": 24.077594757080078, + "learning_rate": 1.5066079295154186e-05, + "loss": 0.4963, + "step": 2925 + }, + { + "epoch": 6.498071625344353, + "grad_norm": 33.84148406982422, + "learning_rate": 1.5004894762604014e-05, + "loss": 0.4715, + "step": 2950 + }, + { + "epoch": 6.553168044077135, + "grad_norm": 22.279882431030273, + "learning_rate": 1.4943710230053843e-05, + "loss": 0.5047, + "step": 2975 + }, + { + "epoch": 6.608264462809918, + "grad_norm": 17.463790893554688, + "learning_rate": 1.4882525697503672e-05, + "loss": 0.4452, + "step": 3000 + }, + { + "epoch": 6.663360881542699, + "grad_norm": 16.56401252746582, + "learning_rate": 1.48213411649535e-05, + "loss": 0.4577, + "step": 3025 + }, + { + "epoch": 6.718457300275482, + "grad_norm": 17.448455810546875, + "learning_rate": 1.476015663240333e-05, + "loss": 0.4442, + "step": 3050 + }, + { + "epoch": 6.773553719008264, + "grad_norm": 21.724430084228516, + "learning_rate": 1.469897209985316e-05, + "loss": 0.481, + "step": 3075 + }, + { + "epoch": 6.828650137741047, + "grad_norm": 40.37847900390625, + "learning_rate": 1.4637787567302988e-05, + "loss": 0.4398, + "step": 3100 + }, + { + "epoch": 6.883746556473829, + "grad_norm": 22.059410095214844, + "learning_rate": 1.4576603034752817e-05, + "loss": 0.4912, + "step": 3125 + }, + { + "epoch": 6.938842975206612, + "grad_norm": 15.54787540435791, + "learning_rate": 1.4515418502202645e-05, + "loss": 0.4943, + "step": 3150 + }, + { + "epoch": 6.993939393939394, + "grad_norm": 22.88062858581543, + "learning_rate": 1.4454233969652474e-05, + "loss": 0.4821, + "step": 3175 + }, + { + "epoch": 7.0, + "eval_accuracy": 0.18137553256238587, + "eval_loss": 1.1318719387054443, + "eval_macro_f1": 0.547079669309891, + "eval_micro_f1": 0.5825033647375505, + "eval_runtime": 4.34, + "eval_samples_per_second": 378.569, + "eval_steps_per_second": 23.733, + "eval_weighted_f1": 0.5844643231764589, + "step": 3178 + }, + { + "epoch": 7.048484848484849, + "grad_norm": 17.657073974609375, + "learning_rate": 1.4393049437102303e-05, + "loss": 0.3667, + "step": 3200 + }, + { + "epoch": 7.103581267217631, + "grad_norm": 11.658130645751953, + "learning_rate": 1.4331864904552131e-05, + "loss": 0.339, + "step": 3225 + }, + { + "epoch": 7.158677685950413, + "grad_norm": 13.482918739318848, + "learning_rate": 1.427068037200196e-05, + "loss": 0.3379, + "step": 3250 + }, + { + "epoch": 7.213774104683195, + "grad_norm": 7.972480773925781, + "learning_rate": 1.4209495839451789e-05, + "loss": 0.3067, + "step": 3275 + }, + { + "epoch": 7.268870523415978, + "grad_norm": 13.777979850769043, + "learning_rate": 1.4148311306901617e-05, + "loss": 0.3032, + "step": 3300 + }, + { + "epoch": 7.32396694214876, + "grad_norm": 13.239235877990723, + "learning_rate": 1.4087126774351444e-05, + "loss": 0.2966, + "step": 3325 + }, + { + "epoch": 7.379063360881543, + "grad_norm": 13.902510643005371, + "learning_rate": 1.4025942241801273e-05, + "loss": 0.3165, + "step": 3350 + }, + { + "epoch": 7.434159779614325, + "grad_norm": 8.007161140441895, + "learning_rate": 1.3964757709251102e-05, + "loss": 0.3151, + "step": 3375 + }, + { + "epoch": 7.489256198347108, + "grad_norm": 22.011545181274414, + "learning_rate": 1.390357317670093e-05, + "loss": 0.3528, + "step": 3400 + }, + { + "epoch": 7.544352617079889, + "grad_norm": 21.754802703857422, + "learning_rate": 1.3842388644150759e-05, + "loss": 0.3255, + "step": 3425 + }, + { + "epoch": 7.599449035812672, + "grad_norm": 11.811413764953613, + "learning_rate": 1.3781204111600588e-05, + "loss": 0.2947, + "step": 3450 + }, + { + "epoch": 7.654545454545454, + "grad_norm": 14.519998550415039, + "learning_rate": 1.3720019579050416e-05, + "loss": 0.3041, + "step": 3475 + }, + { + "epoch": 7.709641873278237, + "grad_norm": 14.036474227905273, + "learning_rate": 1.3658835046500245e-05, + "loss": 0.2913, + "step": 3500 + }, + { + "epoch": 7.764738292011019, + "grad_norm": 10.115962982177734, + "learning_rate": 1.3597650513950073e-05, + "loss": 0.312, + "step": 3525 + }, + { + "epoch": 7.819834710743802, + "grad_norm": 19.058610916137695, + "learning_rate": 1.3536465981399902e-05, + "loss": 0.2998, + "step": 3550 + }, + { + "epoch": 7.874931129476584, + "grad_norm": 13.24122428894043, + "learning_rate": 1.3475281448849732e-05, + "loss": 0.3188, + "step": 3575 + }, + { + "epoch": 7.930027548209367, + "grad_norm": 10.293213844299316, + "learning_rate": 1.3414096916299561e-05, + "loss": 0.2867, + "step": 3600 + }, + { + "epoch": 7.985123966942149, + "grad_norm": 14.120196342468262, + "learning_rate": 1.335291238374939e-05, + "loss": 0.3082, + "step": 3625 + }, + { + "epoch": 8.0, + "eval_accuracy": 0.20998174071819842, + "eval_loss": 1.3308709859848022, + "eval_macro_f1": 0.550203038718693, + "eval_micro_f1": 0.5870346413684059, + "eval_runtime": 4.4629, + "eval_samples_per_second": 368.143, + "eval_steps_per_second": 23.079, + "eval_weighted_f1": 0.5874242663055962, + "step": 3632 + }, + { + "epoch": 8.039669421487604, + "grad_norm": 28.63372039794922, + "learning_rate": 1.3291727851199218e-05, + "loss": 0.2646, + "step": 3650 + }, + { + "epoch": 8.094765840220386, + "grad_norm": 19.73433494567871, + "learning_rate": 1.3230543318649047e-05, + "loss": 0.2289, + "step": 3675 + }, + { + "epoch": 8.149862258953169, + "grad_norm": 7.862354278564453, + "learning_rate": 1.3169358786098876e-05, + "loss": 0.205, + "step": 3700 + }, + { + "epoch": 8.20495867768595, + "grad_norm": 6.76133918762207, + "learning_rate": 1.3108174253548704e-05, + "loss": 0.1921, + "step": 3725 + }, + { + "epoch": 8.260055096418732, + "grad_norm": 18.70047378540039, + "learning_rate": 1.3046989720998533e-05, + "loss": 0.2173, + "step": 3750 + }, + { + "epoch": 8.315151515151515, + "grad_norm": 9.107065200805664, + "learning_rate": 1.2985805188448362e-05, + "loss": 0.2052, + "step": 3775 + }, + { + "epoch": 8.370247933884297, + "grad_norm": 21.070680618286133, + "learning_rate": 1.292462065589819e-05, + "loss": 0.2172, + "step": 3800 + }, + { + "epoch": 8.42534435261708, + "grad_norm": 8.689684867858887, + "learning_rate": 1.2863436123348019e-05, + "loss": 0.1951, + "step": 3825 + }, + { + "epoch": 8.480440771349862, + "grad_norm": 53.31047439575195, + "learning_rate": 1.2802251590797848e-05, + "loss": 0.2149, + "step": 3850 + }, + { + "epoch": 8.535537190082644, + "grad_norm": 8.599199295043945, + "learning_rate": 1.2741067058247676e-05, + "loss": 0.2088, + "step": 3875 + }, + { + "epoch": 8.590633608815427, + "grad_norm": 11.523205757141113, + "learning_rate": 1.2679882525697505e-05, + "loss": 0.2203, + "step": 3900 + }, + { + "epoch": 8.64573002754821, + "grad_norm": 7.496496200561523, + "learning_rate": 1.2618697993147334e-05, + "loss": 0.2012, + "step": 3925 + }, + { + "epoch": 8.700826446280992, + "grad_norm": 13.954992294311523, + "learning_rate": 1.2557513460597162e-05, + "loss": 0.1927, + "step": 3950 + }, + { + "epoch": 8.755922865013774, + "grad_norm": 15.341297149658203, + "learning_rate": 1.2496328928046991e-05, + "loss": 0.1978, + "step": 3975 + }, + { + "epoch": 8.811019283746557, + "grad_norm": 11.205670356750488, + "learning_rate": 1.243514439549682e-05, + "loss": 0.2137, + "step": 4000 + }, + { + "epoch": 8.86611570247934, + "grad_norm": 9.34787368774414, + "learning_rate": 1.2373959862946646e-05, + "loss": 0.1875, + "step": 4025 + }, + { + "epoch": 8.921212121212122, + "grad_norm": 34.0437126159668, + "learning_rate": 1.2312775330396475e-05, + "loss": 0.2181, + "step": 4050 + }, + { + "epoch": 8.976308539944904, + "grad_norm": 14.52303695678711, + "learning_rate": 1.2251590797846304e-05, + "loss": 0.1907, + "step": 4075 + }, + { + "epoch": 9.0, + "eval_accuracy": 0.21119902617163724, + "eval_loss": 1.4581968784332275, + "eval_macro_f1": 0.5448334885010536, + "eval_micro_f1": 0.5822896714306763, + "eval_runtime": 4.3167, + "eval_samples_per_second": 380.615, + "eval_steps_per_second": 23.861, + "eval_weighted_f1": 0.5834233474128085, + "step": 4086 + }, + { + "epoch": 9.030853994490359, + "grad_norm": 4.578360080718994, + "learning_rate": 1.2190406265296136e-05, + "loss": 0.1685, + "step": 4100 + }, + { + "epoch": 9.08595041322314, + "grad_norm": 6.535833835601807, + "learning_rate": 1.2129221732745964e-05, + "loss": 0.1444, + "step": 4125 + }, + { + "epoch": 9.141046831955922, + "grad_norm": 21.845661163330078, + "learning_rate": 1.2068037200195793e-05, + "loss": 0.1415, + "step": 4150 + }, + { + "epoch": 9.196143250688705, + "grad_norm": 6.381717681884766, + "learning_rate": 1.2006852667645622e-05, + "loss": 0.1482, + "step": 4175 + }, + { + "epoch": 9.251239669421487, + "grad_norm": 5.187005996704102, + "learning_rate": 1.1945668135095449e-05, + "loss": 0.1485, + "step": 4200 + }, + { + "epoch": 9.30633608815427, + "grad_norm": 13.340907096862793, + "learning_rate": 1.1884483602545277e-05, + "loss": 0.1329, + "step": 4225 + }, + { + "epoch": 9.361432506887052, + "grad_norm": 36.87532424926758, + "learning_rate": 1.1823299069995106e-05, + "loss": 0.1497, + "step": 4250 + }, + { + "epoch": 9.416528925619835, + "grad_norm": 7.7444353103637695, + "learning_rate": 1.1762114537444935e-05, + "loss": 0.1444, + "step": 4275 + }, + { + "epoch": 9.471625344352617, + "grad_norm": 7.527986526489258, + "learning_rate": 1.1700930004894763e-05, + "loss": 0.1659, + "step": 4300 + }, + { + "epoch": 9.5267217630854, + "grad_norm": 22.333158493041992, + "learning_rate": 1.1639745472344592e-05, + "loss": 0.1475, + "step": 4325 + }, + { + "epoch": 9.581818181818182, + "grad_norm": 18.358163833618164, + "learning_rate": 1.157856093979442e-05, + "loss": 0.1389, + "step": 4350 + }, + { + "epoch": 9.636914600550964, + "grad_norm": 13.2366943359375, + "learning_rate": 1.151737640724425e-05, + "loss": 0.1351, + "step": 4375 + }, + { + "epoch": 9.692011019283747, + "grad_norm": 19.432950973510742, + "learning_rate": 1.1456191874694078e-05, + "loss": 0.1402, + "step": 4400 + }, + { + "epoch": 9.74710743801653, + "grad_norm": 20.012887954711914, + "learning_rate": 1.1395007342143907e-05, + "loss": 0.1273, + "step": 4425 + }, + { + "epoch": 9.802203856749312, + "grad_norm": 10.137861251831055, + "learning_rate": 1.1333822809593735e-05, + "loss": 0.1442, + "step": 4450 + }, + { + "epoch": 9.857300275482094, + "grad_norm": 6.205521106719971, + "learning_rate": 1.1272638277043564e-05, + "loss": 0.1314, + "step": 4475 + }, + { + "epoch": 9.912396694214877, + "grad_norm": 10.11016845703125, + "learning_rate": 1.1211453744493393e-05, + "loss": 0.1291, + "step": 4500 + }, + { + "epoch": 9.967493112947658, + "grad_norm": 11.3107271194458, + "learning_rate": 1.1150269211943221e-05, + "loss": 0.1299, + "step": 4525 + }, + { + "epoch": 10.0, + "eval_accuracy": 0.22580645161290322, + "eval_loss": 1.603807806968689, + "eval_macro_f1": 0.5329114163191738, + "eval_micro_f1": 0.5780066646470766, + "eval_runtime": 4.4406, + "eval_samples_per_second": 369.992, + "eval_steps_per_second": 23.195, + "eval_weighted_f1": 0.5765375104085819, + "step": 4540 + }, + { + "epoch": 10.022038567493112, + "grad_norm": 8.95523738861084, + "learning_rate": 1.108908467939305e-05, + "loss": 0.1119, + "step": 4550 + }, + { + "epoch": 10.077134986225895, + "grad_norm": 4.750641822814941, + "learning_rate": 1.1027900146842878e-05, + "loss": 0.1026, + "step": 4575 + }, + { + "epoch": 10.132231404958677, + "grad_norm": 7.050352096557617, + "learning_rate": 1.0966715614292707e-05, + "loss": 0.1054, + "step": 4600 + }, + { + "epoch": 10.18732782369146, + "grad_norm": 9.429535865783691, + "learning_rate": 1.0905531081742536e-05, + "loss": 0.0882, + "step": 4625 + }, + { + "epoch": 10.242424242424242, + "grad_norm": 9.429491996765137, + "learning_rate": 1.0844346549192366e-05, + "loss": 0.1133, + "step": 4650 + }, + { + "epoch": 10.297520661157025, + "grad_norm": 28.78905487060547, + "learning_rate": 1.0783162016642195e-05, + "loss": 0.101, + "step": 4675 + }, + { + "epoch": 10.352617079889807, + "grad_norm": 21.340988159179688, + "learning_rate": 1.0721977484092023e-05, + "loss": 0.0942, + "step": 4700 + }, + { + "epoch": 10.40771349862259, + "grad_norm": 11.676098823547363, + "learning_rate": 1.0660792951541852e-05, + "loss": 0.0991, + "step": 4725 + }, + { + "epoch": 10.462809917355372, + "grad_norm": 43.355167388916016, + "learning_rate": 1.059960841899168e-05, + "loss": 0.0933, + "step": 4750 + }, + { + "epoch": 10.517906336088155, + "grad_norm": 11.840376853942871, + "learning_rate": 1.053842388644151e-05, + "loss": 0.0977, + "step": 4775 + }, + { + "epoch": 10.573002754820937, + "grad_norm": 15.774160385131836, + "learning_rate": 1.0477239353891338e-05, + "loss": 0.078, + "step": 4800 + }, + { + "epoch": 10.62809917355372, + "grad_norm": 3.623091459274292, + "learning_rate": 1.0416054821341167e-05, + "loss": 0.0919, + "step": 4825 + }, + { + "epoch": 10.683195592286502, + "grad_norm": 7.309466361999512, + "learning_rate": 1.0354870288790995e-05, + "loss": 0.0879, + "step": 4850 + }, + { + "epoch": 10.738292011019285, + "grad_norm": 3.9160447120666504, + "learning_rate": 1.0293685756240824e-05, + "loss": 0.0914, + "step": 4875 + }, + { + "epoch": 10.793388429752067, + "grad_norm": 7.319298267364502, + "learning_rate": 1.0232501223690653e-05, + "loss": 0.09, + "step": 4900 + }, + { + "epoch": 10.848484848484848, + "grad_norm": 21.43614959716797, + "learning_rate": 1.017131669114048e-05, + "loss": 0.0906, + "step": 4925 + }, + { + "epoch": 10.90358126721763, + "grad_norm": 7.134313106536865, + "learning_rate": 1.0110132158590308e-05, + "loss": 0.0792, + "step": 4950 + }, + { + "epoch": 10.958677685950413, + "grad_norm": 6.19244384765625, + "learning_rate": 1.0048947626040137e-05, + "loss": 0.0839, + "step": 4975 + }, + { + "epoch": 11.0, + "eval_accuracy": 0.23006695069993913, + "eval_loss": 1.7840896844863892, + "eval_macro_f1": 0.5385728837141956, + "eval_micro_f1": 0.5828379820149368, + "eval_runtime": 4.2876, + "eval_samples_per_second": 383.197, + "eval_steps_per_second": 24.023, + "eval_weighted_f1": 0.5803754426881687, + "step": 4994 + } + ], + "logging_steps": 25, + "max_steps": 9080, + "num_input_tokens_seen": 0, + "num_train_epochs": 20, + "save_steps": 500, + "stateful_callbacks": { + "EarlyStoppingCallback": { + "args": { + "early_stopping_patience": 3, + "early_stopping_threshold": 0.0 + }, + "attributes": { + "early_stopping_patience_counter": 3 + } + }, + "TrainerControl": { + "args": { + "should_epoch_stop": false, + "should_evaluate": false, + "should_log": false, + "should_save": true, + "should_training_stop": true + }, + "attributes": {} + } + }, + "total_flos": 0.0, + "train_batch_size": 8, + "trial_name": null, + "trial_params": null +} diff --git a/ood-vigoemotions-mmbert-source-source-02-s42-20260905T114918Z-bd528b37/model.safetensors b/ood-vigoemotions-mmbert-source-source-02-s42-20260905T114918Z-bd528b37/model.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..9bd121745f183adee392a52b4a30a991dbfe124d --- /dev/null +++ b/ood-vigoemotions-mmbert-source-source-02-s42-20260905T114918Z-bd528b37/model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:8d72baed5d27d9b3275e2827b9aee8bf96ce789319f8474e20a94bc1a6793167 +size 1227859104 diff --git a/ood-vigoemotions-mmbert-source-source-02-s43-20260905T133226Z-60628050/model.safetensors b/ood-vigoemotions-mmbert-source-source-02-s43-20260905T133226Z-60628050/model.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..c75552eb4125a80d7d21752f35415dc178cb59ff --- /dev/null +++ b/ood-vigoemotions-mmbert-source-source-02-s43-20260905T133226Z-60628050/model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:3e9048c8309b2d113c0242fa31ab59a8d4ed3557e5b9b7950a113dbe1f279185 +size 1227859104 diff --git a/ood-vigoemotions-mmbert-source-source-02-s43-20260905T133226Z-60628050/trainer/checkpoint-5797/special_tokens_map.json b/ood-vigoemotions-mmbert-source-source-02-s43-20260905T133226Z-60628050/trainer/checkpoint-5797/special_tokens_map.json new file mode 100644 index 0000000000000000000000000000000000000000..be4ad795f49075ac5e56d8108ef63be314a1e967 --- /dev/null +++ b/ood-vigoemotions-mmbert-source-source-02-s43-20260905T133226Z-60628050/trainer/checkpoint-5797/special_tokens_map.json @@ -0,0 +1,55 @@ +{ + "additional_special_tokens": [ + "", + "" + ], + "bos_token": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false + }, + "cls_token": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false + }, + "eos_token": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false + }, + "mask_token": { + "content": "", + "lstrip": true, + "normalized": false, + "rstrip": false, + "single_word": false + }, + "pad_token": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false + }, + "sep_token": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false + }, + "unk_token": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false + } +} diff --git a/ood-vigoemotions-mmbert-source-source-02-s43-20260905T133226Z-60628050/trainer/checkpoint-5797/tokenizer_config.json b/ood-vigoemotions-mmbert-source-source-02-s43-20260905T133226Z-60628050/trainer/checkpoint-5797/tokenizer_config.json new file mode 100644 index 0000000000000000000000000000000000000000..b002fb8c2960be8ebf9c373d6e5784b4565d7edc --- /dev/null +++ b/ood-vigoemotions-mmbert-source-source-02-s43-20260905T133226Z-60628050/trainer/checkpoint-5797/tokenizer_config.json @@ -0,0 +1,2018 @@ +{ + "add_bos_token": true, + "added_tokens_decoder": { + "0": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "1": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "2": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "3": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "4": { + "content": "", + "lstrip": true, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "5": { + "content": "<2mass>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "6": { + "content": "[@BOS@]", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "7": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "8": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "9": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "10": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "11": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "12": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "13": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "14": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "15": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "16": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "17": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "18": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "19": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "20": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "21": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "22": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "23": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "24": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "25": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "26": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "27": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "28": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "29": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "30": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "31": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "32": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "33": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "34": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "35": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "36": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "37": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "38": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "39": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "40": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "41": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "42": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "43": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "44": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "45": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "46": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "47": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "48": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "49": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "50": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "51": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "52": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "53": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "54": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "55": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "56": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "57": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "58": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "59": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "60": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "61": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "62": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "63": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "64": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "65": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "66": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "67": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "68": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "69": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "70": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "71": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "72": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "73": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "74": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "75": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "76": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "77": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "78": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "79": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "80": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "81": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "82": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "83": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "84": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "85": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "86": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "87": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "88": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "89": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "90": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "91": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "92": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "93": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "94": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "95": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "96": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "97": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "98": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "99": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "100": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "101": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "102": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "103": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "104": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "105": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "106": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "107": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "108": { + "content": "\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "109": { + "content": "\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "110": { + "content": "\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "111": { + "content": "\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "112": { + "content": "\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "113": { + "content": "\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "114": { + "content": "\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "115": { + "content": "\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "116": { + "content": "\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "117": { + "content": "\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "118": { + "content": "\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "119": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "120": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "121": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "122": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "123": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "124": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "125": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "126": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "127": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "128": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "129": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "130": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "131": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "132": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "133": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "134": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "135": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "136": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "137": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "138": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "139": { + "content": "▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "140": { + "content": "▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "141": { + "content": "▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "142": { + "content": "▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "143": { + "content": "▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "144": { + "content": "▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "145": { + "content": "▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "146": { + "content": "▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "147": { + "content": "▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "148": { + "content": "▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "149": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "150": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "151": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "152": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "153": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "154": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "155": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "156": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "157": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "158": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "159": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "160": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "161": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "162": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "163": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "164": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "165": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "166": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "167": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "168": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "169": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "170": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "172": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "173": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "174": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "175": { + "content": "
", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "171": { + "content": "
", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "176": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "177": { + "content": "
", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "178": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "179": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "180": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "181": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "182": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "183": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "184": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "185": { + "content": "

", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "186": { + "content": "

", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "187": { + "content": "

", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "188": { + "content": "

", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "189": { + "content": "

", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "190": { + "content": "
", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "191": { + "content": "
", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "192": { + "content": "
", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "193": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "194": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "195": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "196": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "197": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "198": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "199": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "200": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "201": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "202": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "203": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "204": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "205": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "206": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "207": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "208": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "209": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "210": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "211": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "212": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "213": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "214": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "215": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "216": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255968": { + "content": "[toxicity=0]", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255969": { + "content": "\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255970": { + "content": "\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255971": { + "content": "\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255972": { + "content": "\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255973": { + "content": "\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255974": { + "content": "\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255975": { + "content": "\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255976": { + "content": "\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255977": { + "content": "\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255978": { + "content": "\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255979": { + "content": "\t\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255980": { + "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255981": { + "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255982": { + "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255983": { + "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255984": { + "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255985": { + "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255986": { + "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255987": { + "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255988": { + "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255989": { + "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255990": { + "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255991": { + "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255992": { + "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255993": { + "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255994": { + "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255995": { + "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255996": { + "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255997": { + "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255998": { + "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255999": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + } + }, + "additional_special_tokens": [ + "", + "" + ], + "bos_token": "", + "clean_up_tokenization_spaces": false, + "cls_token": "", + "eos_token": "", + "extra_special_tokens": {}, + "mask_token": "", + "model_input_names": [ + "input_ids", + "attention_mask" + ], + "model_max_length": 8192, + "pad_token": "", + "padding_side": "right", + "sep_token": "", + "spaces_between_special_tokens": false, + "tokenizer_class": "PreTrainedTokenizerFast", + "unk_token": "" +} diff --git a/ood-vigoemotions-mmbert-source-source-02-s43-20260905T133226Z-60628050/trainer/checkpoint-5797/trainer_state.json b/ood-vigoemotions-mmbert-source-source-02-s43-20260905T133226Z-60628050/trainer/checkpoint-5797/trainer_state.json new file mode 100644 index 0000000000000000000000000000000000000000..5e3c0dcb6db518e29d029979eacbb76db1dfa5bc --- /dev/null +++ b/ood-vigoemotions-mmbert-source-source-02-s43-20260905T133226Z-60628050/trainer/checkpoint-5797/trainer_state.json @@ -0,0 +1,1792 @@ +{ + "best_global_step": 5797, + "best_metric": 0.6045200976640341, + "best_model_checkpoint": "/root/VNEncoder/results/checkpoints/ood-vigoemotions-mmbert-source-source-02-s43-20260905T133226Z-60628050/trainer/checkpoint-5797", + "epoch": 11.0, + "eval_steps": 500, + "global_step": 5797, + "is_hyper_param_search": false, + "is_local_process_zero": true, + "is_world_process_zero": true, + "log_history": [ + { + "epoch": 0.04748338081671415, + "grad_norm": 42.466888427734375, + "learning_rate": 4.554079696394687e-07, + "loss": 6.1416, + "step": 25 + }, + { + "epoch": 0.0949667616334283, + "grad_norm": 38.53072738647461, + "learning_rate": 9.297912713472485e-07, + "loss": 6.3527, + "step": 50 + }, + { + "epoch": 0.14245014245014245, + "grad_norm": 113.28205871582031, + "learning_rate": 1.4041745730550286e-06, + "loss": 6.0832, + "step": 75 + }, + { + "epoch": 0.1899335232668566, + "grad_norm": 73.18839263916016, + "learning_rate": 1.8785578747628086e-06, + "loss": 5.8318, + "step": 100 + }, + { + "epoch": 0.23741690408357075, + "grad_norm": 75.44960021972656, + "learning_rate": 2.3529411764705885e-06, + "loss": 5.7708, + "step": 125 + }, + { + "epoch": 0.2849002849002849, + "grad_norm": 61.81719207763672, + "learning_rate": 2.8273244781783683e-06, + "loss": 5.5055, + "step": 150 + }, + { + "epoch": 0.33238366571699907, + "grad_norm": 64.89173126220703, + "learning_rate": 3.3017077798861485e-06, + "loss": 5.5596, + "step": 175 + }, + { + "epoch": 0.3798670465337132, + "grad_norm": 35.145381927490234, + "learning_rate": 3.7760910815939284e-06, + "loss": 5.5006, + "step": 200 + }, + { + "epoch": 0.42735042735042733, + "grad_norm": 29.304231643676758, + "learning_rate": 4.250474383301708e-06, + "loss": 4.9728, + "step": 225 + }, + { + "epoch": 0.4748338081671415, + "grad_norm": 34.05004119873047, + "learning_rate": 4.724857685009488e-06, + "loss": 4.8357, + "step": 250 + }, + { + "epoch": 0.5223171889838556, + "grad_norm": 42.61710739135742, + "learning_rate": 5.199240986717268e-06, + "loss": 4.5948, + "step": 275 + }, + { + "epoch": 0.5698005698005698, + "grad_norm": 41.6121711730957, + "learning_rate": 5.673624288425048e-06, + "loss": 4.2142, + "step": 300 + }, + { + "epoch": 0.6172839506172839, + "grad_norm": 63.242279052734375, + "learning_rate": 6.1480075901328275e-06, + "loss": 4.0131, + "step": 325 + }, + { + "epoch": 0.6647673314339981, + "grad_norm": 61.35606002807617, + "learning_rate": 6.622390891840608e-06, + "loss": 3.9438, + "step": 350 + }, + { + "epoch": 0.7122507122507122, + "grad_norm": 64.1727523803711, + "learning_rate": 7.096774193548388e-06, + "loss": 3.8263, + "step": 375 + }, + { + "epoch": 0.7597340930674265, + "grad_norm": 48.41128158569336, + "learning_rate": 7.5711574952561674e-06, + "loss": 3.5959, + "step": 400 + }, + { + "epoch": 0.8072174738841406, + "grad_norm": 42.1629753112793, + "learning_rate": 8.045540796963948e-06, + "loss": 3.5873, + "step": 425 + }, + { + "epoch": 0.8547008547008547, + "grad_norm": 36.21388244628906, + "learning_rate": 8.519924098671729e-06, + "loss": 3.4469, + "step": 450 + }, + { + "epoch": 0.9021842355175689, + "grad_norm": 45.33504867553711, + "learning_rate": 8.994307400379507e-06, + "loss": 3.3685, + "step": 475 + }, + { + "epoch": 0.949667616334283, + "grad_norm": 53.4200439453125, + "learning_rate": 9.468690702087288e-06, + "loss": 3.2551, + "step": 500 + }, + { + "epoch": 0.9971509971509972, + "grad_norm": 52.55549621582031, + "learning_rate": 9.943074003795067e-06, + "loss": 3.2891, + "step": 525 + }, + { + "epoch": 1.0, + "eval_accuracy": 0.0016137708445400753, + "eval_loss": 0.7739421129226685, + "eval_macro_f1": 0.36155501356138686, + "eval_micro_f1": 0.37869281045751635, + "eval_runtime": 5.8168, + "eval_samples_per_second": 319.594, + "eval_steps_per_second": 20.114, + "eval_weighted_f1": 0.41253656843404873, + "step": 527 + }, + { + "epoch": 1.043684710351377, + "grad_norm": 38.479248046875, + "learning_rate": 1.0417457305502846e-05, + "loss": 2.8703, + "step": 550 + }, + { + "epoch": 1.0911680911680912, + "grad_norm": 58.546451568603516, + "learning_rate": 1.0891840607210626e-05, + "loss": 2.9037, + "step": 575 + }, + { + "epoch": 1.1386514719848053, + "grad_norm": 80.63526153564453, + "learning_rate": 1.1366223908918409e-05, + "loss": 2.8785, + "step": 600 + }, + { + "epoch": 1.1861348528015194, + "grad_norm": 59.79234313964844, + "learning_rate": 1.1840607210626186e-05, + "loss": 2.9307, + "step": 625 + }, + { + "epoch": 1.2336182336182335, + "grad_norm": 27.5277156829834, + "learning_rate": 1.2314990512333966e-05, + "loss": 2.9974, + "step": 650 + }, + { + "epoch": 1.2811016144349479, + "grad_norm": 47.518856048583984, + "learning_rate": 1.2789373814041749e-05, + "loss": 2.7249, + "step": 675 + }, + { + "epoch": 1.328584995251662, + "grad_norm": 49.7656364440918, + "learning_rate": 1.3263757115749526e-05, + "loss": 2.9259, + "step": 700 + }, + { + "epoch": 1.376068376068376, + "grad_norm": 54.58274841308594, + "learning_rate": 1.3738140417457306e-05, + "loss": 2.7116, + "step": 725 + }, + { + "epoch": 1.4235517568850902, + "grad_norm": 34.950828552246094, + "learning_rate": 1.4212523719165085e-05, + "loss": 2.6866, + "step": 750 + }, + { + "epoch": 1.4710351377018043, + "grad_norm": 52.59558868408203, + "learning_rate": 1.4686907020872866e-05, + "loss": 2.7492, + "step": 775 + }, + { + "epoch": 1.5185185185185186, + "grad_norm": 44.815189361572266, + "learning_rate": 1.5161290322580646e-05, + "loss": 2.7807, + "step": 800 + }, + { + "epoch": 1.5660018993352327, + "grad_norm": 50.69743728637695, + "learning_rate": 1.5635673624288425e-05, + "loss": 2.6976, + "step": 825 + }, + { + "epoch": 1.6134852801519468, + "grad_norm": 50.275367736816406, + "learning_rate": 1.6110056925996208e-05, + "loss": 2.7653, + "step": 850 + }, + { + "epoch": 1.660968660968661, + "grad_norm": 39.01410675048828, + "learning_rate": 1.6584440227703987e-05, + "loss": 2.6573, + "step": 875 + }, + { + "epoch": 1.708452041785375, + "grad_norm": 70.65179443359375, + "learning_rate": 1.7058823529411767e-05, + "loss": 2.9019, + "step": 900 + }, + { + "epoch": 1.7559354226020893, + "grad_norm": 35.23176193237305, + "learning_rate": 1.7533206831119546e-05, + "loss": 2.6363, + "step": 925 + }, + { + "epoch": 1.8034188034188035, + "grad_norm": 45.34317398071289, + "learning_rate": 1.8007590132827325e-05, + "loss": 2.7625, + "step": 950 + }, + { + "epoch": 1.8509021842355176, + "grad_norm": 37.30549240112305, + "learning_rate": 1.8481973434535105e-05, + "loss": 2.7392, + "step": 975 + }, + { + "epoch": 1.8983855650522319, + "grad_norm": 44.2552490234375, + "learning_rate": 1.8956356736242888e-05, + "loss": 2.7414, + "step": 1000 + }, + { + "epoch": 1.9458689458689458, + "grad_norm": 35.74271774291992, + "learning_rate": 1.9430740037950667e-05, + "loss": 2.692, + "step": 1025 + }, + { + "epoch": 1.99335232668566, + "grad_norm": 43.257347106933594, + "learning_rate": 1.9905123339658446e-05, + "loss": 2.6575, + "step": 1050 + }, + { + "epoch": 2.0, + "eval_accuracy": 0.016675632060247445, + "eval_loss": 0.6451638340950012, + "eval_macro_f1": 0.4216957664234933, + "eval_micro_f1": 0.43330427201394944, + "eval_runtime": 4.9871, + "eval_samples_per_second": 372.762, + "eval_steps_per_second": 23.461, + "eval_weighted_f1": 0.4744003458589279, + "step": 1054 + }, + { + "epoch": 2.03988603988604, + "grad_norm": 35.80771255493164, + "learning_rate": 1.9957832595403753e-05, + "loss": 2.0608, + "step": 1075 + }, + { + "epoch": 2.087369420702754, + "grad_norm": 31.177793502807617, + "learning_rate": 1.9905123339658446e-05, + "loss": 1.9354, + "step": 1100 + }, + { + "epoch": 2.134852801519468, + "grad_norm": 21.678997039794922, + "learning_rate": 1.9852414083913136e-05, + "loss": 2.0003, + "step": 1125 + }, + { + "epoch": 2.1823361823361824, + "grad_norm": 38.477752685546875, + "learning_rate": 1.9799704828167827e-05, + "loss": 1.9289, + "step": 1150 + }, + { + "epoch": 2.2298195631528963, + "grad_norm": 31.992578506469727, + "learning_rate": 1.9746995572422517e-05, + "loss": 1.9564, + "step": 1175 + }, + { + "epoch": 2.2773029439696106, + "grad_norm": 69.55426788330078, + "learning_rate": 1.969428631667721e-05, + "loss": 2.0287, + "step": 1200 + }, + { + "epoch": 2.324786324786325, + "grad_norm": 31.55048370361328, + "learning_rate": 1.9641577060931903e-05, + "loss": 2.1207, + "step": 1225 + }, + { + "epoch": 2.372269705603039, + "grad_norm": 51.29478454589844, + "learning_rate": 1.9588867805186593e-05, + "loss": 1.9785, + "step": 1250 + }, + { + "epoch": 2.419753086419753, + "grad_norm": 43.476219177246094, + "learning_rate": 1.9536158549441284e-05, + "loss": 2.0041, + "step": 1275 + }, + { + "epoch": 2.467236467236467, + "grad_norm": 30.011032104492188, + "learning_rate": 1.9483449293695974e-05, + "loss": 2.0289, + "step": 1300 + }, + { + "epoch": 2.5147198480531814, + "grad_norm": 34.02845001220703, + "learning_rate": 1.9430740037950667e-05, + "loss": 1.9928, + "step": 1325 + }, + { + "epoch": 2.5622032288698957, + "grad_norm": 39.73603057861328, + "learning_rate": 1.9378030782205357e-05, + "loss": 2.0899, + "step": 1350 + }, + { + "epoch": 2.6096866096866096, + "grad_norm": 42.63766098022461, + "learning_rate": 1.9325321526460047e-05, + "loss": 1.9751, + "step": 1375 + }, + { + "epoch": 2.657169990503324, + "grad_norm": 33.42068862915039, + "learning_rate": 1.927261227071474e-05, + "loss": 2.0424, + "step": 1400 + }, + { + "epoch": 2.7046533713200382, + "grad_norm": 68.12841796875, + "learning_rate": 1.921990301496943e-05, + "loss": 1.9421, + "step": 1425 + }, + { + "epoch": 2.752136752136752, + "grad_norm": 34.54705047607422, + "learning_rate": 1.916719375922412e-05, + "loss": 1.9273, + "step": 1450 + }, + { + "epoch": 2.799620132953466, + "grad_norm": 26.91336441040039, + "learning_rate": 1.911448450347881e-05, + "loss": 1.9383, + "step": 1475 + }, + { + "epoch": 2.8471035137701803, + "grad_norm": 60.04805374145508, + "learning_rate": 1.9061775247733504e-05, + "loss": 2.0682, + "step": 1500 + }, + { + "epoch": 2.8945868945868947, + "grad_norm": 23.979312896728516, + "learning_rate": 1.9009065991988194e-05, + "loss": 1.9271, + "step": 1525 + }, + { + "epoch": 2.9420702754036085, + "grad_norm": 32.707942962646484, + "learning_rate": 1.8956356736242888e-05, + "loss": 2.0019, + "step": 1550 + }, + { + "epoch": 2.989553656220323, + "grad_norm": 29.024099349975586, + "learning_rate": 1.8903647480497578e-05, + "loss": 1.9078, + "step": 1575 + }, + { + "epoch": 3.0, + "eval_accuracy": 0.04410973641742873, + "eval_loss": 0.640509843826294, + "eval_macro_f1": 0.474925227989289, + "eval_micro_f1": 0.48038184612895396, + "eval_runtime": 5.0213, + "eval_samples_per_second": 370.219, + "eval_steps_per_second": 23.301, + "eval_weighted_f1": 0.5151251185527903, + "step": 1581 + }, + { + "epoch": 3.0360873694207027, + "grad_norm": 24.103885650634766, + "learning_rate": 1.8850938224752268e-05, + "loss": 1.5437, + "step": 1600 + }, + { + "epoch": 3.083570750237417, + "grad_norm": 42.41041946411133, + "learning_rate": 1.879822896900696e-05, + "loss": 1.4517, + "step": 1625 + }, + { + "epoch": 3.131054131054131, + "grad_norm": 15.861024856567383, + "learning_rate": 1.874551971326165e-05, + "loss": 1.3673, + "step": 1650 + }, + { + "epoch": 3.178537511870845, + "grad_norm": 19.538061141967773, + "learning_rate": 1.869281045751634e-05, + "loss": 1.3837, + "step": 1675 + }, + { + "epoch": 3.2260208926875595, + "grad_norm": 46.26189041137695, + "learning_rate": 1.864010120177103e-05, + "loss": 1.3156, + "step": 1700 + }, + { + "epoch": 3.2735042735042734, + "grad_norm": 36.7100715637207, + "learning_rate": 1.8587391946025725e-05, + "loss": 1.3421, + "step": 1725 + }, + { + "epoch": 3.3209876543209877, + "grad_norm": 40.29243850708008, + "learning_rate": 1.8534682690280415e-05, + "loss": 1.3264, + "step": 1750 + }, + { + "epoch": 3.3684710351377016, + "grad_norm": 22.646486282348633, + "learning_rate": 1.8481973434535105e-05, + "loss": 1.355, + "step": 1775 + }, + { + "epoch": 3.415954415954416, + "grad_norm": 24.885631561279297, + "learning_rate": 1.8429264178789795e-05, + "loss": 1.332, + "step": 1800 + }, + { + "epoch": 3.4634377967711303, + "grad_norm": 24.071887969970703, + "learning_rate": 1.8376554923044488e-05, + "loss": 1.258, + "step": 1825 + }, + { + "epoch": 3.510921177587844, + "grad_norm": 30.55681037902832, + "learning_rate": 1.832384566729918e-05, + "loss": 1.2578, + "step": 1850 + }, + { + "epoch": 3.5584045584045585, + "grad_norm": 37.01355743408203, + "learning_rate": 1.8271136411553872e-05, + "loss": 1.3362, + "step": 1875 + }, + { + "epoch": 3.6058879392212724, + "grad_norm": 30.750619888305664, + "learning_rate": 1.8218427155808562e-05, + "loss": 1.329, + "step": 1900 + }, + { + "epoch": 3.6533713200379867, + "grad_norm": 31.818098068237305, + "learning_rate": 1.8165717900063252e-05, + "loss": 1.3591, + "step": 1925 + }, + { + "epoch": 3.700854700854701, + "grad_norm": 26.725540161132812, + "learning_rate": 1.8113008644317945e-05, + "loss": 1.3032, + "step": 1950 + }, + { + "epoch": 3.748338081671415, + "grad_norm": 35.89231872558594, + "learning_rate": 1.8060299388572635e-05, + "loss": 1.3604, + "step": 1975 + }, + { + "epoch": 3.7958214624881292, + "grad_norm": 23.358871459960938, + "learning_rate": 1.8007590132827325e-05, + "loss": 1.2997, + "step": 2000 + }, + { + "epoch": 3.8433048433048436, + "grad_norm": 33.60404586791992, + "learning_rate": 1.7954880877082015e-05, + "loss": 1.2872, + "step": 2025 + }, + { + "epoch": 3.8907882241215574, + "grad_norm": 43.04498291015625, + "learning_rate": 1.790217162133671e-05, + "loss": 1.372, + "step": 2050 + }, + { + "epoch": 3.9382716049382713, + "grad_norm": 34.92791748046875, + "learning_rate": 1.78494623655914e-05, + "loss": 1.3739, + "step": 2075 + }, + { + "epoch": 3.9857549857549857, + "grad_norm": 61.5059814453125, + "learning_rate": 1.779675310984609e-05, + "loss": 1.2661, + "step": 2100 + }, + { + "epoch": 4.0, + "eval_accuracy": 0.07961269499731038, + "eval_loss": 0.7032579183578491, + "eval_macro_f1": 0.5178092021035117, + "eval_micro_f1": 0.5282948622486969, + "eval_runtime": 4.9726, + "eval_samples_per_second": 373.847, + "eval_steps_per_second": 23.529, + "eval_weighted_f1": 0.5480078641240479, + "step": 2108 + }, + { + "epoch": 4.032288698955366, + "grad_norm": 30.800107955932617, + "learning_rate": 1.7744043854100782e-05, + "loss": 1.0502, + "step": 2125 + }, + { + "epoch": 4.07977207977208, + "grad_norm": 25.957460403442383, + "learning_rate": 1.7691334598355472e-05, + "loss": 0.9406, + "step": 2150 + }, + { + "epoch": 4.127255460588794, + "grad_norm": 18.835384368896484, + "learning_rate": 1.7638625342610163e-05, + "loss": 0.909, + "step": 2175 + }, + { + "epoch": 4.174738841405508, + "grad_norm": 34.36490249633789, + "learning_rate": 1.7585916086864856e-05, + "loss": 0.8639, + "step": 2200 + }, + { + "epoch": 4.222222222222222, + "grad_norm": 42.17921829223633, + "learning_rate": 1.7533206831119546e-05, + "loss": 0.986, + "step": 2225 + }, + { + "epoch": 4.269705603038936, + "grad_norm": 18.649545669555664, + "learning_rate": 1.748049757537424e-05, + "loss": 0.9287, + "step": 2250 + }, + { + "epoch": 4.317188983855651, + "grad_norm": 48.732879638671875, + "learning_rate": 1.742778831962893e-05, + "loss": 0.9057, + "step": 2275 + }, + { + "epoch": 4.364672364672365, + "grad_norm": 24.9398193359375, + "learning_rate": 1.737507906388362e-05, + "loss": 0.9474, + "step": 2300 + }, + { + "epoch": 4.412155745489079, + "grad_norm": 11.954604148864746, + "learning_rate": 1.732236980813831e-05, + "loss": 0.9036, + "step": 2325 + }, + { + "epoch": 4.459639126305793, + "grad_norm": 18.34949493408203, + "learning_rate": 1.7269660552393003e-05, + "loss": 0.927, + "step": 2350 + }, + { + "epoch": 4.507122507122507, + "grad_norm": 78.10287475585938, + "learning_rate": 1.7216951296647693e-05, + "loss": 0.9124, + "step": 2375 + }, + { + "epoch": 4.554605887939221, + "grad_norm": 29.533727645874023, + "learning_rate": 1.7164242040902383e-05, + "loss": 0.9532, + "step": 2400 + }, + { + "epoch": 4.602089268755935, + "grad_norm": 20.754230499267578, + "learning_rate": 1.7111532785157073e-05, + "loss": 0.9455, + "step": 2425 + }, + { + "epoch": 4.64957264957265, + "grad_norm": 12.572599411010742, + "learning_rate": 1.7058823529411767e-05, + "loss": 0.9059, + "step": 2450 + }, + { + "epoch": 4.697056030389364, + "grad_norm": 25.214078903198242, + "learning_rate": 1.7006114273666457e-05, + "loss": 0.9483, + "step": 2475 + }, + { + "epoch": 4.744539411206078, + "grad_norm": 28.17908477783203, + "learning_rate": 1.6953405017921147e-05, + "loss": 0.8608, + "step": 2500 + }, + { + "epoch": 4.792022792022792, + "grad_norm": 27.811349868774414, + "learning_rate": 1.690069576217584e-05, + "loss": 0.8963, + "step": 2525 + }, + { + "epoch": 4.839506172839506, + "grad_norm": 34.6890869140625, + "learning_rate": 1.684798650643053e-05, + "loss": 0.9537, + "step": 2550 + }, + { + "epoch": 4.88698955365622, + "grad_norm": 41.03017807006836, + "learning_rate": 1.6795277250685224e-05, + "loss": 0.9382, + "step": 2575 + }, + { + "epoch": 4.934472934472934, + "grad_norm": 20.774999618530273, + "learning_rate": 1.6742567994939914e-05, + "loss": 0.9576, + "step": 2600 + }, + { + "epoch": 4.981956315289649, + "grad_norm": 50.060672760009766, + "learning_rate": 1.6689858739194604e-05, + "loss": 0.8801, + "step": 2625 + }, + { + "epoch": 5.0, + "eval_accuracy": 0.1312533620225928, + "eval_loss": 0.8617390990257263, + "eval_macro_f1": 0.5661674742863365, + "eval_micro_f1": 0.5719966513185433, + "eval_runtime": 4.8508, + "eval_samples_per_second": 383.236, + "eval_steps_per_second": 24.12, + "eval_weighted_f1": 0.5831312572156591, + "step": 2635 + }, + { + "epoch": 5.028490028490029, + "grad_norm": 18.490642547607422, + "learning_rate": 1.6637149483449294e-05, + "loss": 0.7098, + "step": 2650 + }, + { + "epoch": 5.075973409306743, + "grad_norm": 15.921799659729004, + "learning_rate": 1.6584440227703987e-05, + "loss": 0.6619, + "step": 2675 + }, + { + "epoch": 5.1234567901234565, + "grad_norm": 14.693387031555176, + "learning_rate": 1.6531730971958677e-05, + "loss": 0.6622, + "step": 2700 + }, + { + "epoch": 5.170940170940171, + "grad_norm": 13.369607925415039, + "learning_rate": 1.6479021716213367e-05, + "loss": 0.6234, + "step": 2725 + }, + { + "epoch": 5.218423551756885, + "grad_norm": 11.756754875183105, + "learning_rate": 1.642631246046806e-05, + "loss": 0.6504, + "step": 2750 + }, + { + "epoch": 5.265906932573599, + "grad_norm": 32.03675842285156, + "learning_rate": 1.637360320472275e-05, + "loss": 0.6522, + "step": 2775 + }, + { + "epoch": 5.313390313390314, + "grad_norm": 71.17816925048828, + "learning_rate": 1.632089394897744e-05, + "loss": 0.6194, + "step": 2800 + }, + { + "epoch": 5.360873694207028, + "grad_norm": 15.817419052124023, + "learning_rate": 1.626818469323213e-05, + "loss": 0.6312, + "step": 2825 + }, + { + "epoch": 5.4083570750237415, + "grad_norm": 25.708629608154297, + "learning_rate": 1.6215475437486824e-05, + "loss": 0.632, + "step": 2850 + }, + { + "epoch": 5.455840455840455, + "grad_norm": 14.698016166687012, + "learning_rate": 1.6162766181741514e-05, + "loss": 0.6342, + "step": 2875 + }, + { + "epoch": 5.50332383665717, + "grad_norm": 31.988943099975586, + "learning_rate": 1.6110056925996208e-05, + "loss": 0.6859, + "step": 2900 + }, + { + "epoch": 5.550807217473884, + "grad_norm": 14.286497116088867, + "learning_rate": 1.6057347670250898e-05, + "loss": 0.6473, + "step": 2925 + }, + { + "epoch": 5.598290598290598, + "grad_norm": 41.547271728515625, + "learning_rate": 1.6004638414505588e-05, + "loss": 0.6853, + "step": 2950 + }, + { + "epoch": 5.645773979107313, + "grad_norm": 17.2784366607666, + "learning_rate": 1.595192915876028e-05, + "loss": 0.659, + "step": 2975 + }, + { + "epoch": 5.693257359924027, + "grad_norm": 60.431392669677734, + "learning_rate": 1.589921990301497e-05, + "loss": 0.5916, + "step": 3000 + }, + { + "epoch": 5.7407407407407405, + "grad_norm": 27.35256576538086, + "learning_rate": 1.584651064726966e-05, + "loss": 0.6493, + "step": 3025 + }, + { + "epoch": 5.788224121557455, + "grad_norm": 20.633012771606445, + "learning_rate": 1.579380139152435e-05, + "loss": 0.5941, + "step": 3050 + }, + { + "epoch": 5.835707502374169, + "grad_norm": 15.044947624206543, + "learning_rate": 1.5741092135779045e-05, + "loss": 0.6362, + "step": 3075 + }, + { + "epoch": 5.883190883190883, + "grad_norm": 14.263772010803223, + "learning_rate": 1.5688382880033735e-05, + "loss": 0.6214, + "step": 3100 + }, + { + "epoch": 5.930674264007598, + "grad_norm": 34.03923797607422, + "learning_rate": 1.5635673624288425e-05, + "loss": 0.615, + "step": 3125 + }, + { + "epoch": 5.978157644824312, + "grad_norm": 55.2652473449707, + "learning_rate": 1.5582964368543115e-05, + "loss": 0.6469, + "step": 3150 + }, + { + "epoch": 6.0, + "eval_accuracy": 0.17428725121032812, + "eval_loss": 0.930029571056366, + "eval_macro_f1": 0.5744715338078124, + "eval_micro_f1": 0.585635359116022, + "eval_runtime": 4.866, + "eval_samples_per_second": 382.041, + "eval_steps_per_second": 24.045, + "eval_weighted_f1": 0.5939596407496641, + "step": 3162 + }, + { + "epoch": 6.0246913580246915, + "grad_norm": 17.741565704345703, + "learning_rate": 1.553025511279781e-05, + "loss": 0.5214, + "step": 3175 + }, + { + "epoch": 6.072174738841405, + "grad_norm": 22.976102828979492, + "learning_rate": 1.5477545857052502e-05, + "loss": 0.4418, + "step": 3200 + }, + { + "epoch": 6.119658119658119, + "grad_norm": 20.000131607055664, + "learning_rate": 1.5424836601307192e-05, + "loss": 0.4645, + "step": 3225 + }, + { + "epoch": 6.167141500474834, + "grad_norm": 12.409162521362305, + "learning_rate": 1.5372127345561882e-05, + "loss": 0.4078, + "step": 3250 + }, + { + "epoch": 6.214624881291548, + "grad_norm": 11.717437744140625, + "learning_rate": 1.5319418089816572e-05, + "loss": 0.4454, + "step": 3275 + }, + { + "epoch": 6.262108262108262, + "grad_norm": 19.861845016479492, + "learning_rate": 1.5266708834071265e-05, + "loss": 0.4444, + "step": 3300 + }, + { + "epoch": 6.3095916429249765, + "grad_norm": 23.135236740112305, + "learning_rate": 1.5213999578325956e-05, + "loss": 0.3956, + "step": 3325 + }, + { + "epoch": 6.35707502374169, + "grad_norm": 28.478940963745117, + "learning_rate": 1.5161290322580646e-05, + "loss": 0.4448, + "step": 3350 + }, + { + "epoch": 6.404558404558404, + "grad_norm": 17.0195255279541, + "learning_rate": 1.5108581066835337e-05, + "loss": 0.4401, + "step": 3375 + }, + { + "epoch": 6.452041785375119, + "grad_norm": 12.3309965133667, + "learning_rate": 1.5055871811090027e-05, + "loss": 0.4316, + "step": 3400 + }, + { + "epoch": 6.499525166191833, + "grad_norm": 11.332566261291504, + "learning_rate": 1.5003162555344719e-05, + "loss": 0.4079, + "step": 3425 + }, + { + "epoch": 6.547008547008547, + "grad_norm": 26.008562088012695, + "learning_rate": 1.4950453299599411e-05, + "loss": 0.4256, + "step": 3450 + }, + { + "epoch": 6.594491927825262, + "grad_norm": 10.497772216796875, + "learning_rate": 1.4897744043854101e-05, + "loss": 0.4626, + "step": 3475 + }, + { + "epoch": 6.6419753086419755, + "grad_norm": 13.758752822875977, + "learning_rate": 1.4845034788108794e-05, + "loss": 0.4252, + "step": 3500 + }, + { + "epoch": 6.689458689458689, + "grad_norm": 27.361492156982422, + "learning_rate": 1.4792325532363484e-05, + "loss": 0.4509, + "step": 3525 + }, + { + "epoch": 6.736942070275403, + "grad_norm": 14.832262992858887, + "learning_rate": 1.4739616276618176e-05, + "loss": 0.4464, + "step": 3550 + }, + { + "epoch": 6.784425451092118, + "grad_norm": 33.7310791015625, + "learning_rate": 1.4686907020872866e-05, + "loss": 0.4238, + "step": 3575 + }, + { + "epoch": 6.831908831908832, + "grad_norm": 29.504594802856445, + "learning_rate": 1.4634197765127558e-05, + "loss": 0.4699, + "step": 3600 + }, + { + "epoch": 6.879392212725546, + "grad_norm": 42.16069412231445, + "learning_rate": 1.458148850938225e-05, + "loss": 0.4402, + "step": 3625 + }, + { + "epoch": 6.926875593542261, + "grad_norm": 43.56899642944336, + "learning_rate": 1.452877925363694e-05, + "loss": 0.4778, + "step": 3650 + }, + { + "epoch": 6.9743589743589745, + "grad_norm": 13.3931245803833, + "learning_rate": 1.4476069997891631e-05, + "loss": 0.4515, + "step": 3675 + }, + { + "epoch": 7.0, + "eval_accuracy": 0.2087143625605164, + "eval_loss": 1.1090645790100098, + "eval_macro_f1": 0.585298796573478, + "eval_micro_f1": 0.5964482547458665, + "eval_runtime": 4.8885, + "eval_samples_per_second": 380.277, + "eval_steps_per_second": 23.933, + "eval_weighted_f1": 0.6004246347046586, + "step": 3689 + }, + { + "epoch": 7.020892687559354, + "grad_norm": 15.4351224899292, + "learning_rate": 1.4423360742146321e-05, + "loss": 0.3949, + "step": 3700 + }, + { + "epoch": 7.068376068376068, + "grad_norm": 11.938258171081543, + "learning_rate": 1.4370651486401013e-05, + "loss": 0.2897, + "step": 3725 + }, + { + "epoch": 7.115859449192783, + "grad_norm": 19.71200180053711, + "learning_rate": 1.4317942230655703e-05, + "loss": 0.2975, + "step": 3750 + }, + { + "epoch": 7.163342830009497, + "grad_norm": 13.208940505981445, + "learning_rate": 1.4265232974910395e-05, + "loss": 0.3104, + "step": 3775 + }, + { + "epoch": 7.210826210826211, + "grad_norm": 8.498453140258789, + "learning_rate": 1.4212523719165085e-05, + "loss": 0.2945, + "step": 3800 + }, + { + "epoch": 7.258309591642925, + "grad_norm": 9.174602508544922, + "learning_rate": 1.4159814463419779e-05, + "loss": 0.3014, + "step": 3825 + }, + { + "epoch": 7.305792972459639, + "grad_norm": 24.003524780273438, + "learning_rate": 1.410710520767447e-05, + "loss": 0.3163, + "step": 3850 + }, + { + "epoch": 7.353276353276353, + "grad_norm": 19.13205909729004, + "learning_rate": 1.405439595192916e-05, + "loss": 0.3171, + "step": 3875 + }, + { + "epoch": 7.400759734093067, + "grad_norm": 15.435404777526855, + "learning_rate": 1.4001686696183852e-05, + "loss": 0.2934, + "step": 3900 + }, + { + "epoch": 7.448243114909782, + "grad_norm": 10.802959442138672, + "learning_rate": 1.3948977440438542e-05, + "loss": 0.2699, + "step": 3925 + }, + { + "epoch": 7.495726495726496, + "grad_norm": 6.724036693572998, + "learning_rate": 1.3896268184693234e-05, + "loss": 0.2723, + "step": 3950 + }, + { + "epoch": 7.54320987654321, + "grad_norm": 8.457695960998535, + "learning_rate": 1.3843558928947924e-05, + "loss": 0.2864, + "step": 3975 + }, + { + "epoch": 7.590693257359924, + "grad_norm": 59.64891052246094, + "learning_rate": 1.3790849673202616e-05, + "loss": 0.3078, + "step": 4000 + }, + { + "epoch": 7.638176638176638, + "grad_norm": 13.899282455444336, + "learning_rate": 1.3738140417457306e-05, + "loss": 0.2878, + "step": 4025 + }, + { + "epoch": 7.685660018993352, + "grad_norm": 22.03400421142578, + "learning_rate": 1.3685431161711997e-05, + "loss": 0.3049, + "step": 4050 + }, + { + "epoch": 7.733143399810066, + "grad_norm": 10.780784606933594, + "learning_rate": 1.3632721905966687e-05, + "loss": 0.2804, + "step": 4075 + }, + { + "epoch": 7.780626780626781, + "grad_norm": 40.84187698364258, + "learning_rate": 1.358001265022138e-05, + "loss": 0.3037, + "step": 4100 + }, + { + "epoch": 7.828110161443495, + "grad_norm": 16.029813766479492, + "learning_rate": 1.3527303394476071e-05, + "loss": 0.3138, + "step": 4125 + }, + { + "epoch": 7.875593542260209, + "grad_norm": 17.134769439697266, + "learning_rate": 1.3474594138730763e-05, + "loss": 0.3241, + "step": 4150 + }, + { + "epoch": 7.923076923076923, + "grad_norm": 32.032840728759766, + "learning_rate": 1.3421884882985454e-05, + "loss": 0.294, + "step": 4175 + }, + { + "epoch": 7.970560303893637, + "grad_norm": 21.15312385559082, + "learning_rate": 1.3369175627240144e-05, + "loss": 0.2793, + "step": 4200 + }, + { + "epoch": 8.0, + "eval_accuracy": 0.2140935987089833, + "eval_loss": 1.2231533527374268, + "eval_macro_f1": 0.5940205209972464, + "eval_micro_f1": 0.6054839113559535, + "eval_runtime": 4.9611, + "eval_samples_per_second": 374.717, + "eval_steps_per_second": 23.584, + "eval_weighted_f1": 0.6057940584072086, + "step": 4216 + }, + { + "epoch": 8.017094017094017, + "grad_norm": 10.781094551086426, + "learning_rate": 1.3316466371494836e-05, + "loss": 0.2485, + "step": 4225 + }, + { + "epoch": 8.064577397910732, + "grad_norm": 10.791547775268555, + "learning_rate": 1.3263757115749526e-05, + "loss": 0.2238, + "step": 4250 + }, + { + "epoch": 8.112060778727445, + "grad_norm": 19.549699783325195, + "learning_rate": 1.3211047860004218e-05, + "loss": 0.2224, + "step": 4275 + }, + { + "epoch": 8.15954415954416, + "grad_norm": 12.872920036315918, + "learning_rate": 1.315833860425891e-05, + "loss": 0.2067, + "step": 4300 + }, + { + "epoch": 8.207027540360874, + "grad_norm": 7.634912490844727, + "learning_rate": 1.31056293485136e-05, + "loss": 0.2086, + "step": 4325 + }, + { + "epoch": 8.254510921177587, + "grad_norm": 28.830102920532227, + "learning_rate": 1.3052920092768292e-05, + "loss": 0.2017, + "step": 4350 + }, + { + "epoch": 8.301994301994302, + "grad_norm": 5.609751224517822, + "learning_rate": 1.3000210837022982e-05, + "loss": 0.1863, + "step": 4375 + }, + { + "epoch": 8.349477682811017, + "grad_norm": 15.390376091003418, + "learning_rate": 1.2947501581277673e-05, + "loss": 0.1978, + "step": 4400 + }, + { + "epoch": 8.39696106362773, + "grad_norm": 7.242794513702393, + "learning_rate": 1.2894792325532363e-05, + "loss": 0.2013, + "step": 4425 + }, + { + "epoch": 8.444444444444445, + "grad_norm": 33.799224853515625, + "learning_rate": 1.2842083069787055e-05, + "loss": 0.2005, + "step": 4450 + }, + { + "epoch": 8.49192782526116, + "grad_norm": 7.0647358894348145, + "learning_rate": 1.2789373814041749e-05, + "loss": 0.2187, + "step": 4475 + }, + { + "epoch": 8.539411206077872, + "grad_norm": 12.728279113769531, + "learning_rate": 1.2736664558296439e-05, + "loss": 0.2135, + "step": 4500 + }, + { + "epoch": 8.586894586894587, + "grad_norm": 7.330479621887207, + "learning_rate": 1.268395530255113e-05, + "loss": 0.208, + "step": 4525 + }, + { + "epoch": 8.634377967711302, + "grad_norm": 7.347399711608887, + "learning_rate": 1.263124604680582e-05, + "loss": 0.1834, + "step": 4550 + }, + { + "epoch": 8.681861348528015, + "grad_norm": 19.936182022094727, + "learning_rate": 1.2578536791060512e-05, + "loss": 0.1962, + "step": 4575 + }, + { + "epoch": 8.72934472934473, + "grad_norm": 12.831972122192383, + "learning_rate": 1.2525827535315202e-05, + "loss": 0.1847, + "step": 4600 + }, + { + "epoch": 8.776828110161443, + "grad_norm": 6.2358903884887695, + "learning_rate": 1.2473118279569894e-05, + "loss": 0.1993, + "step": 4625 + }, + { + "epoch": 8.824311490978157, + "grad_norm": 9.986373901367188, + "learning_rate": 1.2420409023824584e-05, + "loss": 0.2021, + "step": 4650 + }, + { + "epoch": 8.871794871794872, + "grad_norm": 15.168353080749512, + "learning_rate": 1.2367699768079276e-05, + "loss": 0.2047, + "step": 4675 + }, + { + "epoch": 8.919278252611585, + "grad_norm": 18.249940872192383, + "learning_rate": 1.2314990512333966e-05, + "loss": 0.1867, + "step": 4700 + }, + { + "epoch": 8.9667616334283, + "grad_norm": 8.6532564163208, + "learning_rate": 1.2262281256588657e-05, + "loss": 0.1818, + "step": 4725 + }, + { + "epoch": 9.0, + "eval_accuracy": 0.23830016137708446, + "eval_loss": 1.4221587181091309, + "eval_macro_f1": 0.5937747690570366, + "eval_micro_f1": 0.6078904991948471, + "eval_runtime": 5.0553, + "eval_samples_per_second": 367.734, + "eval_steps_per_second": 23.144, + "eval_weighted_f1": 0.6074643571973501, + "step": 4743 + }, + { + "epoch": 9.01329534662868, + "grad_norm": 7.653168201446533, + "learning_rate": 1.2209572000843348e-05, + "loss": 0.1608, + "step": 4750 + }, + { + "epoch": 9.060778727445394, + "grad_norm": 25.898069381713867, + "learning_rate": 1.215686274509804e-05, + "loss": 0.1589, + "step": 4775 + }, + { + "epoch": 9.108262108262108, + "grad_norm": 7.836034297943115, + "learning_rate": 1.2104153489352733e-05, + "loss": 0.1556, + "step": 4800 + }, + { + "epoch": 9.155745489078823, + "grad_norm": 8.399602890014648, + "learning_rate": 1.2051444233607423e-05, + "loss": 0.1274, + "step": 4825 + }, + { + "epoch": 9.203228869895536, + "grad_norm": 15.443283081054688, + "learning_rate": 1.1998734977862115e-05, + "loss": 0.136, + "step": 4850 + }, + { + "epoch": 9.250712250712251, + "grad_norm": 25.787002563476562, + "learning_rate": 1.1946025722116805e-05, + "loss": 0.1332, + "step": 4875 + }, + { + "epoch": 9.298195631528966, + "grad_norm": 9.80220890045166, + "learning_rate": 1.1893316466371496e-05, + "loss": 0.1303, + "step": 4900 + }, + { + "epoch": 9.345679012345679, + "grad_norm": 8.279167175292969, + "learning_rate": 1.1840607210626186e-05, + "loss": 0.1354, + "step": 4925 + }, + { + "epoch": 9.393162393162394, + "grad_norm": 9.350815773010254, + "learning_rate": 1.1787897954880878e-05, + "loss": 0.1297, + "step": 4950 + }, + { + "epoch": 9.440645773979107, + "grad_norm": 41.99497604370117, + "learning_rate": 1.173518869913557e-05, + "loss": 0.1265, + "step": 4975 + }, + { + "epoch": 9.488129154795821, + "grad_norm": 12.936697959899902, + "learning_rate": 1.168247944339026e-05, + "loss": 0.1418, + "step": 5000 + }, + { + "epoch": 9.535612535612536, + "grad_norm": 12.372526168823242, + "learning_rate": 1.1629770187644952e-05, + "loss": 0.1522, + "step": 5025 + }, + { + "epoch": 9.583095916429249, + "grad_norm": 4.182272434234619, + "learning_rate": 1.1577060931899642e-05, + "loss": 0.1573, + "step": 5050 + }, + { + "epoch": 9.630579297245964, + "grad_norm": 10.066529273986816, + "learning_rate": 1.1524351676154333e-05, + "loss": 0.1371, + "step": 5075 + }, + { + "epoch": 9.678062678062679, + "grad_norm": 4.693755149841309, + "learning_rate": 1.1471642420409023e-05, + "loss": 0.1268, + "step": 5100 + }, + { + "epoch": 9.725546058879392, + "grad_norm": 9.667073249816895, + "learning_rate": 1.1418933164663717e-05, + "loss": 0.1328, + "step": 5125 + }, + { + "epoch": 9.773029439696106, + "grad_norm": 9.26330852508545, + "learning_rate": 1.1366223908918409e-05, + "loss": 0.1226, + "step": 5150 + }, + { + "epoch": 9.820512820512821, + "grad_norm": 15.474223136901855, + "learning_rate": 1.1313514653173099e-05, + "loss": 0.1275, + "step": 5175 + }, + { + "epoch": 9.867996201329534, + "grad_norm": 19.436092376708984, + "learning_rate": 1.126080539742779e-05, + "loss": 0.1285, + "step": 5200 + }, + { + "epoch": 9.915479582146249, + "grad_norm": 10.770730972290039, + "learning_rate": 1.120809614168248e-05, + "loss": 0.12, + "step": 5225 + }, + { + "epoch": 9.962962962962964, + "grad_norm": 7.493702411651611, + "learning_rate": 1.1155386885937172e-05, + "loss": 0.1158, + "step": 5250 + }, + { + "epoch": 10.0, + "eval_accuracy": 0.24045185583647122, + "eval_loss": 1.5478484630584717, + "eval_macro_f1": 0.5937281220361948, + "eval_micro_f1": 0.610705596107056, + "eval_runtime": 4.8236, + "eval_samples_per_second": 385.396, + "eval_steps_per_second": 24.256, + "eval_weighted_f1": 0.609383947645994, + "step": 5270 + }, + { + "epoch": 10.009496676163343, + "grad_norm": 3.4874448776245117, + "learning_rate": 1.1102677630191862e-05, + "loss": 0.1098, + "step": 5275 + }, + { + "epoch": 10.056980056980057, + "grad_norm": 5.664300441741943, + "learning_rate": 1.1049968374446554e-05, + "loss": 0.0934, + "step": 5300 + }, + { + "epoch": 10.10446343779677, + "grad_norm": 7.604928016662598, + "learning_rate": 1.0997259118701244e-05, + "loss": 0.1023, + "step": 5325 + }, + { + "epoch": 10.151946818613485, + "grad_norm": 10.689448356628418, + "learning_rate": 1.0944549862955936e-05, + "loss": 0.0998, + "step": 5350 + }, + { + "epoch": 10.1994301994302, + "grad_norm": 13.132899284362793, + "learning_rate": 1.0891840607210626e-05, + "loss": 0.1039, + "step": 5375 + }, + { + "epoch": 10.246913580246913, + "grad_norm": 12.537591934204102, + "learning_rate": 1.0839131351465318e-05, + "loss": 0.087, + "step": 5400 + }, + { + "epoch": 10.294396961063628, + "grad_norm": 3.18717360496521, + "learning_rate": 1.0786422095720008e-05, + "loss": 0.1003, + "step": 5425 + }, + { + "epoch": 10.341880341880342, + "grad_norm": 3.4345011711120605, + "learning_rate": 1.0733712839974701e-05, + "loss": 0.0851, + "step": 5450 + }, + { + "epoch": 10.389363722697055, + "grad_norm": 9.32591724395752, + "learning_rate": 1.0681003584229393e-05, + "loss": 0.0844, + "step": 5475 + }, + { + "epoch": 10.43684710351377, + "grad_norm": 6.039343357086182, + "learning_rate": 1.0628294328484083e-05, + "loss": 0.0802, + "step": 5500 + }, + { + "epoch": 10.484330484330485, + "grad_norm": 2.382885694503784, + "learning_rate": 1.0575585072738775e-05, + "loss": 0.0923, + "step": 5525 + }, + { + "epoch": 10.531813865147198, + "grad_norm": 16.99856948852539, + "learning_rate": 1.0522875816993465e-05, + "loss": 0.078, + "step": 5550 + }, + { + "epoch": 10.579297245963913, + "grad_norm": 9.307934761047363, + "learning_rate": 1.0470166561248156e-05, + "loss": 0.0776, + "step": 5575 + }, + { + "epoch": 10.626780626780628, + "grad_norm": 6.536963939666748, + "learning_rate": 1.0417457305502846e-05, + "loss": 0.0979, + "step": 5600 + }, + { + "epoch": 10.67426400759734, + "grad_norm": 10.15799331665039, + "learning_rate": 1.0364748049757538e-05, + "loss": 0.079, + "step": 5625 + }, + { + "epoch": 10.721747388414055, + "grad_norm": 4.942262172698975, + "learning_rate": 1.031203879401223e-05, + "loss": 0.087, + "step": 5650 + }, + { + "epoch": 10.76923076923077, + "grad_norm": 21.63833236694336, + "learning_rate": 1.025932953826692e-05, + "loss": 0.0847, + "step": 5675 + }, + { + "epoch": 10.816714150047483, + "grad_norm": 5.988709449768066, + "learning_rate": 1.0206620282521612e-05, + "loss": 0.0939, + "step": 5700 + }, + { + "epoch": 10.864197530864198, + "grad_norm": 5.350837707519531, + "learning_rate": 1.0153911026776302e-05, + "loss": 0.0755, + "step": 5725 + }, + { + "epoch": 10.91168091168091, + "grad_norm": 3.8369336128234863, + "learning_rate": 1.0101201771030993e-05, + "loss": 0.0735, + "step": 5750 + }, + { + "epoch": 10.959164292497626, + "grad_norm": 19.231178283691406, + "learning_rate": 1.0048492515285685e-05, + "loss": 0.0793, + "step": 5775 + }, + { + "epoch": 11.0, + "eval_accuracy": 0.24744486282947822, + "eval_loss": 1.6193442344665527, + "eval_macro_f1": 0.6045200976640341, + "eval_micro_f1": 0.619672574753078, + "eval_runtime": 5.0136, + "eval_samples_per_second": 370.788, + "eval_steps_per_second": 23.336, + "eval_weighted_f1": 0.6189624499875899, + "step": 5797 + } + ], + "logging_steps": 25, + "max_steps": 10540, + "num_input_tokens_seen": 0, + "num_train_epochs": 20, + "save_steps": 500, + "stateful_callbacks": { + "EarlyStoppingCallback": { + "args": { + "early_stopping_patience": 3, + "early_stopping_threshold": 0.0 + }, + "attributes": { + "early_stopping_patience_counter": 0 + } + }, + "TrainerControl": { + "args": { + "should_epoch_stop": false, + "should_evaluate": false, + "should_log": false, + "should_save": true, + "should_training_stop": false + }, + "attributes": {} + } + }, + "total_flos": 0.0, + "train_batch_size": 8, + "trial_name": null, + "trial_params": null +} diff --git a/ood-vigoemotions-mmbert-source-source-02-s43-20260905T133226Z-60628050/trainer/checkpoint-7378/trainer_state.json b/ood-vigoemotions-mmbert-source-source-02-s43-20260905T133226Z-60628050/trainer/checkpoint-7378/trainer_state.json new file mode 100644 index 0000000000000000000000000000000000000000..b6e5579e670462644d3920a9ceca404f9aa3dbab --- /dev/null +++ b/ood-vigoemotions-mmbert-source-source-02-s43-20260905T133226Z-60628050/trainer/checkpoint-7378/trainer_state.json @@ -0,0 +1,2276 @@ +{ + "best_global_step": 5797, + "best_metric": 0.6045200976640341, + "best_model_checkpoint": "/root/VNEncoder/results/checkpoints/ood-vigoemotions-mmbert-source-source-02-s43-20260905T133226Z-60628050/trainer/checkpoint-5797", + "epoch": 14.0, + "eval_steps": 500, + "global_step": 7378, + "is_hyper_param_search": false, + "is_local_process_zero": true, + "is_world_process_zero": true, + "log_history": [ + { + "epoch": 0.04748338081671415, + "grad_norm": 42.466888427734375, + "learning_rate": 4.554079696394687e-07, + "loss": 6.1416, + "step": 25 + }, + { + "epoch": 0.0949667616334283, + "grad_norm": 38.53072738647461, + "learning_rate": 9.297912713472485e-07, + "loss": 6.3527, + "step": 50 + }, + { + "epoch": 0.14245014245014245, + "grad_norm": 113.28205871582031, + "learning_rate": 1.4041745730550286e-06, + "loss": 6.0832, + "step": 75 + }, + { + "epoch": 0.1899335232668566, + "grad_norm": 73.18839263916016, + "learning_rate": 1.8785578747628086e-06, + "loss": 5.8318, + "step": 100 + }, + { + "epoch": 0.23741690408357075, + "grad_norm": 75.44960021972656, + "learning_rate": 2.3529411764705885e-06, + "loss": 5.7708, + "step": 125 + }, + { + "epoch": 0.2849002849002849, + "grad_norm": 61.81719207763672, + "learning_rate": 2.8273244781783683e-06, + "loss": 5.5055, + "step": 150 + }, + { + "epoch": 0.33238366571699907, + "grad_norm": 64.89173126220703, + "learning_rate": 3.3017077798861485e-06, + "loss": 5.5596, + "step": 175 + }, + { + "epoch": 0.3798670465337132, + "grad_norm": 35.145381927490234, + "learning_rate": 3.7760910815939284e-06, + "loss": 5.5006, + "step": 200 + }, + { + "epoch": 0.42735042735042733, + "grad_norm": 29.304231643676758, + "learning_rate": 4.250474383301708e-06, + "loss": 4.9728, + "step": 225 + }, + { + "epoch": 0.4748338081671415, + "grad_norm": 34.05004119873047, + "learning_rate": 4.724857685009488e-06, + "loss": 4.8357, + "step": 250 + }, + { + "epoch": 0.5223171889838556, + "grad_norm": 42.61710739135742, + "learning_rate": 5.199240986717268e-06, + "loss": 4.5948, + "step": 275 + }, + { + "epoch": 0.5698005698005698, + "grad_norm": 41.6121711730957, + "learning_rate": 5.673624288425048e-06, + "loss": 4.2142, + "step": 300 + }, + { + "epoch": 0.6172839506172839, + "grad_norm": 63.242279052734375, + "learning_rate": 6.1480075901328275e-06, + "loss": 4.0131, + "step": 325 + }, + { + "epoch": 0.6647673314339981, + "grad_norm": 61.35606002807617, + "learning_rate": 6.622390891840608e-06, + "loss": 3.9438, + "step": 350 + }, + { + "epoch": 0.7122507122507122, + "grad_norm": 64.1727523803711, + "learning_rate": 7.096774193548388e-06, + "loss": 3.8263, + "step": 375 + }, + { + "epoch": 0.7597340930674265, + "grad_norm": 48.41128158569336, + "learning_rate": 7.5711574952561674e-06, + "loss": 3.5959, + "step": 400 + }, + { + "epoch": 0.8072174738841406, + "grad_norm": 42.1629753112793, + "learning_rate": 8.045540796963948e-06, + "loss": 3.5873, + "step": 425 + }, + { + "epoch": 0.8547008547008547, + "grad_norm": 36.21388244628906, + "learning_rate": 8.519924098671729e-06, + "loss": 3.4469, + "step": 450 + }, + { + "epoch": 0.9021842355175689, + "grad_norm": 45.33504867553711, + "learning_rate": 8.994307400379507e-06, + "loss": 3.3685, + "step": 475 + }, + { + "epoch": 0.949667616334283, + "grad_norm": 53.4200439453125, + "learning_rate": 9.468690702087288e-06, + "loss": 3.2551, + "step": 500 + }, + { + "epoch": 0.9971509971509972, + "grad_norm": 52.55549621582031, + "learning_rate": 9.943074003795067e-06, + "loss": 3.2891, + "step": 525 + }, + { + "epoch": 1.0, + "eval_accuracy": 0.0016137708445400753, + "eval_loss": 0.7739421129226685, + "eval_macro_f1": 0.36155501356138686, + "eval_micro_f1": 0.37869281045751635, + "eval_runtime": 5.8168, + "eval_samples_per_second": 319.594, + "eval_steps_per_second": 20.114, + "eval_weighted_f1": 0.41253656843404873, + "step": 527 + }, + { + "epoch": 1.043684710351377, + "grad_norm": 38.479248046875, + "learning_rate": 1.0417457305502846e-05, + "loss": 2.8703, + "step": 550 + }, + { + "epoch": 1.0911680911680912, + "grad_norm": 58.546451568603516, + "learning_rate": 1.0891840607210626e-05, + "loss": 2.9037, + "step": 575 + }, + { + "epoch": 1.1386514719848053, + "grad_norm": 80.63526153564453, + "learning_rate": 1.1366223908918409e-05, + "loss": 2.8785, + "step": 600 + }, + { + "epoch": 1.1861348528015194, + "grad_norm": 59.79234313964844, + "learning_rate": 1.1840607210626186e-05, + "loss": 2.9307, + "step": 625 + }, + { + "epoch": 1.2336182336182335, + "grad_norm": 27.5277156829834, + "learning_rate": 1.2314990512333966e-05, + "loss": 2.9974, + "step": 650 + }, + { + "epoch": 1.2811016144349479, + "grad_norm": 47.518856048583984, + "learning_rate": 1.2789373814041749e-05, + "loss": 2.7249, + "step": 675 + }, + { + "epoch": 1.328584995251662, + "grad_norm": 49.7656364440918, + "learning_rate": 1.3263757115749526e-05, + "loss": 2.9259, + "step": 700 + }, + { + "epoch": 1.376068376068376, + "grad_norm": 54.58274841308594, + "learning_rate": 1.3738140417457306e-05, + "loss": 2.7116, + "step": 725 + }, + { + "epoch": 1.4235517568850902, + "grad_norm": 34.950828552246094, + "learning_rate": 1.4212523719165085e-05, + "loss": 2.6866, + "step": 750 + }, + { + "epoch": 1.4710351377018043, + "grad_norm": 52.59558868408203, + "learning_rate": 1.4686907020872866e-05, + "loss": 2.7492, + "step": 775 + }, + { + "epoch": 1.5185185185185186, + "grad_norm": 44.815189361572266, + "learning_rate": 1.5161290322580646e-05, + "loss": 2.7807, + "step": 800 + }, + { + "epoch": 1.5660018993352327, + "grad_norm": 50.69743728637695, + "learning_rate": 1.5635673624288425e-05, + "loss": 2.6976, + "step": 825 + }, + { + "epoch": 1.6134852801519468, + "grad_norm": 50.275367736816406, + "learning_rate": 1.6110056925996208e-05, + "loss": 2.7653, + "step": 850 + }, + { + "epoch": 1.660968660968661, + "grad_norm": 39.01410675048828, + "learning_rate": 1.6584440227703987e-05, + "loss": 2.6573, + "step": 875 + }, + { + "epoch": 1.708452041785375, + "grad_norm": 70.65179443359375, + "learning_rate": 1.7058823529411767e-05, + "loss": 2.9019, + "step": 900 + }, + { + "epoch": 1.7559354226020893, + "grad_norm": 35.23176193237305, + "learning_rate": 1.7533206831119546e-05, + "loss": 2.6363, + "step": 925 + }, + { + "epoch": 1.8034188034188035, + "grad_norm": 45.34317398071289, + "learning_rate": 1.8007590132827325e-05, + "loss": 2.7625, + "step": 950 + }, + { + "epoch": 1.8509021842355176, + "grad_norm": 37.30549240112305, + "learning_rate": 1.8481973434535105e-05, + "loss": 2.7392, + "step": 975 + }, + { + "epoch": 1.8983855650522319, + "grad_norm": 44.2552490234375, + "learning_rate": 1.8956356736242888e-05, + "loss": 2.7414, + "step": 1000 + }, + { + "epoch": 1.9458689458689458, + "grad_norm": 35.74271774291992, + "learning_rate": 1.9430740037950667e-05, + "loss": 2.692, + "step": 1025 + }, + { + "epoch": 1.99335232668566, + "grad_norm": 43.257347106933594, + "learning_rate": 1.9905123339658446e-05, + "loss": 2.6575, + "step": 1050 + }, + { + "epoch": 2.0, + "eval_accuracy": 0.016675632060247445, + "eval_loss": 0.6451638340950012, + "eval_macro_f1": 0.4216957664234933, + "eval_micro_f1": 0.43330427201394944, + "eval_runtime": 4.9871, + "eval_samples_per_second": 372.762, + "eval_steps_per_second": 23.461, + "eval_weighted_f1": 0.4744003458589279, + "step": 1054 + }, + { + "epoch": 2.03988603988604, + "grad_norm": 35.80771255493164, + "learning_rate": 1.9957832595403753e-05, + "loss": 2.0608, + "step": 1075 + }, + { + "epoch": 2.087369420702754, + "grad_norm": 31.177793502807617, + "learning_rate": 1.9905123339658446e-05, + "loss": 1.9354, + "step": 1100 + }, + { + "epoch": 2.134852801519468, + "grad_norm": 21.678997039794922, + "learning_rate": 1.9852414083913136e-05, + "loss": 2.0003, + "step": 1125 + }, + { + "epoch": 2.1823361823361824, + "grad_norm": 38.477752685546875, + "learning_rate": 1.9799704828167827e-05, + "loss": 1.9289, + "step": 1150 + }, + { + "epoch": 2.2298195631528963, + "grad_norm": 31.992578506469727, + "learning_rate": 1.9746995572422517e-05, + "loss": 1.9564, + "step": 1175 + }, + { + "epoch": 2.2773029439696106, + "grad_norm": 69.55426788330078, + "learning_rate": 1.969428631667721e-05, + "loss": 2.0287, + "step": 1200 + }, + { + "epoch": 2.324786324786325, + "grad_norm": 31.55048370361328, + "learning_rate": 1.9641577060931903e-05, + "loss": 2.1207, + "step": 1225 + }, + { + "epoch": 2.372269705603039, + "grad_norm": 51.29478454589844, + "learning_rate": 1.9588867805186593e-05, + "loss": 1.9785, + "step": 1250 + }, + { + "epoch": 2.419753086419753, + "grad_norm": 43.476219177246094, + "learning_rate": 1.9536158549441284e-05, + "loss": 2.0041, + "step": 1275 + }, + { + "epoch": 2.467236467236467, + "grad_norm": 30.011032104492188, + "learning_rate": 1.9483449293695974e-05, + "loss": 2.0289, + "step": 1300 + }, + { + "epoch": 2.5147198480531814, + "grad_norm": 34.02845001220703, + "learning_rate": 1.9430740037950667e-05, + "loss": 1.9928, + "step": 1325 + }, + { + "epoch": 2.5622032288698957, + "grad_norm": 39.73603057861328, + "learning_rate": 1.9378030782205357e-05, + "loss": 2.0899, + "step": 1350 + }, + { + "epoch": 2.6096866096866096, + "grad_norm": 42.63766098022461, + "learning_rate": 1.9325321526460047e-05, + "loss": 1.9751, + "step": 1375 + }, + { + "epoch": 2.657169990503324, + "grad_norm": 33.42068862915039, + "learning_rate": 1.927261227071474e-05, + "loss": 2.0424, + "step": 1400 + }, + { + "epoch": 2.7046533713200382, + "grad_norm": 68.12841796875, + "learning_rate": 1.921990301496943e-05, + "loss": 1.9421, + "step": 1425 + }, + { + "epoch": 2.752136752136752, + "grad_norm": 34.54705047607422, + "learning_rate": 1.916719375922412e-05, + "loss": 1.9273, + "step": 1450 + }, + { + "epoch": 2.799620132953466, + "grad_norm": 26.91336441040039, + "learning_rate": 1.911448450347881e-05, + "loss": 1.9383, + "step": 1475 + }, + { + "epoch": 2.8471035137701803, + "grad_norm": 60.04805374145508, + "learning_rate": 1.9061775247733504e-05, + "loss": 2.0682, + "step": 1500 + }, + { + "epoch": 2.8945868945868947, + "grad_norm": 23.979312896728516, + "learning_rate": 1.9009065991988194e-05, + "loss": 1.9271, + "step": 1525 + }, + { + "epoch": 2.9420702754036085, + "grad_norm": 32.707942962646484, + "learning_rate": 1.8956356736242888e-05, + "loss": 2.0019, + "step": 1550 + }, + { + "epoch": 2.989553656220323, + "grad_norm": 29.024099349975586, + "learning_rate": 1.8903647480497578e-05, + "loss": 1.9078, + "step": 1575 + }, + { + "epoch": 3.0, + "eval_accuracy": 0.04410973641742873, + "eval_loss": 0.640509843826294, + "eval_macro_f1": 0.474925227989289, + "eval_micro_f1": 0.48038184612895396, + "eval_runtime": 5.0213, + "eval_samples_per_second": 370.219, + "eval_steps_per_second": 23.301, + "eval_weighted_f1": 0.5151251185527903, + "step": 1581 + }, + { + "epoch": 3.0360873694207027, + "grad_norm": 24.103885650634766, + "learning_rate": 1.8850938224752268e-05, + "loss": 1.5437, + "step": 1600 + }, + { + "epoch": 3.083570750237417, + "grad_norm": 42.41041946411133, + "learning_rate": 1.879822896900696e-05, + "loss": 1.4517, + "step": 1625 + }, + { + "epoch": 3.131054131054131, + "grad_norm": 15.861024856567383, + "learning_rate": 1.874551971326165e-05, + "loss": 1.3673, + "step": 1650 + }, + { + "epoch": 3.178537511870845, + "grad_norm": 19.538061141967773, + "learning_rate": 1.869281045751634e-05, + "loss": 1.3837, + "step": 1675 + }, + { + "epoch": 3.2260208926875595, + "grad_norm": 46.26189041137695, + "learning_rate": 1.864010120177103e-05, + "loss": 1.3156, + "step": 1700 + }, + { + "epoch": 3.2735042735042734, + "grad_norm": 36.7100715637207, + "learning_rate": 1.8587391946025725e-05, + "loss": 1.3421, + "step": 1725 + }, + { + "epoch": 3.3209876543209877, + "grad_norm": 40.29243850708008, + "learning_rate": 1.8534682690280415e-05, + "loss": 1.3264, + "step": 1750 + }, + { + "epoch": 3.3684710351377016, + "grad_norm": 22.646486282348633, + "learning_rate": 1.8481973434535105e-05, + "loss": 1.355, + "step": 1775 + }, + { + "epoch": 3.415954415954416, + "grad_norm": 24.885631561279297, + "learning_rate": 1.8429264178789795e-05, + "loss": 1.332, + "step": 1800 + }, + { + "epoch": 3.4634377967711303, + "grad_norm": 24.071887969970703, + "learning_rate": 1.8376554923044488e-05, + "loss": 1.258, + "step": 1825 + }, + { + "epoch": 3.510921177587844, + "grad_norm": 30.55681037902832, + "learning_rate": 1.832384566729918e-05, + "loss": 1.2578, + "step": 1850 + }, + { + "epoch": 3.5584045584045585, + "grad_norm": 37.01355743408203, + "learning_rate": 1.8271136411553872e-05, + "loss": 1.3362, + "step": 1875 + }, + { + "epoch": 3.6058879392212724, + "grad_norm": 30.750619888305664, + "learning_rate": 1.8218427155808562e-05, + "loss": 1.329, + "step": 1900 + }, + { + "epoch": 3.6533713200379867, + "grad_norm": 31.818098068237305, + "learning_rate": 1.8165717900063252e-05, + "loss": 1.3591, + "step": 1925 + }, + { + "epoch": 3.700854700854701, + "grad_norm": 26.725540161132812, + "learning_rate": 1.8113008644317945e-05, + "loss": 1.3032, + "step": 1950 + }, + { + "epoch": 3.748338081671415, + "grad_norm": 35.89231872558594, + "learning_rate": 1.8060299388572635e-05, + "loss": 1.3604, + "step": 1975 + }, + { + "epoch": 3.7958214624881292, + "grad_norm": 23.358871459960938, + "learning_rate": 1.8007590132827325e-05, + "loss": 1.2997, + "step": 2000 + }, + { + "epoch": 3.8433048433048436, + "grad_norm": 33.60404586791992, + "learning_rate": 1.7954880877082015e-05, + "loss": 1.2872, + "step": 2025 + }, + { + "epoch": 3.8907882241215574, + "grad_norm": 43.04498291015625, + "learning_rate": 1.790217162133671e-05, + "loss": 1.372, + "step": 2050 + }, + { + "epoch": 3.9382716049382713, + "grad_norm": 34.92791748046875, + "learning_rate": 1.78494623655914e-05, + "loss": 1.3739, + "step": 2075 + }, + { + "epoch": 3.9857549857549857, + "grad_norm": 61.5059814453125, + "learning_rate": 1.779675310984609e-05, + "loss": 1.2661, + "step": 2100 + }, + { + "epoch": 4.0, + "eval_accuracy": 0.07961269499731038, + "eval_loss": 0.7032579183578491, + "eval_macro_f1": 0.5178092021035117, + "eval_micro_f1": 0.5282948622486969, + "eval_runtime": 4.9726, + "eval_samples_per_second": 373.847, + "eval_steps_per_second": 23.529, + "eval_weighted_f1": 0.5480078641240479, + "step": 2108 + }, + { + "epoch": 4.032288698955366, + "grad_norm": 30.800107955932617, + "learning_rate": 1.7744043854100782e-05, + "loss": 1.0502, + "step": 2125 + }, + { + "epoch": 4.07977207977208, + "grad_norm": 25.957460403442383, + "learning_rate": 1.7691334598355472e-05, + "loss": 0.9406, + "step": 2150 + }, + { + "epoch": 4.127255460588794, + "grad_norm": 18.835384368896484, + "learning_rate": 1.7638625342610163e-05, + "loss": 0.909, + "step": 2175 + }, + { + "epoch": 4.174738841405508, + "grad_norm": 34.36490249633789, + "learning_rate": 1.7585916086864856e-05, + "loss": 0.8639, + "step": 2200 + }, + { + "epoch": 4.222222222222222, + "grad_norm": 42.17921829223633, + "learning_rate": 1.7533206831119546e-05, + "loss": 0.986, + "step": 2225 + }, + { + "epoch": 4.269705603038936, + "grad_norm": 18.649545669555664, + "learning_rate": 1.748049757537424e-05, + "loss": 0.9287, + "step": 2250 + }, + { + "epoch": 4.317188983855651, + "grad_norm": 48.732879638671875, + "learning_rate": 1.742778831962893e-05, + "loss": 0.9057, + "step": 2275 + }, + { + "epoch": 4.364672364672365, + "grad_norm": 24.9398193359375, + "learning_rate": 1.737507906388362e-05, + "loss": 0.9474, + "step": 2300 + }, + { + "epoch": 4.412155745489079, + "grad_norm": 11.954604148864746, + "learning_rate": 1.732236980813831e-05, + "loss": 0.9036, + "step": 2325 + }, + { + "epoch": 4.459639126305793, + "grad_norm": 18.34949493408203, + "learning_rate": 1.7269660552393003e-05, + "loss": 0.927, + "step": 2350 + }, + { + "epoch": 4.507122507122507, + "grad_norm": 78.10287475585938, + "learning_rate": 1.7216951296647693e-05, + "loss": 0.9124, + "step": 2375 + }, + { + "epoch": 4.554605887939221, + "grad_norm": 29.533727645874023, + "learning_rate": 1.7164242040902383e-05, + "loss": 0.9532, + "step": 2400 + }, + { + "epoch": 4.602089268755935, + "grad_norm": 20.754230499267578, + "learning_rate": 1.7111532785157073e-05, + "loss": 0.9455, + "step": 2425 + }, + { + "epoch": 4.64957264957265, + "grad_norm": 12.572599411010742, + "learning_rate": 1.7058823529411767e-05, + "loss": 0.9059, + "step": 2450 + }, + { + "epoch": 4.697056030389364, + "grad_norm": 25.214078903198242, + "learning_rate": 1.7006114273666457e-05, + "loss": 0.9483, + "step": 2475 + }, + { + "epoch": 4.744539411206078, + "grad_norm": 28.17908477783203, + "learning_rate": 1.6953405017921147e-05, + "loss": 0.8608, + "step": 2500 + }, + { + "epoch": 4.792022792022792, + "grad_norm": 27.811349868774414, + "learning_rate": 1.690069576217584e-05, + "loss": 0.8963, + "step": 2525 + }, + { + "epoch": 4.839506172839506, + "grad_norm": 34.6890869140625, + "learning_rate": 1.684798650643053e-05, + "loss": 0.9537, + "step": 2550 + }, + { + "epoch": 4.88698955365622, + "grad_norm": 41.03017807006836, + "learning_rate": 1.6795277250685224e-05, + "loss": 0.9382, + "step": 2575 + }, + { + "epoch": 4.934472934472934, + "grad_norm": 20.774999618530273, + "learning_rate": 1.6742567994939914e-05, + "loss": 0.9576, + "step": 2600 + }, + { + "epoch": 4.981956315289649, + "grad_norm": 50.060672760009766, + "learning_rate": 1.6689858739194604e-05, + "loss": 0.8801, + "step": 2625 + }, + { + "epoch": 5.0, + "eval_accuracy": 0.1312533620225928, + "eval_loss": 0.8617390990257263, + "eval_macro_f1": 0.5661674742863365, + "eval_micro_f1": 0.5719966513185433, + "eval_runtime": 4.8508, + "eval_samples_per_second": 383.236, + "eval_steps_per_second": 24.12, + "eval_weighted_f1": 0.5831312572156591, + "step": 2635 + }, + { + "epoch": 5.028490028490029, + "grad_norm": 18.490642547607422, + "learning_rate": 1.6637149483449294e-05, + "loss": 0.7098, + "step": 2650 + }, + { + "epoch": 5.075973409306743, + "grad_norm": 15.921799659729004, + "learning_rate": 1.6584440227703987e-05, + "loss": 0.6619, + "step": 2675 + }, + { + "epoch": 5.1234567901234565, + "grad_norm": 14.693387031555176, + "learning_rate": 1.6531730971958677e-05, + "loss": 0.6622, + "step": 2700 + }, + { + "epoch": 5.170940170940171, + "grad_norm": 13.369607925415039, + "learning_rate": 1.6479021716213367e-05, + "loss": 0.6234, + "step": 2725 + }, + { + "epoch": 5.218423551756885, + "grad_norm": 11.756754875183105, + "learning_rate": 1.642631246046806e-05, + "loss": 0.6504, + "step": 2750 + }, + { + "epoch": 5.265906932573599, + "grad_norm": 32.03675842285156, + "learning_rate": 1.637360320472275e-05, + "loss": 0.6522, + "step": 2775 + }, + { + "epoch": 5.313390313390314, + "grad_norm": 71.17816925048828, + "learning_rate": 1.632089394897744e-05, + "loss": 0.6194, + "step": 2800 + }, + { + "epoch": 5.360873694207028, + "grad_norm": 15.817419052124023, + "learning_rate": 1.626818469323213e-05, + "loss": 0.6312, + "step": 2825 + }, + { + "epoch": 5.4083570750237415, + "grad_norm": 25.708629608154297, + "learning_rate": 1.6215475437486824e-05, + "loss": 0.632, + "step": 2850 + }, + { + "epoch": 5.455840455840455, + "grad_norm": 14.698016166687012, + "learning_rate": 1.6162766181741514e-05, + "loss": 0.6342, + "step": 2875 + }, + { + "epoch": 5.50332383665717, + "grad_norm": 31.988943099975586, + "learning_rate": 1.6110056925996208e-05, + "loss": 0.6859, + "step": 2900 + }, + { + "epoch": 5.550807217473884, + "grad_norm": 14.286497116088867, + "learning_rate": 1.6057347670250898e-05, + "loss": 0.6473, + "step": 2925 + }, + { + "epoch": 5.598290598290598, + "grad_norm": 41.547271728515625, + "learning_rate": 1.6004638414505588e-05, + "loss": 0.6853, + "step": 2950 + }, + { + "epoch": 5.645773979107313, + "grad_norm": 17.2784366607666, + "learning_rate": 1.595192915876028e-05, + "loss": 0.659, + "step": 2975 + }, + { + "epoch": 5.693257359924027, + "grad_norm": 60.431392669677734, + "learning_rate": 1.589921990301497e-05, + "loss": 0.5916, + "step": 3000 + }, + { + "epoch": 5.7407407407407405, + "grad_norm": 27.35256576538086, + "learning_rate": 1.584651064726966e-05, + "loss": 0.6493, + "step": 3025 + }, + { + "epoch": 5.788224121557455, + "grad_norm": 20.633012771606445, + "learning_rate": 1.579380139152435e-05, + "loss": 0.5941, + "step": 3050 + }, + { + "epoch": 5.835707502374169, + "grad_norm": 15.044947624206543, + "learning_rate": 1.5741092135779045e-05, + "loss": 0.6362, + "step": 3075 + }, + { + "epoch": 5.883190883190883, + "grad_norm": 14.263772010803223, + "learning_rate": 1.5688382880033735e-05, + "loss": 0.6214, + "step": 3100 + }, + { + "epoch": 5.930674264007598, + "grad_norm": 34.03923797607422, + "learning_rate": 1.5635673624288425e-05, + "loss": 0.615, + "step": 3125 + }, + { + "epoch": 5.978157644824312, + "grad_norm": 55.2652473449707, + "learning_rate": 1.5582964368543115e-05, + "loss": 0.6469, + "step": 3150 + }, + { + "epoch": 6.0, + "eval_accuracy": 0.17428725121032812, + "eval_loss": 0.930029571056366, + "eval_macro_f1": 0.5744715338078124, + "eval_micro_f1": 0.585635359116022, + "eval_runtime": 4.866, + "eval_samples_per_second": 382.041, + "eval_steps_per_second": 24.045, + "eval_weighted_f1": 0.5939596407496641, + "step": 3162 + }, + { + "epoch": 6.0246913580246915, + "grad_norm": 17.741565704345703, + "learning_rate": 1.553025511279781e-05, + "loss": 0.5214, + "step": 3175 + }, + { + "epoch": 6.072174738841405, + "grad_norm": 22.976102828979492, + "learning_rate": 1.5477545857052502e-05, + "loss": 0.4418, + "step": 3200 + }, + { + "epoch": 6.119658119658119, + "grad_norm": 20.000131607055664, + "learning_rate": 1.5424836601307192e-05, + "loss": 0.4645, + "step": 3225 + }, + { + "epoch": 6.167141500474834, + "grad_norm": 12.409162521362305, + "learning_rate": 1.5372127345561882e-05, + "loss": 0.4078, + "step": 3250 + }, + { + "epoch": 6.214624881291548, + "grad_norm": 11.717437744140625, + "learning_rate": 1.5319418089816572e-05, + "loss": 0.4454, + "step": 3275 + }, + { + "epoch": 6.262108262108262, + "grad_norm": 19.861845016479492, + "learning_rate": 1.5266708834071265e-05, + "loss": 0.4444, + "step": 3300 + }, + { + "epoch": 6.3095916429249765, + "grad_norm": 23.135236740112305, + "learning_rate": 1.5213999578325956e-05, + "loss": 0.3956, + "step": 3325 + }, + { + "epoch": 6.35707502374169, + "grad_norm": 28.478940963745117, + "learning_rate": 1.5161290322580646e-05, + "loss": 0.4448, + "step": 3350 + }, + { + "epoch": 6.404558404558404, + "grad_norm": 17.0195255279541, + "learning_rate": 1.5108581066835337e-05, + "loss": 0.4401, + "step": 3375 + }, + { + "epoch": 6.452041785375119, + "grad_norm": 12.3309965133667, + "learning_rate": 1.5055871811090027e-05, + "loss": 0.4316, + "step": 3400 + }, + { + "epoch": 6.499525166191833, + "grad_norm": 11.332566261291504, + "learning_rate": 1.5003162555344719e-05, + "loss": 0.4079, + "step": 3425 + }, + { + "epoch": 6.547008547008547, + "grad_norm": 26.008562088012695, + "learning_rate": 1.4950453299599411e-05, + "loss": 0.4256, + "step": 3450 + }, + { + "epoch": 6.594491927825262, + "grad_norm": 10.497772216796875, + "learning_rate": 1.4897744043854101e-05, + "loss": 0.4626, + "step": 3475 + }, + { + "epoch": 6.6419753086419755, + "grad_norm": 13.758752822875977, + "learning_rate": 1.4845034788108794e-05, + "loss": 0.4252, + "step": 3500 + }, + { + "epoch": 6.689458689458689, + "grad_norm": 27.361492156982422, + "learning_rate": 1.4792325532363484e-05, + "loss": 0.4509, + "step": 3525 + }, + { + "epoch": 6.736942070275403, + "grad_norm": 14.832262992858887, + "learning_rate": 1.4739616276618176e-05, + "loss": 0.4464, + "step": 3550 + }, + { + "epoch": 6.784425451092118, + "grad_norm": 33.7310791015625, + "learning_rate": 1.4686907020872866e-05, + "loss": 0.4238, + "step": 3575 + }, + { + "epoch": 6.831908831908832, + "grad_norm": 29.504594802856445, + "learning_rate": 1.4634197765127558e-05, + "loss": 0.4699, + "step": 3600 + }, + { + "epoch": 6.879392212725546, + "grad_norm": 42.16069412231445, + "learning_rate": 1.458148850938225e-05, + "loss": 0.4402, + "step": 3625 + }, + { + "epoch": 6.926875593542261, + "grad_norm": 43.56899642944336, + "learning_rate": 1.452877925363694e-05, + "loss": 0.4778, + "step": 3650 + }, + { + "epoch": 6.9743589743589745, + "grad_norm": 13.3931245803833, + "learning_rate": 1.4476069997891631e-05, + "loss": 0.4515, + "step": 3675 + }, + { + "epoch": 7.0, + "eval_accuracy": 0.2087143625605164, + "eval_loss": 1.1090645790100098, + "eval_macro_f1": 0.585298796573478, + "eval_micro_f1": 0.5964482547458665, + "eval_runtime": 4.8885, + "eval_samples_per_second": 380.277, + "eval_steps_per_second": 23.933, + "eval_weighted_f1": 0.6004246347046586, + "step": 3689 + }, + { + "epoch": 7.020892687559354, + "grad_norm": 15.4351224899292, + "learning_rate": 1.4423360742146321e-05, + "loss": 0.3949, + "step": 3700 + }, + { + "epoch": 7.068376068376068, + "grad_norm": 11.938258171081543, + "learning_rate": 1.4370651486401013e-05, + "loss": 0.2897, + "step": 3725 + }, + { + "epoch": 7.115859449192783, + "grad_norm": 19.71200180053711, + "learning_rate": 1.4317942230655703e-05, + "loss": 0.2975, + "step": 3750 + }, + { + "epoch": 7.163342830009497, + "grad_norm": 13.208940505981445, + "learning_rate": 1.4265232974910395e-05, + "loss": 0.3104, + "step": 3775 + }, + { + "epoch": 7.210826210826211, + "grad_norm": 8.498453140258789, + "learning_rate": 1.4212523719165085e-05, + "loss": 0.2945, + "step": 3800 + }, + { + "epoch": 7.258309591642925, + "grad_norm": 9.174602508544922, + "learning_rate": 1.4159814463419779e-05, + "loss": 0.3014, + "step": 3825 + }, + { + "epoch": 7.305792972459639, + "grad_norm": 24.003524780273438, + "learning_rate": 1.410710520767447e-05, + "loss": 0.3163, + "step": 3850 + }, + { + "epoch": 7.353276353276353, + "grad_norm": 19.13205909729004, + "learning_rate": 1.405439595192916e-05, + "loss": 0.3171, + "step": 3875 + }, + { + "epoch": 7.400759734093067, + "grad_norm": 15.435404777526855, + "learning_rate": 1.4001686696183852e-05, + "loss": 0.2934, + "step": 3900 + }, + { + "epoch": 7.448243114909782, + "grad_norm": 10.802959442138672, + "learning_rate": 1.3948977440438542e-05, + "loss": 0.2699, + "step": 3925 + }, + { + "epoch": 7.495726495726496, + "grad_norm": 6.724036693572998, + "learning_rate": 1.3896268184693234e-05, + "loss": 0.2723, + "step": 3950 + }, + { + "epoch": 7.54320987654321, + "grad_norm": 8.457695960998535, + "learning_rate": 1.3843558928947924e-05, + "loss": 0.2864, + "step": 3975 + }, + { + "epoch": 7.590693257359924, + "grad_norm": 59.64891052246094, + "learning_rate": 1.3790849673202616e-05, + "loss": 0.3078, + "step": 4000 + }, + { + "epoch": 7.638176638176638, + "grad_norm": 13.899282455444336, + "learning_rate": 1.3738140417457306e-05, + "loss": 0.2878, + "step": 4025 + }, + { + "epoch": 7.685660018993352, + "grad_norm": 22.03400421142578, + "learning_rate": 1.3685431161711997e-05, + "loss": 0.3049, + "step": 4050 + }, + { + "epoch": 7.733143399810066, + "grad_norm": 10.780784606933594, + "learning_rate": 1.3632721905966687e-05, + "loss": 0.2804, + "step": 4075 + }, + { + "epoch": 7.780626780626781, + "grad_norm": 40.84187698364258, + "learning_rate": 1.358001265022138e-05, + "loss": 0.3037, + "step": 4100 + }, + { + "epoch": 7.828110161443495, + "grad_norm": 16.029813766479492, + "learning_rate": 1.3527303394476071e-05, + "loss": 0.3138, + "step": 4125 + }, + { + "epoch": 7.875593542260209, + "grad_norm": 17.134769439697266, + "learning_rate": 1.3474594138730763e-05, + "loss": 0.3241, + "step": 4150 + }, + { + "epoch": 7.923076923076923, + "grad_norm": 32.032840728759766, + "learning_rate": 1.3421884882985454e-05, + "loss": 0.294, + "step": 4175 + }, + { + "epoch": 7.970560303893637, + "grad_norm": 21.15312385559082, + "learning_rate": 1.3369175627240144e-05, + "loss": 0.2793, + "step": 4200 + }, + { + "epoch": 8.0, + "eval_accuracy": 0.2140935987089833, + "eval_loss": 1.2231533527374268, + "eval_macro_f1": 0.5940205209972464, + "eval_micro_f1": 0.6054839113559535, + "eval_runtime": 4.9611, + "eval_samples_per_second": 374.717, + "eval_steps_per_second": 23.584, + "eval_weighted_f1": 0.6057940584072086, + "step": 4216 + }, + { + "epoch": 8.017094017094017, + "grad_norm": 10.781094551086426, + "learning_rate": 1.3316466371494836e-05, + "loss": 0.2485, + "step": 4225 + }, + { + "epoch": 8.064577397910732, + "grad_norm": 10.791547775268555, + "learning_rate": 1.3263757115749526e-05, + "loss": 0.2238, + "step": 4250 + }, + { + "epoch": 8.112060778727445, + "grad_norm": 19.549699783325195, + "learning_rate": 1.3211047860004218e-05, + "loss": 0.2224, + "step": 4275 + }, + { + "epoch": 8.15954415954416, + "grad_norm": 12.872920036315918, + "learning_rate": 1.315833860425891e-05, + "loss": 0.2067, + "step": 4300 + }, + { + "epoch": 8.207027540360874, + "grad_norm": 7.634912490844727, + "learning_rate": 1.31056293485136e-05, + "loss": 0.2086, + "step": 4325 + }, + { + "epoch": 8.254510921177587, + "grad_norm": 28.830102920532227, + "learning_rate": 1.3052920092768292e-05, + "loss": 0.2017, + "step": 4350 + }, + { + "epoch": 8.301994301994302, + "grad_norm": 5.609751224517822, + "learning_rate": 1.3000210837022982e-05, + "loss": 0.1863, + "step": 4375 + }, + { + "epoch": 8.349477682811017, + "grad_norm": 15.390376091003418, + "learning_rate": 1.2947501581277673e-05, + "loss": 0.1978, + "step": 4400 + }, + { + "epoch": 8.39696106362773, + "grad_norm": 7.242794513702393, + "learning_rate": 1.2894792325532363e-05, + "loss": 0.2013, + "step": 4425 + }, + { + "epoch": 8.444444444444445, + "grad_norm": 33.799224853515625, + "learning_rate": 1.2842083069787055e-05, + "loss": 0.2005, + "step": 4450 + }, + { + "epoch": 8.49192782526116, + "grad_norm": 7.0647358894348145, + "learning_rate": 1.2789373814041749e-05, + "loss": 0.2187, + "step": 4475 + }, + { + "epoch": 8.539411206077872, + "grad_norm": 12.728279113769531, + "learning_rate": 1.2736664558296439e-05, + "loss": 0.2135, + "step": 4500 + }, + { + "epoch": 8.586894586894587, + "grad_norm": 7.330479621887207, + "learning_rate": 1.268395530255113e-05, + "loss": 0.208, + "step": 4525 + }, + { + "epoch": 8.634377967711302, + "grad_norm": 7.347399711608887, + "learning_rate": 1.263124604680582e-05, + "loss": 0.1834, + "step": 4550 + }, + { + "epoch": 8.681861348528015, + "grad_norm": 19.936182022094727, + "learning_rate": 1.2578536791060512e-05, + "loss": 0.1962, + "step": 4575 + }, + { + "epoch": 8.72934472934473, + "grad_norm": 12.831972122192383, + "learning_rate": 1.2525827535315202e-05, + "loss": 0.1847, + "step": 4600 + }, + { + "epoch": 8.776828110161443, + "grad_norm": 6.2358903884887695, + "learning_rate": 1.2473118279569894e-05, + "loss": 0.1993, + "step": 4625 + }, + { + "epoch": 8.824311490978157, + "grad_norm": 9.986373901367188, + "learning_rate": 1.2420409023824584e-05, + "loss": 0.2021, + "step": 4650 + }, + { + "epoch": 8.871794871794872, + "grad_norm": 15.168353080749512, + "learning_rate": 1.2367699768079276e-05, + "loss": 0.2047, + "step": 4675 + }, + { + "epoch": 8.919278252611585, + "grad_norm": 18.249940872192383, + "learning_rate": 1.2314990512333966e-05, + "loss": 0.1867, + "step": 4700 + }, + { + "epoch": 8.9667616334283, + "grad_norm": 8.6532564163208, + "learning_rate": 1.2262281256588657e-05, + "loss": 0.1818, + "step": 4725 + }, + { + "epoch": 9.0, + "eval_accuracy": 0.23830016137708446, + "eval_loss": 1.4221587181091309, + "eval_macro_f1": 0.5937747690570366, + "eval_micro_f1": 0.6078904991948471, + "eval_runtime": 5.0553, + "eval_samples_per_second": 367.734, + "eval_steps_per_second": 23.144, + "eval_weighted_f1": 0.6074643571973501, + "step": 4743 + }, + { + "epoch": 9.01329534662868, + "grad_norm": 7.653168201446533, + "learning_rate": 1.2209572000843348e-05, + "loss": 0.1608, + "step": 4750 + }, + { + "epoch": 9.060778727445394, + "grad_norm": 25.898069381713867, + "learning_rate": 1.215686274509804e-05, + "loss": 0.1589, + "step": 4775 + }, + { + "epoch": 9.108262108262108, + "grad_norm": 7.836034297943115, + "learning_rate": 1.2104153489352733e-05, + "loss": 0.1556, + "step": 4800 + }, + { + "epoch": 9.155745489078823, + "grad_norm": 8.399602890014648, + "learning_rate": 1.2051444233607423e-05, + "loss": 0.1274, + "step": 4825 + }, + { + "epoch": 9.203228869895536, + "grad_norm": 15.443283081054688, + "learning_rate": 1.1998734977862115e-05, + "loss": 0.136, + "step": 4850 + }, + { + "epoch": 9.250712250712251, + "grad_norm": 25.787002563476562, + "learning_rate": 1.1946025722116805e-05, + "loss": 0.1332, + "step": 4875 + }, + { + "epoch": 9.298195631528966, + "grad_norm": 9.80220890045166, + "learning_rate": 1.1893316466371496e-05, + "loss": 0.1303, + "step": 4900 + }, + { + "epoch": 9.345679012345679, + "grad_norm": 8.279167175292969, + "learning_rate": 1.1840607210626186e-05, + "loss": 0.1354, + "step": 4925 + }, + { + "epoch": 9.393162393162394, + "grad_norm": 9.350815773010254, + "learning_rate": 1.1787897954880878e-05, + "loss": 0.1297, + "step": 4950 + }, + { + "epoch": 9.440645773979107, + "grad_norm": 41.99497604370117, + "learning_rate": 1.173518869913557e-05, + "loss": 0.1265, + "step": 4975 + }, + { + "epoch": 9.488129154795821, + "grad_norm": 12.936697959899902, + "learning_rate": 1.168247944339026e-05, + "loss": 0.1418, + "step": 5000 + }, + { + "epoch": 9.535612535612536, + "grad_norm": 12.372526168823242, + "learning_rate": 1.1629770187644952e-05, + "loss": 0.1522, + "step": 5025 + }, + { + "epoch": 9.583095916429249, + "grad_norm": 4.182272434234619, + "learning_rate": 1.1577060931899642e-05, + "loss": 0.1573, + "step": 5050 + }, + { + "epoch": 9.630579297245964, + "grad_norm": 10.066529273986816, + "learning_rate": 1.1524351676154333e-05, + "loss": 0.1371, + "step": 5075 + }, + { + "epoch": 9.678062678062679, + "grad_norm": 4.693755149841309, + "learning_rate": 1.1471642420409023e-05, + "loss": 0.1268, + "step": 5100 + }, + { + "epoch": 9.725546058879392, + "grad_norm": 9.667073249816895, + "learning_rate": 1.1418933164663717e-05, + "loss": 0.1328, + "step": 5125 + }, + { + "epoch": 9.773029439696106, + "grad_norm": 9.26330852508545, + "learning_rate": 1.1366223908918409e-05, + "loss": 0.1226, + "step": 5150 + }, + { + "epoch": 9.820512820512821, + "grad_norm": 15.474223136901855, + "learning_rate": 1.1313514653173099e-05, + "loss": 0.1275, + "step": 5175 + }, + { + "epoch": 9.867996201329534, + "grad_norm": 19.436092376708984, + "learning_rate": 1.126080539742779e-05, + "loss": 0.1285, + "step": 5200 + }, + { + "epoch": 9.915479582146249, + "grad_norm": 10.770730972290039, + "learning_rate": 1.120809614168248e-05, + "loss": 0.12, + "step": 5225 + }, + { + "epoch": 9.962962962962964, + "grad_norm": 7.493702411651611, + "learning_rate": 1.1155386885937172e-05, + "loss": 0.1158, + "step": 5250 + }, + { + "epoch": 10.0, + "eval_accuracy": 0.24045185583647122, + "eval_loss": 1.5478484630584717, + "eval_macro_f1": 0.5937281220361948, + "eval_micro_f1": 0.610705596107056, + "eval_runtime": 4.8236, + "eval_samples_per_second": 385.396, + "eval_steps_per_second": 24.256, + "eval_weighted_f1": 0.609383947645994, + "step": 5270 + }, + { + "epoch": 10.009496676163343, + "grad_norm": 3.4874448776245117, + "learning_rate": 1.1102677630191862e-05, + "loss": 0.1098, + "step": 5275 + }, + { + "epoch": 10.056980056980057, + "grad_norm": 5.664300441741943, + "learning_rate": 1.1049968374446554e-05, + "loss": 0.0934, + "step": 5300 + }, + { + "epoch": 10.10446343779677, + "grad_norm": 7.604928016662598, + "learning_rate": 1.0997259118701244e-05, + "loss": 0.1023, + "step": 5325 + }, + { + "epoch": 10.151946818613485, + "grad_norm": 10.689448356628418, + "learning_rate": 1.0944549862955936e-05, + "loss": 0.0998, + "step": 5350 + }, + { + "epoch": 10.1994301994302, + "grad_norm": 13.132899284362793, + "learning_rate": 1.0891840607210626e-05, + "loss": 0.1039, + "step": 5375 + }, + { + "epoch": 10.246913580246913, + "grad_norm": 12.537591934204102, + "learning_rate": 1.0839131351465318e-05, + "loss": 0.087, + "step": 5400 + }, + { + "epoch": 10.294396961063628, + "grad_norm": 3.18717360496521, + "learning_rate": 1.0786422095720008e-05, + "loss": 0.1003, + "step": 5425 + }, + { + "epoch": 10.341880341880342, + "grad_norm": 3.4345011711120605, + "learning_rate": 1.0733712839974701e-05, + "loss": 0.0851, + "step": 5450 + }, + { + "epoch": 10.389363722697055, + "grad_norm": 9.32591724395752, + "learning_rate": 1.0681003584229393e-05, + "loss": 0.0844, + "step": 5475 + }, + { + "epoch": 10.43684710351377, + "grad_norm": 6.039343357086182, + "learning_rate": 1.0628294328484083e-05, + "loss": 0.0802, + "step": 5500 + }, + { + "epoch": 10.484330484330485, + "grad_norm": 2.382885694503784, + "learning_rate": 1.0575585072738775e-05, + "loss": 0.0923, + "step": 5525 + }, + { + "epoch": 10.531813865147198, + "grad_norm": 16.99856948852539, + "learning_rate": 1.0522875816993465e-05, + "loss": 0.078, + "step": 5550 + }, + { + "epoch": 10.579297245963913, + "grad_norm": 9.307934761047363, + "learning_rate": 1.0470166561248156e-05, + "loss": 0.0776, + "step": 5575 + }, + { + "epoch": 10.626780626780628, + "grad_norm": 6.536963939666748, + "learning_rate": 1.0417457305502846e-05, + "loss": 0.0979, + "step": 5600 + }, + { + "epoch": 10.67426400759734, + "grad_norm": 10.15799331665039, + "learning_rate": 1.0364748049757538e-05, + "loss": 0.079, + "step": 5625 + }, + { + "epoch": 10.721747388414055, + "grad_norm": 4.942262172698975, + "learning_rate": 1.031203879401223e-05, + "loss": 0.087, + "step": 5650 + }, + { + "epoch": 10.76923076923077, + "grad_norm": 21.63833236694336, + "learning_rate": 1.025932953826692e-05, + "loss": 0.0847, + "step": 5675 + }, + { + "epoch": 10.816714150047483, + "grad_norm": 5.988709449768066, + "learning_rate": 1.0206620282521612e-05, + "loss": 0.0939, + "step": 5700 + }, + { + "epoch": 10.864197530864198, + "grad_norm": 5.350837707519531, + "learning_rate": 1.0153911026776302e-05, + "loss": 0.0755, + "step": 5725 + }, + { + "epoch": 10.91168091168091, + "grad_norm": 3.8369336128234863, + "learning_rate": 1.0101201771030993e-05, + "loss": 0.0735, + "step": 5750 + }, + { + "epoch": 10.959164292497626, + "grad_norm": 19.231178283691406, + "learning_rate": 1.0048492515285685e-05, + "loss": 0.0793, + "step": 5775 + }, + { + "epoch": 11.0, + "eval_accuracy": 0.24744486282947822, + "eval_loss": 1.6193442344665527, + "eval_macro_f1": 0.6045200976640341, + "eval_micro_f1": 0.619672574753078, + "eval_runtime": 5.0136, + "eval_samples_per_second": 370.788, + "eval_steps_per_second": 23.336, + "eval_weighted_f1": 0.6189624499875899, + "step": 5797 + }, + { + "epoch": 11.005698005698006, + "grad_norm": 6.854330062866211, + "learning_rate": 9.995783259540375e-06, + "loss": 0.0649, + "step": 5800 + }, + { + "epoch": 11.05318138651472, + "grad_norm": 6.888882160186768, + "learning_rate": 9.943074003795067e-06, + "loss": 0.0741, + "step": 5825 + }, + { + "epoch": 11.100664767331434, + "grad_norm": 4.7677106857299805, + "learning_rate": 9.890364748049759e-06, + "loss": 0.0688, + "step": 5850 + }, + { + "epoch": 11.148148148148149, + "grad_norm": 2.838637590408325, + "learning_rate": 9.83765549230445e-06, + "loss": 0.06, + "step": 5875 + }, + { + "epoch": 11.195631528964862, + "grad_norm": 20.290132522583008, + "learning_rate": 9.78494623655914e-06, + "loss": 0.0639, + "step": 5900 + }, + { + "epoch": 11.243114909781577, + "grad_norm": 2.2312605381011963, + "learning_rate": 9.732236980813832e-06, + "loss": 0.054, + "step": 5925 + }, + { + "epoch": 11.290598290598291, + "grad_norm": 3.7514896392822266, + "learning_rate": 9.679527725068522e-06, + "loss": 0.0538, + "step": 5950 + }, + { + "epoch": 11.338081671415004, + "grad_norm": 6.979944705963135, + "learning_rate": 9.626818469323214e-06, + "loss": 0.0776, + "step": 5975 + }, + { + "epoch": 11.385565052231719, + "grad_norm": 4.903264045715332, + "learning_rate": 9.574109213577904e-06, + "loss": 0.0634, + "step": 6000 + }, + { + "epoch": 11.433048433048434, + "grad_norm": 6.619261741638184, + "learning_rate": 9.521399957832596e-06, + "loss": 0.0564, + "step": 6025 + }, + { + "epoch": 11.480531813865147, + "grad_norm": 7.268097877502441, + "learning_rate": 9.468690702087288e-06, + "loss": 0.061, + "step": 6050 + }, + { + "epoch": 11.528015194681862, + "grad_norm": 6.525518417358398, + "learning_rate": 9.41598144634198e-06, + "loss": 0.0528, + "step": 6075 + }, + { + "epoch": 11.575498575498575, + "grad_norm": 19.773109436035156, + "learning_rate": 9.36327219059667e-06, + "loss": 0.0647, + "step": 6100 + }, + { + "epoch": 11.62298195631529, + "grad_norm": 1.7680853605270386, + "learning_rate": 9.310562934851361e-06, + "loss": 0.0707, + "step": 6125 + }, + { + "epoch": 11.670465337132004, + "grad_norm": 23.178977966308594, + "learning_rate": 9.257853679106051e-06, + "loss": 0.0678, + "step": 6150 + }, + { + "epoch": 11.717948717948717, + "grad_norm": 4.943007469177246, + "learning_rate": 9.205144423360743e-06, + "loss": 0.055, + "step": 6175 + }, + { + "epoch": 11.765432098765432, + "grad_norm": 11.310662269592285, + "learning_rate": 9.152435167615435e-06, + "loss": 0.0483, + "step": 6200 + }, + { + "epoch": 11.812915479582147, + "grad_norm": 16.74103546142578, + "learning_rate": 9.099725911870125e-06, + "loss": 0.0575, + "step": 6225 + }, + { + "epoch": 11.86039886039886, + "grad_norm": 6.127476215362549, + "learning_rate": 9.047016656124816e-06, + "loss": 0.0502, + "step": 6250 + }, + { + "epoch": 11.907882241215574, + "grad_norm": 2.1509368419647217, + "learning_rate": 8.994307400379507e-06, + "loss": 0.0478, + "step": 6275 + }, + { + "epoch": 11.95536562203229, + "grad_norm": 4.104255676269531, + "learning_rate": 8.941598144634198e-06, + "loss": 0.0478, + "step": 6300 + }, + { + "epoch": 12.0, + "eval_accuracy": 0.25766541151156536, + "eval_loss": 1.7872278690338135, + "eval_macro_f1": 0.6021417191968793, + "eval_micro_f1": 0.6147103136556636, + "eval_runtime": 4.9818, + "eval_samples_per_second": 373.155, + "eval_steps_per_second": 23.485, + "eval_weighted_f1": 0.6141078636029691, + "step": 6324 + }, + { + "epoch": 12.001899335232668, + "grad_norm": 2.3589696884155273, + "learning_rate": 8.888888888888888e-06, + "loss": 0.0572, + "step": 6325 + }, + { + "epoch": 12.049382716049383, + "grad_norm": 8.248597145080566, + "learning_rate": 8.836179633143582e-06, + "loss": 0.0418, + "step": 6350 + }, + { + "epoch": 12.096866096866098, + "grad_norm": 1.2999802827835083, + "learning_rate": 8.783470377398272e-06, + "loss": 0.0388, + "step": 6375 + }, + { + "epoch": 12.14434947768281, + "grad_norm": 1.5620146989822388, + "learning_rate": 8.730761121652964e-06, + "loss": 0.0386, + "step": 6400 + }, + { + "epoch": 12.191832858499525, + "grad_norm": 12.313676834106445, + "learning_rate": 8.678051865907654e-06, + "loss": 0.0533, + "step": 6425 + }, + { + "epoch": 12.239316239316238, + "grad_norm": 8.87389087677002, + "learning_rate": 8.625342610162345e-06, + "loss": 0.0401, + "step": 6450 + }, + { + "epoch": 12.286799620132953, + "grad_norm": 3.5049731731414795, + "learning_rate": 8.572633354417035e-06, + "loss": 0.0365, + "step": 6475 + }, + { + "epoch": 12.334283000949668, + "grad_norm": 1.5245490074157715, + "learning_rate": 8.519924098671729e-06, + "loss": 0.0447, + "step": 6500 + }, + { + "epoch": 12.381766381766381, + "grad_norm": 10.846375465393066, + "learning_rate": 8.467214842926419e-06, + "loss": 0.0475, + "step": 6525 + }, + { + "epoch": 12.429249762583096, + "grad_norm": 29.654497146606445, + "learning_rate": 8.41450558718111e-06, + "loss": 0.0543, + "step": 6550 + }, + { + "epoch": 12.47673314339981, + "grad_norm": 1.8981044292449951, + "learning_rate": 8.3617963314358e-06, + "loss": 0.0405, + "step": 6575 + }, + { + "epoch": 12.524216524216524, + "grad_norm": 4.788665771484375, + "learning_rate": 8.309087075690492e-06, + "loss": 0.0372, + "step": 6600 + }, + { + "epoch": 12.571699905033238, + "grad_norm": 4.4697394371032715, + "learning_rate": 8.256377819945182e-06, + "loss": 0.0524, + "step": 6625 + }, + { + "epoch": 12.619183285849953, + "grad_norm": 18.409130096435547, + "learning_rate": 8.203668564199874e-06, + "loss": 0.0427, + "step": 6650 + }, + { + "epoch": 12.666666666666666, + "grad_norm": 5.236688613891602, + "learning_rate": 8.150959308454566e-06, + "loss": 0.0363, + "step": 6675 + }, + { + "epoch": 12.71415004748338, + "grad_norm": 6.301948547363281, + "learning_rate": 8.098250052709256e-06, + "loss": 0.0414, + "step": 6700 + }, + { + "epoch": 12.761633428300096, + "grad_norm": 5.281355381011963, + "learning_rate": 8.045540796963948e-06, + "loss": 0.0384, + "step": 6725 + }, + { + "epoch": 12.809116809116809, + "grad_norm": 2.6161389350891113, + "learning_rate": 7.992831541218638e-06, + "loss": 0.0341, + "step": 6750 + }, + { + "epoch": 12.856600189933523, + "grad_norm": 4.726546764373779, + "learning_rate": 7.94012228547333e-06, + "loss": 0.0323, + "step": 6775 + }, + { + "epoch": 12.904083570750238, + "grad_norm": 1.8252311944961548, + "learning_rate": 7.887413029728021e-06, + "loss": 0.0435, + "step": 6800 + }, + { + "epoch": 12.951566951566951, + "grad_norm": 5.838278293609619, + "learning_rate": 7.834703773982713e-06, + "loss": 0.0466, + "step": 6825 + }, + { + "epoch": 12.999050332383666, + "grad_norm": 2.512367010116577, + "learning_rate": 7.781994518237403e-06, + "loss": 0.0331, + "step": 6850 + }, + { + "epoch": 13.0, + "eval_accuracy": 0.26573426573426573, + "eval_loss": 1.9706891775131226, + "eval_macro_f1": 0.6032090695110145, + "eval_micro_f1": 0.6180674671240709, + "eval_runtime": 5.0383, + "eval_samples_per_second": 368.977, + "eval_steps_per_second": 23.222, + "eval_weighted_f1": 0.614801620920999, + "step": 6851 + }, + { + "epoch": 13.045584045584045, + "grad_norm": 10.964371681213379, + "learning_rate": 7.729285262492095e-06, + "loss": 0.0287, + "step": 6875 + }, + { + "epoch": 13.09306742640076, + "grad_norm": 3.414742946624756, + "learning_rate": 7.676576006746785e-06, + "loss": 0.0233, + "step": 6900 + }, + { + "epoch": 13.140550807217474, + "grad_norm": 2.452306032180786, + "learning_rate": 7.6238667510014765e-06, + "loss": 0.0276, + "step": 6925 + }, + { + "epoch": 13.188034188034187, + "grad_norm": 13.015519142150879, + "learning_rate": 7.5711574952561674e-06, + "loss": 0.0275, + "step": 6950 + }, + { + "epoch": 13.235517568850902, + "grad_norm": 1.6397945880889893, + "learning_rate": 7.518448239510858e-06, + "loss": 0.0357, + "step": 6975 + }, + { + "epoch": 13.283000949667617, + "grad_norm": 2.658342123031616, + "learning_rate": 7.46573898376555e-06, + "loss": 0.0309, + "step": 7000 + }, + { + "epoch": 13.33048433048433, + "grad_norm": 2.3749077320098877, + "learning_rate": 7.413029728020241e-06, + "loss": 0.0254, + "step": 7025 + }, + { + "epoch": 13.377967711301045, + "grad_norm": 8.341275215148926, + "learning_rate": 7.360320472274932e-06, + "loss": 0.0289, + "step": 7050 + }, + { + "epoch": 13.42545109211776, + "grad_norm": 21.5805606842041, + "learning_rate": 7.307611216529623e-06, + "loss": 0.0392, + "step": 7075 + }, + { + "epoch": 13.472934472934472, + "grad_norm": 4.2061767578125, + "learning_rate": 7.2549019607843145e-06, + "loss": 0.0225, + "step": 7100 + }, + { + "epoch": 13.520417853751187, + "grad_norm": 1.3556793928146362, + "learning_rate": 7.202192705039005e-06, + "loss": 0.0253, + "step": 7125 + }, + { + "epoch": 13.567901234567902, + "grad_norm": 3.0510809421539307, + "learning_rate": 7.149483449293697e-06, + "loss": 0.0287, + "step": 7150 + }, + { + "epoch": 13.615384615384615, + "grad_norm": 2.111264228820801, + "learning_rate": 7.096774193548388e-06, + "loss": 0.0226, + "step": 7175 + }, + { + "epoch": 13.66286799620133, + "grad_norm": 1.2025902271270752, + "learning_rate": 7.044064937803079e-06, + "loss": 0.0305, + "step": 7200 + }, + { + "epoch": 13.710351377018045, + "grad_norm": 3.1950247287750244, + "learning_rate": 6.99135568205777e-06, + "loss": 0.0351, + "step": 7225 + }, + { + "epoch": 13.757834757834758, + "grad_norm": 2.3655929565429688, + "learning_rate": 6.938646426312461e-06, + "loss": 0.0247, + "step": 7250 + }, + { + "epoch": 13.805318138651472, + "grad_norm": 1.4691689014434814, + "learning_rate": 6.885937170567152e-06, + "loss": 0.0239, + "step": 7275 + }, + { + "epoch": 13.852801519468187, + "grad_norm": 63.52339553833008, + "learning_rate": 6.8332279148218425e-06, + "loss": 0.0285, + "step": 7300 + }, + { + "epoch": 13.9002849002849, + "grad_norm": 2.136404275894165, + "learning_rate": 6.780518659076535e-06, + "loss": 0.0303, + "step": 7325 + }, + { + "epoch": 13.947768281101615, + "grad_norm": 0.9162741899490356, + "learning_rate": 6.727809403331226e-06, + "loss": 0.0316, + "step": 7350 + }, + { + "epoch": 13.995251661918328, + "grad_norm": 5.802157402038574, + "learning_rate": 6.675100147585917e-06, + "loss": 0.0207, + "step": 7375 + }, + { + "epoch": 14.0, + "eval_accuracy": 0.2608929532006455, + "eval_loss": 2.0579302310943604, + "eval_macro_f1": 0.600143920444938, + "eval_micro_f1": 0.6170791017021885, + "eval_runtime": 4.841, + "eval_samples_per_second": 384.014, + "eval_steps_per_second": 24.169, + "eval_weighted_f1": 0.6139209499452548, + "step": 7378 + } + ], + "logging_steps": 25, + "max_steps": 10540, + "num_input_tokens_seen": 0, + "num_train_epochs": 20, + "save_steps": 500, + "stateful_callbacks": { + "EarlyStoppingCallback": { + "args": { + "early_stopping_patience": 3, + "early_stopping_threshold": 0.0 + }, + "attributes": { + "early_stopping_patience_counter": 3 + } + }, + "TrainerControl": { + "args": { + "should_epoch_stop": false, + "should_evaluate": false, + "should_log": false, + "should_save": true, + "should_training_stop": true + }, + "attributes": {} + } + }, + "total_flos": 0.0, + "train_batch_size": 8, + "trial_name": null, + "trial_params": null +} diff --git a/ood-vigoemotions-mmbert-source-source-02-s44-20260905T152208Z-37344fee/model.safetensors b/ood-vigoemotions-mmbert-source-source-02-s44-20260905T152208Z-37344fee/model.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..711449d58187169f6c06077975bee95c080e6ffc --- /dev/null +++ b/ood-vigoemotions-mmbert-source-source-02-s44-20260905T152208Z-37344fee/model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:d476b73b769518a9dbfaa3b06140855d3f4e1c1543679881721bd3d0515c72e1 +size 1227859104 diff --git a/ood-vigoemotions-mmbert-source-source-02-s44-20260905T152208Z-37344fee/trainer/checkpoint-7905/optimizer.pt b/ood-vigoemotions-mmbert-source-source-02-s44-20260905T152208Z-37344fee/trainer/checkpoint-7905/optimizer.pt new file mode 100644 index 0000000000000000000000000000000000000000..b1f0d53a1dad7f35751e5a510283c89398519cc9 --- /dev/null +++ b/ood-vigoemotions-mmbert-source-source-02-s44-20260905T152208Z-37344fee/trainer/checkpoint-7905/optimizer.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:7b40d1db0ba10c92b35f6f357620ed612da600544362f820a4a7ebc576c6475c +size 2455803898 diff --git a/ood-vigoemotions-mmbert-source-source-02-s44-20260905T152208Z-37344fee/trainer/checkpoint-7905/special_tokens_map.json b/ood-vigoemotions-mmbert-source-source-02-s44-20260905T152208Z-37344fee/trainer/checkpoint-7905/special_tokens_map.json new file mode 100644 index 0000000000000000000000000000000000000000..be4ad795f49075ac5e56d8108ef63be314a1e967 --- /dev/null +++ b/ood-vigoemotions-mmbert-source-source-02-s44-20260905T152208Z-37344fee/trainer/checkpoint-7905/special_tokens_map.json @@ -0,0 +1,55 @@ +{ + "additional_special_tokens": [ + "", + "" + ], + "bos_token": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false + }, + "cls_token": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false + }, + "eos_token": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false + }, + "mask_token": { + "content": "", + "lstrip": true, + "normalized": false, + "rstrip": false, + "single_word": false + }, + "pad_token": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false + }, + "sep_token": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false + }, + "unk_token": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false + } +} diff --git a/ood-vigoemotions-mmbert-source-source-02-s44-20260905T152208Z-37344fee/trainer/checkpoint-7905/tokenizer_config.json b/ood-vigoemotions-mmbert-source-source-02-s44-20260905T152208Z-37344fee/trainer/checkpoint-7905/tokenizer_config.json new file mode 100644 index 0000000000000000000000000000000000000000..b002fb8c2960be8ebf9c373d6e5784b4565d7edc --- /dev/null +++ b/ood-vigoemotions-mmbert-source-source-02-s44-20260905T152208Z-37344fee/trainer/checkpoint-7905/tokenizer_config.json @@ -0,0 +1,2018 @@ +{ + "add_bos_token": true, + "added_tokens_decoder": { + "0": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "1": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "2": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "3": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "4": { + "content": "", + "lstrip": true, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "5": { + "content": "<2mass>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "6": { + "content": "[@BOS@]", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "7": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "8": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "9": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "10": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "11": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "12": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "13": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "14": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "15": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "16": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "17": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "18": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "19": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "20": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "21": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "22": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "23": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "24": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "25": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "26": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "27": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "28": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "29": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "30": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "31": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "32": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "33": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "34": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "35": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "36": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "37": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "38": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "39": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "40": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "41": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "42": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "43": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "44": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "45": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "46": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "47": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "48": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "49": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "50": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "51": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "52": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "53": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "54": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "55": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "56": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "57": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "58": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "59": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "60": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "61": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "62": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "63": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "64": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "65": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "66": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "67": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "68": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "69": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "70": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "71": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "72": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "73": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "74": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "75": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "76": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "77": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "78": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "79": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "80": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "81": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "82": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "83": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "84": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "85": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "86": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "87": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "88": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "89": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "90": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "91": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "92": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "93": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "94": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "95": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "96": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "97": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "98": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "99": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "100": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "101": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "102": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "103": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "104": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "105": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "106": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "107": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "108": { + "content": "\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "109": { + "content": "\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "110": { + "content": "\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "111": { + "content": "\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "112": { + "content": "\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "113": { + "content": "\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "114": { + "content": "\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "115": { + "content": "\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "116": { + "content": "\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "117": { + "content": "\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "118": { + "content": "\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "119": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "120": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "121": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "122": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "123": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "124": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "125": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "126": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "127": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "128": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "129": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "130": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "131": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "132": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "133": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "134": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "135": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "136": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "137": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "138": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "139": { + "content": "▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "140": { + "content": "▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "141": { + "content": "▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "142": { + "content": "▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "143": { + "content": "▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "144": { + "content": "▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "145": { + "content": "▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "146": { + "content": "▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "147": { + "content": "▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "148": { + "content": "▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "149": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "150": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "151": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "152": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "153": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "154": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "155": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "156": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "157": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "158": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "159": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "160": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "161": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "162": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "163": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "164": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "165": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "166": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "167": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "168": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "169": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "170": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "172": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "173": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "174": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "175": { + "content": "
", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "171": { + "content": "
", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "176": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "177": { + "content": "
", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "178": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "179": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "180": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "181": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "182": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "183": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "184": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "185": { + "content": "

", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "186": { + "content": "

", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "187": { + "content": "

", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "188": { + "content": "

", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "189": { + "content": "

", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "190": { + "content": "
", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "191": { + "content": "
", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "192": { + "content": "
", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "193": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "194": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "195": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "196": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "197": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "198": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "199": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "200": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "201": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "202": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "203": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "204": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "205": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "206": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "207": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "208": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "209": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "210": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "211": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "212": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "213": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "214": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "215": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "216": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255968": { + "content": "[toxicity=0]", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255969": { + "content": "\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255970": { + "content": "\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255971": { + "content": "\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255972": { + "content": "\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255973": { + "content": "\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255974": { + "content": "\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255975": { + "content": "\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255976": { + "content": "\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255977": { + "content": "\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255978": { + "content": "\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255979": { + "content": "\t\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255980": { + "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255981": { + "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255982": { + "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255983": { + "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255984": { + "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255985": { + "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255986": { + "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255987": { + "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255988": { + "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255989": { + "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255990": { + "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255991": { + "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255992": { + "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255993": { + "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255994": { + "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255995": { + "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255996": { + "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255997": { + "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255998": { + "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255999": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + } + }, + "additional_special_tokens": [ + "", + "" + ], + "bos_token": "", + "clean_up_tokenization_spaces": false, + "cls_token": "", + "eos_token": "", + "extra_special_tokens": {}, + "mask_token": "", + "model_input_names": [ + "input_ids", + "attention_mask" + ], + "model_max_length": 8192, + "pad_token": "", + "padding_side": "right", + "sep_token": "", + "spaces_between_special_tokens": false, + "tokenizer_class": "PreTrainedTokenizerFast", + "unk_token": "" +} diff --git a/ood-vigoemotions-mmbert-source-source-02-s44-20260905T152208Z-37344fee/trainer/checkpoint-7905/trainer_state.json b/ood-vigoemotions-mmbert-source-source-02-s44-20260905T152208Z-37344fee/trainer/checkpoint-7905/trainer_state.json new file mode 100644 index 0000000000000000000000000000000000000000..76bc5c2475e69dbc8dd02ec25df4856b7e1e8cea --- /dev/null +++ b/ood-vigoemotions-mmbert-source-source-02-s44-20260905T152208Z-37344fee/trainer/checkpoint-7905/trainer_state.json @@ -0,0 +1,2435 @@ +{ + "best_global_step": 7905, + "best_metric": 0.5907126666859718, + "best_model_checkpoint": "/root/VNEncoder/results/checkpoints/ood-vigoemotions-mmbert-source-source-02-s44-20260905T152208Z-37344fee/trainer/checkpoint-7905", + "epoch": 15.0, + "eval_steps": 500, + "global_step": 7905, + "is_hyper_param_search": false, + "is_local_process_zero": true, + "is_world_process_zero": true, + "log_history": [ + { + "epoch": 0.04748338081671415, + "grad_norm": 34.162723541259766, + "learning_rate": 4.554079696394687e-07, + "loss": 6.1444, + "step": 25 + }, + { + "epoch": 0.0949667616334283, + "grad_norm": 36.55830383300781, + "learning_rate": 9.297912713472485e-07, + "loss": 6.2258, + "step": 50 + }, + { + "epoch": 0.14245014245014245, + "grad_norm": 33.597232818603516, + "learning_rate": 1.4041745730550286e-06, + "loss": 6.2708, + "step": 75 + }, + { + "epoch": 0.1899335232668566, + "grad_norm": 39.089359283447266, + "learning_rate": 1.8785578747628086e-06, + "loss": 6.2674, + "step": 100 + }, + { + "epoch": 0.23741690408357075, + "grad_norm": 30.89531135559082, + "learning_rate": 2.3529411764705885e-06, + "loss": 6.2874, + "step": 125 + }, + { + "epoch": 0.2849002849002849, + "grad_norm": 58.9684944152832, + "learning_rate": 2.8273244781783683e-06, + "loss": 6.0477, + "step": 150 + }, + { + "epoch": 0.33238366571699907, + "grad_norm": 142.00091552734375, + "learning_rate": 3.3017077798861485e-06, + "loss": 5.3108, + "step": 175 + }, + { + "epoch": 0.3798670465337132, + "grad_norm": 32.22747039794922, + "learning_rate": 3.7760910815939284e-06, + "loss": 5.1682, + "step": 200 + }, + { + "epoch": 0.42735042735042733, + "grad_norm": 30.62629508972168, + "learning_rate": 4.250474383301708e-06, + "loss": 4.9753, + "step": 225 + }, + { + "epoch": 0.4748338081671415, + "grad_norm": 52.48829650878906, + "learning_rate": 4.724857685009488e-06, + "loss": 4.8218, + "step": 250 + }, + { + "epoch": 0.5223171889838556, + "grad_norm": 48.24330139160156, + "learning_rate": 5.199240986717268e-06, + "loss": 4.5239, + "step": 275 + }, + { + "epoch": 0.5698005698005698, + "grad_norm": 79.2529296875, + "learning_rate": 5.673624288425048e-06, + "loss": 4.2647, + "step": 300 + }, + { + "epoch": 0.6172839506172839, + "grad_norm": 66.08687591552734, + "learning_rate": 6.1480075901328275e-06, + "loss": 4.1624, + "step": 325 + }, + { + "epoch": 0.6647673314339981, + "grad_norm": 66.99076080322266, + "learning_rate": 6.622390891840608e-06, + "loss": 3.8941, + "step": 350 + }, + { + "epoch": 0.7122507122507122, + "grad_norm": 51.146968841552734, + "learning_rate": 7.096774193548388e-06, + "loss": 3.8293, + "step": 375 + }, + { + "epoch": 0.7597340930674265, + "grad_norm": 37.30567932128906, + "learning_rate": 7.5711574952561674e-06, + "loss": 3.7049, + "step": 400 + }, + { + "epoch": 0.8072174738841406, + "grad_norm": 43.886505126953125, + "learning_rate": 8.045540796963948e-06, + "loss": 3.656, + "step": 425 + }, + { + "epoch": 0.8547008547008547, + "grad_norm": 42.3010368347168, + "learning_rate": 8.519924098671729e-06, + "loss": 3.6309, + "step": 450 + }, + { + "epoch": 0.9021842355175689, + "grad_norm": 45.34294509887695, + "learning_rate": 8.994307400379507e-06, + "loss": 3.4031, + "step": 475 + }, + { + "epoch": 0.949667616334283, + "grad_norm": 54.65419387817383, + "learning_rate": 9.468690702087288e-06, + "loss": 3.2897, + "step": 500 + }, + { + "epoch": 0.9971509971509972, + "grad_norm": 55.79121017456055, + "learning_rate": 9.943074003795067e-06, + "loss": 3.3332, + "step": 525 + }, + { + "epoch": 1.0, + "eval_accuracy": 0.0005379236148466917, + "eval_loss": 0.8047798871994019, + "eval_macro_f1": 0.3364472078136643, + "eval_micro_f1": 0.3479386208171566, + "eval_runtime": 5.6709, + "eval_samples_per_second": 327.814, + "eval_steps_per_second": 20.632, + "eval_weighted_f1": 0.40064642675810264, + "step": 527 + }, + { + "epoch": 1.043684710351377, + "grad_norm": 50.92436599731445, + "learning_rate": 1.0417457305502846e-05, + "loss": 3.0012, + "step": 550 + }, + { + "epoch": 1.0911680911680912, + "grad_norm": 51.1168327331543, + "learning_rate": 1.0891840607210626e-05, + "loss": 2.8467, + "step": 575 + }, + { + "epoch": 1.1386514719848053, + "grad_norm": 40.4506950378418, + "learning_rate": 1.1366223908918409e-05, + "loss": 2.8875, + "step": 600 + }, + { + "epoch": 1.1861348528015194, + "grad_norm": 46.12940979003906, + "learning_rate": 1.1840607210626186e-05, + "loss": 2.8581, + "step": 625 + }, + { + "epoch": 1.2336182336182335, + "grad_norm": 57.1807975769043, + "learning_rate": 1.2314990512333966e-05, + "loss": 2.8958, + "step": 650 + }, + { + "epoch": 1.2811016144349479, + "grad_norm": 63.2441520690918, + "learning_rate": 1.2789373814041749e-05, + "loss": 2.9974, + "step": 675 + }, + { + "epoch": 1.328584995251662, + "grad_norm": 35.746822357177734, + "learning_rate": 1.3263757115749526e-05, + "loss": 3.0096, + "step": 700 + }, + { + "epoch": 1.376068376068376, + "grad_norm": 140.2974395751953, + "learning_rate": 1.3738140417457306e-05, + "loss": 2.8375, + "step": 725 + }, + { + "epoch": 1.4235517568850902, + "grad_norm": 86.68020629882812, + "learning_rate": 1.4212523719165085e-05, + "loss": 2.8921, + "step": 750 + }, + { + "epoch": 1.4710351377018043, + "grad_norm": 55.23737335205078, + "learning_rate": 1.4686907020872866e-05, + "loss": 3.0217, + "step": 775 + }, + { + "epoch": 1.5185185185185186, + "grad_norm": 48.04639434814453, + "learning_rate": 1.5161290322580646e-05, + "loss": 2.7802, + "step": 800 + }, + { + "epoch": 1.5660018993352327, + "grad_norm": 31.351240158081055, + "learning_rate": 1.5635673624288425e-05, + "loss": 2.7824, + "step": 825 + }, + { + "epoch": 1.6134852801519468, + "grad_norm": 48.16146469116211, + "learning_rate": 1.6110056925996208e-05, + "loss": 2.7755, + "step": 850 + }, + { + "epoch": 1.660968660968661, + "grad_norm": 59.989097595214844, + "learning_rate": 1.6584440227703987e-05, + "loss": 2.8255, + "step": 875 + }, + { + "epoch": 1.708452041785375, + "grad_norm": 39.0864372253418, + "learning_rate": 1.7058823529411767e-05, + "loss": 2.6891, + "step": 900 + }, + { + "epoch": 1.7559354226020893, + "grad_norm": 35.95862579345703, + "learning_rate": 1.7533206831119546e-05, + "loss": 2.7272, + "step": 925 + }, + { + "epoch": 1.8034188034188035, + "grad_norm": 59.74723815917969, + "learning_rate": 1.8007590132827325e-05, + "loss": 2.6682, + "step": 950 + }, + { + "epoch": 1.8509021842355176, + "grad_norm": 37.617794036865234, + "learning_rate": 1.8481973434535105e-05, + "loss": 2.8645, + "step": 975 + }, + { + "epoch": 1.8983855650522319, + "grad_norm": 33.81266784667969, + "learning_rate": 1.8956356736242888e-05, + "loss": 2.7383, + "step": 1000 + }, + { + "epoch": 1.9458689458689458, + "grad_norm": 75.7305908203125, + "learning_rate": 1.9430740037950667e-05, + "loss": 2.7314, + "step": 1025 + }, + { + "epoch": 1.99335232668566, + "grad_norm": 33.35398864746094, + "learning_rate": 1.9905123339658446e-05, + "loss": 2.5902, + "step": 1050 + }, + { + "epoch": 2.0, + "eval_accuracy": 0.01775147928994083, + "eval_loss": 0.6891770958900452, + "eval_macro_f1": 0.41180038828201304, + "eval_micro_f1": 0.4208560758402758, + "eval_runtime": 4.9435, + "eval_samples_per_second": 376.051, + "eval_steps_per_second": 23.668, + "eval_weighted_f1": 0.4598716002150326, + "step": 1054 + }, + { + "epoch": 2.03988603988604, + "grad_norm": 71.47340393066406, + "learning_rate": 1.9957832595403753e-05, + "loss": 2.1893, + "step": 1075 + }, + { + "epoch": 2.087369420702754, + "grad_norm": 34.373592376708984, + "learning_rate": 1.9905123339658446e-05, + "loss": 2.1511, + "step": 1100 + }, + { + "epoch": 2.134852801519468, + "grad_norm": 18.752532958984375, + "learning_rate": 1.9852414083913136e-05, + "loss": 1.9717, + "step": 1125 + }, + { + "epoch": 2.1823361823361824, + "grad_norm": 40.68827819824219, + "learning_rate": 1.9799704828167827e-05, + "loss": 2.1369, + "step": 1150 + }, + { + "epoch": 2.2298195631528963, + "grad_norm": 29.010814666748047, + "learning_rate": 1.9746995572422517e-05, + "loss": 2.037, + "step": 1175 + }, + { + "epoch": 2.2773029439696106, + "grad_norm": 33.29954528808594, + "learning_rate": 1.969428631667721e-05, + "loss": 2.053, + "step": 1200 + }, + { + "epoch": 2.324786324786325, + "grad_norm": 32.961490631103516, + "learning_rate": 1.9641577060931903e-05, + "loss": 2.043, + "step": 1225 + }, + { + "epoch": 2.372269705603039, + "grad_norm": 61.129783630371094, + "learning_rate": 1.9588867805186593e-05, + "loss": 2.0912, + "step": 1250 + }, + { + "epoch": 2.419753086419753, + "grad_norm": 49.88624572753906, + "learning_rate": 1.9536158549441284e-05, + "loss": 1.995, + "step": 1275 + }, + { + "epoch": 2.467236467236467, + "grad_norm": 54.424903869628906, + "learning_rate": 1.9483449293695974e-05, + "loss": 1.9819, + "step": 1300 + }, + { + "epoch": 2.5147198480531814, + "grad_norm": 43.5580940246582, + "learning_rate": 1.9430740037950667e-05, + "loss": 2.0496, + "step": 1325 + }, + { + "epoch": 2.5622032288698957, + "grad_norm": 28.404251098632812, + "learning_rate": 1.9378030782205357e-05, + "loss": 2.0856, + "step": 1350 + }, + { + "epoch": 2.6096866096866096, + "grad_norm": 56.033836364746094, + "learning_rate": 1.9325321526460047e-05, + "loss": 2.1025, + "step": 1375 + }, + { + "epoch": 2.657169990503324, + "grad_norm": 30.688547134399414, + "learning_rate": 1.927261227071474e-05, + "loss": 2.0668, + "step": 1400 + }, + { + "epoch": 2.7046533713200382, + "grad_norm": 51.40027618408203, + "learning_rate": 1.921990301496943e-05, + "loss": 2.2163, + "step": 1425 + }, + { + "epoch": 2.752136752136752, + "grad_norm": 44.746917724609375, + "learning_rate": 1.916719375922412e-05, + "loss": 1.9534, + "step": 1450 + }, + { + "epoch": 2.799620132953466, + "grad_norm": 23.693490982055664, + "learning_rate": 1.911448450347881e-05, + "loss": 1.9513, + "step": 1475 + }, + { + "epoch": 2.8471035137701803, + "grad_norm": 40.34550094604492, + "learning_rate": 1.9061775247733504e-05, + "loss": 1.9943, + "step": 1500 + }, + { + "epoch": 2.8945868945868947, + "grad_norm": 43.02825927734375, + "learning_rate": 1.9009065991988194e-05, + "loss": 2.0602, + "step": 1525 + }, + { + "epoch": 2.9420702754036085, + "grad_norm": 42.245784759521484, + "learning_rate": 1.8956356736242888e-05, + "loss": 1.9576, + "step": 1550 + }, + { + "epoch": 2.989553656220323, + "grad_norm": 38.256168365478516, + "learning_rate": 1.8903647480497578e-05, + "loss": 1.9699, + "step": 1575 + }, + { + "epoch": 3.0, + "eval_accuracy": 0.0328133405056482, + "eval_loss": 0.6396704912185669, + "eval_macro_f1": 0.4600545960071464, + "eval_micro_f1": 0.46655478302565334, + "eval_runtime": 4.8346, + "eval_samples_per_second": 384.521, + "eval_steps_per_second": 24.201, + "eval_weighted_f1": 0.5101439509128096, + "step": 1581 + }, + { + "epoch": 3.0360873694207027, + "grad_norm": 20.20168685913086, + "learning_rate": 1.8850938224752268e-05, + "loss": 1.5347, + "step": 1600 + }, + { + "epoch": 3.083570750237417, + "grad_norm": 28.101213455200195, + "learning_rate": 1.879822896900696e-05, + "loss": 1.3689, + "step": 1625 + }, + { + "epoch": 3.131054131054131, + "grad_norm": 33.39235305786133, + "learning_rate": 1.874551971326165e-05, + "loss": 1.4377, + "step": 1650 + }, + { + "epoch": 3.178537511870845, + "grad_norm": 30.524776458740234, + "learning_rate": 1.869281045751634e-05, + "loss": 1.3264, + "step": 1675 + }, + { + "epoch": 3.2260208926875595, + "grad_norm": 26.046913146972656, + "learning_rate": 1.864010120177103e-05, + "loss": 1.4242, + "step": 1700 + }, + { + "epoch": 3.2735042735042734, + "grad_norm": 31.101991653442383, + "learning_rate": 1.8587391946025725e-05, + "loss": 1.3763, + "step": 1725 + }, + { + "epoch": 3.3209876543209877, + "grad_norm": 20.603694915771484, + "learning_rate": 1.8534682690280415e-05, + "loss": 1.3973, + "step": 1750 + }, + { + "epoch": 3.3684710351377016, + "grad_norm": 30.091970443725586, + "learning_rate": 1.8481973434535105e-05, + "loss": 1.3262, + "step": 1775 + }, + { + "epoch": 3.415954415954416, + "grad_norm": 30.08812713623047, + "learning_rate": 1.8429264178789795e-05, + "loss": 1.3596, + "step": 1800 + }, + { + "epoch": 3.4634377967711303, + "grad_norm": 28.665847778320312, + "learning_rate": 1.8376554923044488e-05, + "loss": 1.3754, + "step": 1825 + }, + { + "epoch": 3.510921177587844, + "grad_norm": 23.131080627441406, + "learning_rate": 1.832384566729918e-05, + "loss": 1.4043, + "step": 1850 + }, + { + "epoch": 3.5584045584045585, + "grad_norm": 15.830599784851074, + "learning_rate": 1.8271136411553872e-05, + "loss": 1.3662, + "step": 1875 + }, + { + "epoch": 3.6058879392212724, + "grad_norm": 26.26952362060547, + "learning_rate": 1.8218427155808562e-05, + "loss": 1.3602, + "step": 1900 + }, + { + "epoch": 3.6533713200379867, + "grad_norm": 31.234373092651367, + "learning_rate": 1.8165717900063252e-05, + "loss": 1.3459, + "step": 1925 + }, + { + "epoch": 3.700854700854701, + "grad_norm": 34.2337532043457, + "learning_rate": 1.8113008644317945e-05, + "loss": 1.4278, + "step": 1950 + }, + { + "epoch": 3.748338081671415, + "grad_norm": 36.79152297973633, + "learning_rate": 1.8060299388572635e-05, + "loss": 1.3326, + "step": 1975 + }, + { + "epoch": 3.7958214624881292, + "grad_norm": 40.74213409423828, + "learning_rate": 1.8007590132827325e-05, + "loss": 1.3891, + "step": 2000 + }, + { + "epoch": 3.8433048433048436, + "grad_norm": 100.57337188720703, + "learning_rate": 1.7954880877082015e-05, + "loss": 1.4088, + "step": 2025 + }, + { + "epoch": 3.8907882241215574, + "grad_norm": 23.349496841430664, + "learning_rate": 1.790217162133671e-05, + "loss": 1.3784, + "step": 2050 + }, + { + "epoch": 3.9382716049382713, + "grad_norm": 31.829254150390625, + "learning_rate": 1.78494623655914e-05, + "loss": 1.492, + "step": 2075 + }, + { + "epoch": 3.9857549857549857, + "grad_norm": 25.376209259033203, + "learning_rate": 1.779675310984609e-05, + "loss": 1.3584, + "step": 2100 + }, + { + "epoch": 4.0, + "eval_accuracy": 0.08230231307154384, + "eval_loss": 0.7609887719154358, + "eval_macro_f1": 0.5337177326494273, + "eval_micro_f1": 0.5383452965621458, + "eval_runtime": 4.899, + "eval_samples_per_second": 379.465, + "eval_steps_per_second": 23.882, + "eval_weighted_f1": 0.5553705232275817, + "step": 2108 + }, + { + "epoch": 4.032288698955366, + "grad_norm": 28.834596633911133, + "learning_rate": 1.7744043854100782e-05, + "loss": 1.0238, + "step": 2125 + }, + { + "epoch": 4.07977207977208, + "grad_norm": 27.219921112060547, + "learning_rate": 1.7691334598355472e-05, + "loss": 0.9589, + "step": 2150 + }, + { + "epoch": 4.127255460588794, + "grad_norm": 9.110738754272461, + "learning_rate": 1.7638625342610163e-05, + "loss": 0.94, + "step": 2175 + }, + { + "epoch": 4.174738841405508, + "grad_norm": 32.859806060791016, + "learning_rate": 1.7585916086864856e-05, + "loss": 0.9117, + "step": 2200 + }, + { + "epoch": 4.222222222222222, + "grad_norm": 22.447908401489258, + "learning_rate": 1.7533206831119546e-05, + "loss": 0.9674, + "step": 2225 + }, + { + "epoch": 4.269705603038936, + "grad_norm": 24.79280662536621, + "learning_rate": 1.748049757537424e-05, + "loss": 0.9744, + "step": 2250 + }, + { + "epoch": 4.317188983855651, + "grad_norm": 21.35533905029297, + "learning_rate": 1.742778831962893e-05, + "loss": 0.9193, + "step": 2275 + }, + { + "epoch": 4.364672364672365, + "grad_norm": 25.3253173828125, + "learning_rate": 1.737507906388362e-05, + "loss": 0.9302, + "step": 2300 + }, + { + "epoch": 4.412155745489079, + "grad_norm": 29.509456634521484, + "learning_rate": 1.732236980813831e-05, + "loss": 0.9541, + "step": 2325 + }, + { + "epoch": 4.459639126305793, + "grad_norm": 26.883155822753906, + "learning_rate": 1.7269660552393003e-05, + "loss": 0.9354, + "step": 2350 + }, + { + "epoch": 4.507122507122507, + "grad_norm": 44.8468132019043, + "learning_rate": 1.7216951296647693e-05, + "loss": 0.9985, + "step": 2375 + }, + { + "epoch": 4.554605887939221, + "grad_norm": 22.781490325927734, + "learning_rate": 1.7164242040902383e-05, + "loss": 1.0022, + "step": 2400 + }, + { + "epoch": 4.602089268755935, + "grad_norm": 29.72208023071289, + "learning_rate": 1.7111532785157073e-05, + "loss": 1.0015, + "step": 2425 + }, + { + "epoch": 4.64957264957265, + "grad_norm": 27.31267738342285, + "learning_rate": 1.7058823529411767e-05, + "loss": 0.8904, + "step": 2450 + }, + { + "epoch": 4.697056030389364, + "grad_norm": 23.065595626831055, + "learning_rate": 1.7006114273666457e-05, + "loss": 0.8698, + "step": 2475 + }, + { + "epoch": 4.744539411206078, + "grad_norm": 39.216705322265625, + "learning_rate": 1.6953405017921147e-05, + "loss": 0.9714, + "step": 2500 + }, + { + "epoch": 4.792022792022792, + "grad_norm": 45.64983367919922, + "learning_rate": 1.690069576217584e-05, + "loss": 0.8886, + "step": 2525 + }, + { + "epoch": 4.839506172839506, + "grad_norm": 24.30289649963379, + "learning_rate": 1.684798650643053e-05, + "loss": 0.9215, + "step": 2550 + }, + { + "epoch": 4.88698955365622, + "grad_norm": 55.504310607910156, + "learning_rate": 1.6795277250685224e-05, + "loss": 0.9372, + "step": 2575 + }, + { + "epoch": 4.934472934472934, + "grad_norm": 25.460933685302734, + "learning_rate": 1.6742567994939914e-05, + "loss": 0.9582, + "step": 2600 + }, + { + "epoch": 4.981956315289649, + "grad_norm": 24.881301879882812, + "learning_rate": 1.6689858739194604e-05, + "loss": 0.9422, + "step": 2625 + }, + { + "epoch": 5.0, + "eval_accuracy": 0.12641204948897256, + "eval_loss": 0.8357337713241577, + "eval_macro_f1": 0.5648626881399819, + "eval_micro_f1": 0.5698924731182796, + "eval_runtime": 4.9764, + "eval_samples_per_second": 373.561, + "eval_steps_per_second": 23.511, + "eval_weighted_f1": 0.584282614713785, + "step": 2635 + }, + { + "epoch": 5.028490028490029, + "grad_norm": 27.813451766967773, + "learning_rate": 1.6637149483449294e-05, + "loss": 0.7289, + "step": 2650 + }, + { + "epoch": 5.075973409306743, + "grad_norm": 21.82946014404297, + "learning_rate": 1.6584440227703987e-05, + "loss": 0.7127, + "step": 2675 + }, + { + "epoch": 5.1234567901234565, + "grad_norm": 19.29052734375, + "learning_rate": 1.6531730971958677e-05, + "loss": 0.6762, + "step": 2700 + }, + { + "epoch": 5.170940170940171, + "grad_norm": 42.652442932128906, + "learning_rate": 1.6479021716213367e-05, + "loss": 0.6371, + "step": 2725 + }, + { + "epoch": 5.218423551756885, + "grad_norm": 18.059337615966797, + "learning_rate": 1.642631246046806e-05, + "loss": 0.6415, + "step": 2750 + }, + { + "epoch": 5.265906932573599, + "grad_norm": 14.32496452331543, + "learning_rate": 1.637360320472275e-05, + "loss": 0.6583, + "step": 2775 + }, + { + "epoch": 5.313390313390314, + "grad_norm": 41.976932525634766, + "learning_rate": 1.632089394897744e-05, + "loss": 0.678, + "step": 2800 + }, + { + "epoch": 5.360873694207028, + "grad_norm": 38.39088439941406, + "learning_rate": 1.626818469323213e-05, + "loss": 0.6329, + "step": 2825 + }, + { + "epoch": 5.4083570750237415, + "grad_norm": 21.257617950439453, + "learning_rate": 1.6215475437486824e-05, + "loss": 0.6527, + "step": 2850 + }, + { + "epoch": 5.455840455840455, + "grad_norm": 15.757341384887695, + "learning_rate": 1.6162766181741514e-05, + "loss": 0.6476, + "step": 2875 + }, + { + "epoch": 5.50332383665717, + "grad_norm": 11.70419979095459, + "learning_rate": 1.6110056925996208e-05, + "loss": 0.6269, + "step": 2900 + }, + { + "epoch": 5.550807217473884, + "grad_norm": 44.25235366821289, + "learning_rate": 1.6057347670250898e-05, + "loss": 0.6537, + "step": 2925 + }, + { + "epoch": 5.598290598290598, + "grad_norm": 25.33641242980957, + "learning_rate": 1.6004638414505588e-05, + "loss": 0.6865, + "step": 2950 + }, + { + "epoch": 5.645773979107313, + "grad_norm": 16.503864288330078, + "learning_rate": 1.595192915876028e-05, + "loss": 0.6372, + "step": 2975 + }, + { + "epoch": 5.693257359924027, + "grad_norm": 21.774084091186523, + "learning_rate": 1.589921990301497e-05, + "loss": 0.6898, + "step": 3000 + }, + { + "epoch": 5.7407407407407405, + "grad_norm": 39.14348220825195, + "learning_rate": 1.584651064726966e-05, + "loss": 0.6671, + "step": 3025 + }, + { + "epoch": 5.788224121557455, + "grad_norm": 28.4456787109375, + "learning_rate": 1.579380139152435e-05, + "loss": 0.6059, + "step": 3050 + }, + { + "epoch": 5.835707502374169, + "grad_norm": 25.33370018005371, + "learning_rate": 1.5741092135779045e-05, + "loss": 0.6202, + "step": 3075 + }, + { + "epoch": 5.883190883190883, + "grad_norm": 30.26043128967285, + "learning_rate": 1.5688382880033735e-05, + "loss": 0.6552, + "step": 3100 + }, + { + "epoch": 5.930674264007598, + "grad_norm": 18.209794998168945, + "learning_rate": 1.5635673624288425e-05, + "loss": 0.6824, + "step": 3125 + }, + { + "epoch": 5.978157644824312, + "grad_norm": 38.43061828613281, + "learning_rate": 1.5582964368543115e-05, + "loss": 0.7002, + "step": 3150 + }, + { + "epoch": 6.0, + "eval_accuracy": 0.15438407746100055, + "eval_loss": 0.9505584836006165, + "eval_macro_f1": 0.5611208876450059, + "eval_micro_f1": 0.574826560951437, + "eval_runtime": 4.8418, + "eval_samples_per_second": 383.951, + "eval_steps_per_second": 24.165, + "eval_weighted_f1": 0.5844921772307327, + "step": 3162 + }, + { + "epoch": 6.0246913580246915, + "grad_norm": 18.99205780029297, + "learning_rate": 1.553025511279781e-05, + "loss": 0.5742, + "step": 3175 + }, + { + "epoch": 6.072174738841405, + "grad_norm": 9.61410903930664, + "learning_rate": 1.5477545857052502e-05, + "loss": 0.481, + "step": 3200 + }, + { + "epoch": 6.119658119658119, + "grad_norm": 16.063552856445312, + "learning_rate": 1.5424836601307192e-05, + "loss": 0.4474, + "step": 3225 + }, + { + "epoch": 6.167141500474834, + "grad_norm": 17.643842697143555, + "learning_rate": 1.5372127345561882e-05, + "loss": 0.4736, + "step": 3250 + }, + { + "epoch": 6.214624881291548, + "grad_norm": 41.78548049926758, + "learning_rate": 1.5319418089816572e-05, + "loss": 0.4752, + "step": 3275 + }, + { + "epoch": 6.262108262108262, + "grad_norm": 28.454172134399414, + "learning_rate": 1.5266708834071265e-05, + "loss": 0.4468, + "step": 3300 + }, + { + "epoch": 6.3095916429249765, + "grad_norm": 11.904889106750488, + "learning_rate": 1.5213999578325956e-05, + "loss": 0.4544, + "step": 3325 + }, + { + "epoch": 6.35707502374169, + "grad_norm": 16.409883499145508, + "learning_rate": 1.5161290322580646e-05, + "loss": 0.4544, + "step": 3350 + }, + { + "epoch": 6.404558404558404, + "grad_norm": 9.785941123962402, + "learning_rate": 1.5108581066835337e-05, + "loss": 0.4191, + "step": 3375 + }, + { + "epoch": 6.452041785375119, + "grad_norm": 32.8699836730957, + "learning_rate": 1.5055871811090027e-05, + "loss": 0.425, + "step": 3400 + }, + { + "epoch": 6.499525166191833, + "grad_norm": 32.70186233520508, + "learning_rate": 1.5003162555344719e-05, + "loss": 0.4595, + "step": 3425 + }, + { + "epoch": 6.547008547008547, + "grad_norm": 35.721683502197266, + "learning_rate": 1.4950453299599411e-05, + "loss": 0.4743, + "step": 3450 + }, + { + "epoch": 6.594491927825262, + "grad_norm": 12.760897636413574, + "learning_rate": 1.4897744043854101e-05, + "loss": 0.4509, + "step": 3475 + }, + { + "epoch": 6.6419753086419755, + "grad_norm": 10.61234188079834, + "learning_rate": 1.4845034788108794e-05, + "loss": 0.4619, + "step": 3500 + }, + { + "epoch": 6.689458689458689, + "grad_norm": 22.560327529907227, + "learning_rate": 1.4792325532363484e-05, + "loss": 0.4604, + "step": 3525 + }, + { + "epoch": 6.736942070275403, + "grad_norm": 15.491840362548828, + "learning_rate": 1.4739616276618176e-05, + "loss": 0.4701, + "step": 3550 + }, + { + "epoch": 6.784425451092118, + "grad_norm": 37.8037223815918, + "learning_rate": 1.4686907020872866e-05, + "loss": 0.4648, + "step": 3575 + }, + { + "epoch": 6.831908831908832, + "grad_norm": 16.956642150878906, + "learning_rate": 1.4634197765127558e-05, + "loss": 0.4579, + "step": 3600 + }, + { + "epoch": 6.879392212725546, + "grad_norm": 14.967536926269531, + "learning_rate": 1.458148850938225e-05, + "loss": 0.4411, + "step": 3625 + }, + { + "epoch": 6.926875593542261, + "grad_norm": 9.597579956054688, + "learning_rate": 1.452877925363694e-05, + "loss": 0.445, + "step": 3650 + }, + { + "epoch": 6.9743589743589745, + "grad_norm": 27.15288734436035, + "learning_rate": 1.4476069997891631e-05, + "loss": 0.4283, + "step": 3675 + }, + { + "epoch": 7.0, + "eval_accuracy": 0.18450779989241528, + "eval_loss": 1.1455060243606567, + "eval_macro_f1": 0.5782723196884059, + "eval_micro_f1": 0.5810215501845458, + "eval_runtime": 4.9967, + "eval_samples_per_second": 372.048, + "eval_steps_per_second": 23.416, + "eval_weighted_f1": 0.5939735287884271, + "step": 3689 + }, + { + "epoch": 7.020892687559354, + "grad_norm": 15.282934188842773, + "learning_rate": 1.4423360742146321e-05, + "loss": 0.3855, + "step": 3700 + }, + { + "epoch": 7.068376068376068, + "grad_norm": 11.879530906677246, + "learning_rate": 1.4370651486401013e-05, + "loss": 0.3347, + "step": 3725 + }, + { + "epoch": 7.115859449192783, + "grad_norm": 9.561882019042969, + "learning_rate": 1.4317942230655703e-05, + "loss": 0.3209, + "step": 3750 + }, + { + "epoch": 7.163342830009497, + "grad_norm": 13.189828872680664, + "learning_rate": 1.4265232974910395e-05, + "loss": 0.3162, + "step": 3775 + }, + { + "epoch": 7.210826210826211, + "grad_norm": 7.225050926208496, + "learning_rate": 1.4212523719165085e-05, + "loss": 0.3047, + "step": 3800 + }, + { + "epoch": 7.258309591642925, + "grad_norm": 9.511372566223145, + "learning_rate": 1.4159814463419779e-05, + "loss": 0.2889, + "step": 3825 + }, + { + "epoch": 7.305792972459639, + "grad_norm": 9.01153564453125, + "learning_rate": 1.410710520767447e-05, + "loss": 0.2788, + "step": 3850 + }, + { + "epoch": 7.353276353276353, + "grad_norm": 16.266386032104492, + "learning_rate": 1.405439595192916e-05, + "loss": 0.3126, + "step": 3875 + }, + { + "epoch": 7.400759734093067, + "grad_norm": 17.959672927856445, + "learning_rate": 1.4001686696183852e-05, + "loss": 0.2955, + "step": 3900 + }, + { + "epoch": 7.448243114909782, + "grad_norm": 18.825002670288086, + "learning_rate": 1.3948977440438542e-05, + "loss": 0.2955, + "step": 3925 + }, + { + "epoch": 7.495726495726496, + "grad_norm": 16.141721725463867, + "learning_rate": 1.3896268184693234e-05, + "loss": 0.2851, + "step": 3950 + }, + { + "epoch": 7.54320987654321, + "grad_norm": 19.40777587890625, + "learning_rate": 1.3843558928947924e-05, + "loss": 0.2869, + "step": 3975 + }, + { + "epoch": 7.590693257359924, + "grad_norm": 7.993305206298828, + "learning_rate": 1.3790849673202616e-05, + "loss": 0.3241, + "step": 4000 + }, + { + "epoch": 7.638176638176638, + "grad_norm": 24.663816452026367, + "learning_rate": 1.3738140417457306e-05, + "loss": 0.2923, + "step": 4025 + }, + { + "epoch": 7.685660018993352, + "grad_norm": 33.08135986328125, + "learning_rate": 1.3685431161711997e-05, + "loss": 0.2951, + "step": 4050 + }, + { + "epoch": 7.733143399810066, + "grad_norm": 11.239387512207031, + "learning_rate": 1.3632721905966687e-05, + "loss": 0.3102, + "step": 4075 + }, + { + "epoch": 7.780626780626781, + "grad_norm": 29.755128860473633, + "learning_rate": 1.358001265022138e-05, + "loss": 0.3141, + "step": 4100 + }, + { + "epoch": 7.828110161443495, + "grad_norm": 19.09785270690918, + "learning_rate": 1.3527303394476071e-05, + "loss": 0.3105, + "step": 4125 + }, + { + "epoch": 7.875593542260209, + "grad_norm": 18.331241607666016, + "learning_rate": 1.3474594138730763e-05, + "loss": 0.3177, + "step": 4150 + }, + { + "epoch": 7.923076923076923, + "grad_norm": 12.62179183959961, + "learning_rate": 1.3421884882985454e-05, + "loss": 0.2848, + "step": 4175 + }, + { + "epoch": 7.970560303893637, + "grad_norm": 45.56703186035156, + "learning_rate": 1.3369175627240144e-05, + "loss": 0.2963, + "step": 4200 + }, + { + "epoch": 8.0, + "eval_accuracy": 0.21732114039806347, + "eval_loss": 1.3072589635849, + "eval_macro_f1": 0.5765370791626004, + "eval_micro_f1": 0.5921717171717171, + "eval_runtime": 5.0047, + "eval_samples_per_second": 371.448, + "eval_steps_per_second": 23.378, + "eval_weighted_f1": 0.5926505086354816, + "step": 4216 + }, + { + "epoch": 8.017094017094017, + "grad_norm": 9.57081127166748, + "learning_rate": 1.3316466371494836e-05, + "loss": 0.2372, + "step": 4225 + }, + { + "epoch": 8.064577397910732, + "grad_norm": 6.422524452209473, + "learning_rate": 1.3263757115749526e-05, + "loss": 0.2301, + "step": 4250 + }, + { + "epoch": 8.112060778727445, + "grad_norm": 43.28707504272461, + "learning_rate": 1.3211047860004218e-05, + "loss": 0.1943, + "step": 4275 + }, + { + "epoch": 8.15954415954416, + "grad_norm": 13.590718269348145, + "learning_rate": 1.315833860425891e-05, + "loss": 0.2152, + "step": 4300 + }, + { + "epoch": 8.207027540360874, + "grad_norm": 15.272283554077148, + "learning_rate": 1.31056293485136e-05, + "loss": 0.2347, + "step": 4325 + }, + { + "epoch": 8.254510921177587, + "grad_norm": 7.845763206481934, + "learning_rate": 1.3052920092768292e-05, + "loss": 0.184, + "step": 4350 + }, + { + "epoch": 8.301994301994302, + "grad_norm": 6.99273681640625, + "learning_rate": 1.3000210837022982e-05, + "loss": 0.1949, + "step": 4375 + }, + { + "epoch": 8.349477682811017, + "grad_norm": 9.851614952087402, + "learning_rate": 1.2947501581277673e-05, + "loss": 0.2184, + "step": 4400 + }, + { + "epoch": 8.39696106362773, + "grad_norm": 21.555150985717773, + "learning_rate": 1.2894792325532363e-05, + "loss": 0.1959, + "step": 4425 + }, + { + "epoch": 8.444444444444445, + "grad_norm": 5.740550518035889, + "learning_rate": 1.2842083069787055e-05, + "loss": 0.1973, + "step": 4450 + }, + { + "epoch": 8.49192782526116, + "grad_norm": 51.750831604003906, + "learning_rate": 1.2789373814041749e-05, + "loss": 0.2322, + "step": 4475 + }, + { + "epoch": 8.539411206077872, + "grad_norm": 18.474130630493164, + "learning_rate": 1.2736664558296439e-05, + "loss": 0.2058, + "step": 4500 + }, + { + "epoch": 8.586894586894587, + "grad_norm": 14.638848304748535, + "learning_rate": 1.268395530255113e-05, + "loss": 0.1931, + "step": 4525 + }, + { + "epoch": 8.634377967711302, + "grad_norm": 20.490144729614258, + "learning_rate": 1.263124604680582e-05, + "loss": 0.2042, + "step": 4550 + }, + { + "epoch": 8.681861348528015, + "grad_norm": 10.135281562805176, + "learning_rate": 1.2578536791060512e-05, + "loss": 0.2113, + "step": 4575 + }, + { + "epoch": 8.72934472934473, + "grad_norm": 15.780980110168457, + "learning_rate": 1.2525827535315202e-05, + "loss": 0.2051, + "step": 4600 + }, + { + "epoch": 8.776828110161443, + "grad_norm": 36.281375885009766, + "learning_rate": 1.2473118279569894e-05, + "loss": 0.2284, + "step": 4625 + }, + { + "epoch": 8.824311490978157, + "grad_norm": 10.127740859985352, + "learning_rate": 1.2420409023824584e-05, + "loss": 0.1895, + "step": 4650 + }, + { + "epoch": 8.871794871794872, + "grad_norm": 24.48247718811035, + "learning_rate": 1.2367699768079276e-05, + "loss": 0.2153, + "step": 4675 + }, + { + "epoch": 8.919278252611585, + "grad_norm": 28.687623977661133, + "learning_rate": 1.2314990512333966e-05, + "loss": 0.2224, + "step": 4700 + }, + { + "epoch": 8.9667616334283, + "grad_norm": 23.58778190612793, + "learning_rate": 1.2262281256588657e-05, + "loss": 0.2009, + "step": 4725 + }, + { + "epoch": 9.0, + "eval_accuracy": 0.23991393222162452, + "eval_loss": 1.4571115970611572, + "eval_macro_f1": 0.5877494809891864, + "eval_micro_f1": 0.600444967936134, + "eval_runtime": 4.8234, + "eval_samples_per_second": 385.412, + "eval_steps_per_second": 24.257, + "eval_weighted_f1": 0.6004586928382261, + "step": 4743 + }, + { + "epoch": 9.01329534662868, + "grad_norm": 42.445255279541016, + "learning_rate": 1.2209572000843348e-05, + "loss": 0.1844, + "step": 4750 + }, + { + "epoch": 9.060778727445394, + "grad_norm": 10.499454498291016, + "learning_rate": 1.215686274509804e-05, + "loss": 0.1227, + "step": 4775 + }, + { + "epoch": 9.108262108262108, + "grad_norm": 74.01554107666016, + "learning_rate": 1.2104153489352733e-05, + "loss": 0.1515, + "step": 4800 + }, + { + "epoch": 9.155745489078823, + "grad_norm": 31.23480987548828, + "learning_rate": 1.2051444233607423e-05, + "loss": 0.1447, + "step": 4825 + }, + { + "epoch": 9.203228869895536, + "grad_norm": 7.121490478515625, + "learning_rate": 1.1998734977862115e-05, + "loss": 0.1555, + "step": 4850 + }, + { + "epoch": 9.250712250712251, + "grad_norm": 9.050018310546875, + "learning_rate": 1.1946025722116805e-05, + "loss": 0.1354, + "step": 4875 + }, + { + "epoch": 9.298195631528966, + "grad_norm": 7.522151947021484, + "learning_rate": 1.1893316466371496e-05, + "loss": 0.1349, + "step": 4900 + }, + { + "epoch": 9.345679012345679, + "grad_norm": 19.035409927368164, + "learning_rate": 1.1840607210626186e-05, + "loss": 0.1239, + "step": 4925 + }, + { + "epoch": 9.393162393162394, + "grad_norm": 7.271290302276611, + "learning_rate": 1.1787897954880878e-05, + "loss": 0.1152, + "step": 4950 + }, + { + "epoch": 9.440645773979107, + "grad_norm": 11.508015632629395, + "learning_rate": 1.173518869913557e-05, + "loss": 0.1185, + "step": 4975 + }, + { + "epoch": 9.488129154795821, + "grad_norm": 21.420225143432617, + "learning_rate": 1.168247944339026e-05, + "loss": 0.131, + "step": 5000 + }, + { + "epoch": 9.535612535612536, + "grad_norm": 21.15029525756836, + "learning_rate": 1.1629770187644952e-05, + "loss": 0.1384, + "step": 5025 + }, + { + "epoch": 9.583095916429249, + "grad_norm": 9.246801376342773, + "learning_rate": 1.1577060931899642e-05, + "loss": 0.1332, + "step": 5050 + }, + { + "epoch": 9.630579297245964, + "grad_norm": 5.654879093170166, + "learning_rate": 1.1524351676154333e-05, + "loss": 0.1607, + "step": 5075 + }, + { + "epoch": 9.678062678062679, + "grad_norm": 13.765711784362793, + "learning_rate": 1.1471642420409023e-05, + "loss": 0.1269, + "step": 5100 + }, + { + "epoch": 9.725546058879392, + "grad_norm": 11.771772384643555, + "learning_rate": 1.1418933164663717e-05, + "loss": 0.1225, + "step": 5125 + }, + { + "epoch": 9.773029439696106, + "grad_norm": 9.324225425720215, + "learning_rate": 1.1366223908918409e-05, + "loss": 0.1261, + "step": 5150 + }, + { + "epoch": 9.820512820512821, + "grad_norm": 12.03601360321045, + "learning_rate": 1.1313514653173099e-05, + "loss": 0.1357, + "step": 5175 + }, + { + "epoch": 9.867996201329534, + "grad_norm": 10.868319511413574, + "learning_rate": 1.126080539742779e-05, + "loss": 0.1235, + "step": 5200 + }, + { + "epoch": 9.915479582146249, + "grad_norm": 10.426804542541504, + "learning_rate": 1.120809614168248e-05, + "loss": 0.1277, + "step": 5225 + }, + { + "epoch": 9.962962962962964, + "grad_norm": 6.7754225730896, + "learning_rate": 1.1155386885937172e-05, + "loss": 0.1195, + "step": 5250 + }, + { + "epoch": 10.0, + "eval_accuracy": 0.2490586336740183, + "eval_loss": 1.572389841079712, + "eval_macro_f1": 0.5855009716659698, + "eval_micro_f1": 0.6038342941413058, + "eval_runtime": 4.9291, + "eval_samples_per_second": 377.152, + "eval_steps_per_second": 23.737, + "eval_weighted_f1": 0.6024200138375176, + "step": 5270 + }, + { + "epoch": 10.009496676163343, + "grad_norm": 4.07381010055542, + "learning_rate": 1.1102677630191862e-05, + "loss": 0.1133, + "step": 5275 + }, + { + "epoch": 10.056980056980057, + "grad_norm": 4.855506420135498, + "learning_rate": 1.1049968374446554e-05, + "loss": 0.0899, + "step": 5300 + }, + { + "epoch": 10.10446343779677, + "grad_norm": 6.390931129455566, + "learning_rate": 1.0997259118701244e-05, + "loss": 0.1037, + "step": 5325 + }, + { + "epoch": 10.151946818613485, + "grad_norm": 19.132612228393555, + "learning_rate": 1.0944549862955936e-05, + "loss": 0.0888, + "step": 5350 + }, + { + "epoch": 10.1994301994302, + "grad_norm": 7.906110763549805, + "learning_rate": 1.0891840607210626e-05, + "loss": 0.0909, + "step": 5375 + }, + { + "epoch": 10.246913580246913, + "grad_norm": 7.56940221786499, + "learning_rate": 1.0839131351465318e-05, + "loss": 0.1005, + "step": 5400 + }, + { + "epoch": 10.294396961063628, + "grad_norm": 17.0421142578125, + "learning_rate": 1.0786422095720008e-05, + "loss": 0.1084, + "step": 5425 + }, + { + "epoch": 10.341880341880342, + "grad_norm": 11.954485893249512, + "learning_rate": 1.0733712839974701e-05, + "loss": 0.0945, + "step": 5450 + }, + { + "epoch": 10.389363722697055, + "grad_norm": 5.614569187164307, + "learning_rate": 1.0681003584229393e-05, + "loss": 0.0911, + "step": 5475 + }, + { + "epoch": 10.43684710351377, + "grad_norm": 6.112778663635254, + "learning_rate": 1.0628294328484083e-05, + "loss": 0.0974, + "step": 5500 + }, + { + "epoch": 10.484330484330485, + "grad_norm": 24.94161033630371, + "learning_rate": 1.0575585072738775e-05, + "loss": 0.1222, + "step": 5525 + }, + { + "epoch": 10.531813865147198, + "grad_norm": 1.951548457145691, + "learning_rate": 1.0522875816993465e-05, + "loss": 0.1037, + "step": 5550 + }, + { + "epoch": 10.579297245963913, + "grad_norm": 4.477534294128418, + "learning_rate": 1.0470166561248156e-05, + "loss": 0.0811, + "step": 5575 + }, + { + "epoch": 10.626780626780628, + "grad_norm": 9.41921329498291, + "learning_rate": 1.0417457305502846e-05, + "loss": 0.0802, + "step": 5600 + }, + { + "epoch": 10.67426400759734, + "grad_norm": 13.954259872436523, + "learning_rate": 1.0364748049757538e-05, + "loss": 0.1071, + "step": 5625 + }, + { + "epoch": 10.721747388414055, + "grad_norm": 5.14282751083374, + "learning_rate": 1.031203879401223e-05, + "loss": 0.0902, + "step": 5650 + }, + { + "epoch": 10.76923076923077, + "grad_norm": 6.13380765914917, + "learning_rate": 1.025932953826692e-05, + "loss": 0.0817, + "step": 5675 + }, + { + "epoch": 10.816714150047483, + "grad_norm": 4.080665111541748, + "learning_rate": 1.0206620282521612e-05, + "loss": 0.0872, + "step": 5700 + }, + { + "epoch": 10.864197530864198, + "grad_norm": 12.922618865966797, + "learning_rate": 1.0153911026776302e-05, + "loss": 0.0816, + "step": 5725 + }, + { + "epoch": 10.91168091168091, + "grad_norm": 24.40860939025879, + "learning_rate": 1.0101201771030993e-05, + "loss": 0.0875, + "step": 5750 + }, + { + "epoch": 10.959164292497626, + "grad_norm": 8.776957511901855, + "learning_rate": 1.0048492515285685e-05, + "loss": 0.0789, + "step": 5775 + }, + { + "epoch": 11.0, + "eval_accuracy": 0.2528240989779451, + "eval_loss": 1.7437480688095093, + "eval_macro_f1": 0.5863809123209947, + "eval_micro_f1": 0.6008548186957121, + "eval_runtime": 4.8242, + "eval_samples_per_second": 385.349, + "eval_steps_per_second": 24.253, + "eval_weighted_f1": 0.599604351453644, + "step": 5797 + }, + { + "epoch": 11.005698005698006, + "grad_norm": 9.598529815673828, + "learning_rate": 9.995783259540375e-06, + "loss": 0.1007, + "step": 5800 + }, + { + "epoch": 11.05318138651472, + "grad_norm": 17.230466842651367, + "learning_rate": 9.943074003795067e-06, + "loss": 0.0626, + "step": 5825 + }, + { + "epoch": 11.100664767331434, + "grad_norm": 4.226415634155273, + "learning_rate": 9.890364748049759e-06, + "loss": 0.0668, + "step": 5850 + }, + { + "epoch": 11.148148148148149, + "grad_norm": 4.840086460113525, + "learning_rate": 9.83765549230445e-06, + "loss": 0.0613, + "step": 5875 + }, + { + "epoch": 11.195631528964862, + "grad_norm": 7.025901794433594, + "learning_rate": 9.78494623655914e-06, + "loss": 0.0671, + "step": 5900 + }, + { + "epoch": 11.243114909781577, + "grad_norm": 4.434010982513428, + "learning_rate": 9.732236980813832e-06, + "loss": 0.0698, + "step": 5925 + }, + { + "epoch": 11.290598290598291, + "grad_norm": 54.92803955078125, + "learning_rate": 9.679527725068522e-06, + "loss": 0.08, + "step": 5950 + }, + { + "epoch": 11.338081671415004, + "grad_norm": 6.1089630126953125, + "learning_rate": 9.626818469323214e-06, + "loss": 0.0519, + "step": 5975 + }, + { + "epoch": 11.385565052231719, + "grad_norm": 2.2659831047058105, + "learning_rate": 9.574109213577904e-06, + "loss": 0.0487, + "step": 6000 + }, + { + "epoch": 11.433048433048434, + "grad_norm": 3.403841257095337, + "learning_rate": 9.521399957832596e-06, + "loss": 0.0663, + "step": 6025 + }, + { + "epoch": 11.480531813865147, + "grad_norm": 3.4662554264068604, + "learning_rate": 9.468690702087288e-06, + "loss": 0.0663, + "step": 6050 + }, + { + "epoch": 11.528015194681862, + "grad_norm": 4.2505035400390625, + "learning_rate": 9.41598144634198e-06, + "loss": 0.0489, + "step": 6075 + }, + { + "epoch": 11.575498575498575, + "grad_norm": 11.110034942626953, + "learning_rate": 9.36327219059667e-06, + "loss": 0.0618, + "step": 6100 + }, + { + "epoch": 11.62298195631529, + "grad_norm": 4.673676490783691, + "learning_rate": 9.310562934851361e-06, + "loss": 0.0645, + "step": 6125 + }, + { + "epoch": 11.670465337132004, + "grad_norm": 4.6900553703308105, + "learning_rate": 9.257853679106051e-06, + "loss": 0.0614, + "step": 6150 + }, + { + "epoch": 11.717948717948717, + "grad_norm": 3.712474822998047, + "learning_rate": 9.205144423360743e-06, + "loss": 0.0643, + "step": 6175 + }, + { + "epoch": 11.765432098765432, + "grad_norm": 9.34663200378418, + "learning_rate": 9.152435167615435e-06, + "loss": 0.0638, + "step": 6200 + }, + { + "epoch": 11.812915479582147, + "grad_norm": 5.74763298034668, + "learning_rate": 9.099725911870125e-06, + "loss": 0.0519, + "step": 6225 + }, + { + "epoch": 11.86039886039886, + "grad_norm": 10.277482032775879, + "learning_rate": 9.047016656124816e-06, + "loss": 0.0488, + "step": 6250 + }, + { + "epoch": 11.907882241215574, + "grad_norm": 3.5318892002105713, + "learning_rate": 8.994307400379507e-06, + "loss": 0.053, + "step": 6275 + }, + { + "epoch": 11.95536562203229, + "grad_norm": 8.122844696044922, + "learning_rate": 8.941598144634198e-06, + "loss": 0.0692, + "step": 6300 + }, + { + "epoch": 12.0, + "eval_accuracy": 0.24529316837009144, + "eval_loss": 1.8085319995880127, + "eval_macro_f1": 0.5892817059465624, + "eval_micro_f1": 0.6006609749380336, + "eval_runtime": 4.9571, + "eval_samples_per_second": 375.019, + "eval_steps_per_second": 23.603, + "eval_weighted_f1": 0.6017616341149384, + "step": 6324 + }, + { + "epoch": 12.001899335232668, + "grad_norm": 1.5765128135681152, + "learning_rate": 8.888888888888888e-06, + "loss": 0.0597, + "step": 6325 + }, + { + "epoch": 12.049382716049383, + "grad_norm": 14.969993591308594, + "learning_rate": 8.836179633143582e-06, + "loss": 0.0483, + "step": 6350 + }, + { + "epoch": 12.096866096866098, + "grad_norm": 21.624101638793945, + "learning_rate": 8.783470377398272e-06, + "loss": 0.0435, + "step": 6375 + }, + { + "epoch": 12.14434947768281, + "grad_norm": 2.6236233711242676, + "learning_rate": 8.730761121652964e-06, + "loss": 0.0409, + "step": 6400 + }, + { + "epoch": 12.191832858499525, + "grad_norm": 2.5968286991119385, + "learning_rate": 8.678051865907654e-06, + "loss": 0.0423, + "step": 6425 + }, + { + "epoch": 12.239316239316238, + "grad_norm": 1.7756624221801758, + "learning_rate": 8.625342610162345e-06, + "loss": 0.0474, + "step": 6450 + }, + { + "epoch": 12.286799620132953, + "grad_norm": 3.3645403385162354, + "learning_rate": 8.572633354417035e-06, + "loss": 0.0349, + "step": 6475 + }, + { + "epoch": 12.334283000949668, + "grad_norm": 1.835996150970459, + "learning_rate": 8.519924098671729e-06, + "loss": 0.0396, + "step": 6500 + }, + { + "epoch": 12.381766381766381, + "grad_norm": 2.93648624420166, + "learning_rate": 8.467214842926419e-06, + "loss": 0.0472, + "step": 6525 + }, + { + "epoch": 12.429249762583096, + "grad_norm": 23.154329299926758, + "learning_rate": 8.41450558718111e-06, + "loss": 0.0563, + "step": 6550 + }, + { + "epoch": 12.47673314339981, + "grad_norm": 49.2928352355957, + "learning_rate": 8.3617963314358e-06, + "loss": 0.0407, + "step": 6575 + }, + { + "epoch": 12.524216524216524, + "grad_norm": 2.346856117248535, + "learning_rate": 8.309087075690492e-06, + "loss": 0.0343, + "step": 6600 + }, + { + "epoch": 12.571699905033238, + "grad_norm": 19.43263816833496, + "learning_rate": 8.256377819945182e-06, + "loss": 0.0479, + "step": 6625 + }, + { + "epoch": 12.619183285849953, + "grad_norm": 9.567797660827637, + "learning_rate": 8.203668564199874e-06, + "loss": 0.0333, + "step": 6650 + }, + { + "epoch": 12.666666666666666, + "grad_norm": 3.1313788890838623, + "learning_rate": 8.150959308454566e-06, + "loss": 0.036, + "step": 6675 + }, + { + "epoch": 12.71415004748338, + "grad_norm": 2.256866455078125, + "learning_rate": 8.098250052709256e-06, + "loss": 0.0401, + "step": 6700 + }, + { + "epoch": 12.761633428300096, + "grad_norm": 4.29447603225708, + "learning_rate": 8.045540796963948e-06, + "loss": 0.0387, + "step": 6725 + }, + { + "epoch": 12.809116809116809, + "grad_norm": 9.302247047424316, + "learning_rate": 7.992831541218638e-06, + "loss": 0.0336, + "step": 6750 + }, + { + "epoch": 12.856600189933523, + "grad_norm": 1.325979232788086, + "learning_rate": 7.94012228547333e-06, + "loss": 0.0354, + "step": 6775 + }, + { + "epoch": 12.904083570750238, + "grad_norm": 6.64365816116333, + "learning_rate": 7.887413029728021e-06, + "loss": 0.0313, + "step": 6800 + }, + { + "epoch": 12.951566951566951, + "grad_norm": 12.782024383544922, + "learning_rate": 7.834703773982713e-06, + "loss": 0.0392, + "step": 6825 + }, + { + "epoch": 12.999050332383666, + "grad_norm": 4.69736385345459, + "learning_rate": 7.781994518237403e-06, + "loss": 0.0347, + "step": 6850 + }, + { + "epoch": 13.0, + "eval_accuracy": 0.2528240989779451, + "eval_loss": 1.9986724853515625, + "eval_macro_f1": 0.5857716436175565, + "eval_micro_f1": 0.6015121814617754, + "eval_runtime": 4.8425, + "eval_samples_per_second": 383.894, + "eval_steps_per_second": 24.161, + "eval_weighted_f1": 0.6002283261093951, + "step": 6851 + }, + { + "epoch": 13.045584045584045, + "grad_norm": 0.9540898203849792, + "learning_rate": 7.729285262492095e-06, + "loss": 0.033, + "step": 6875 + }, + { + "epoch": 13.09306742640076, + "grad_norm": 2.6230499744415283, + "learning_rate": 7.676576006746785e-06, + "loss": 0.0225, + "step": 6900 + }, + { + "epoch": 13.140550807217474, + "grad_norm": 12.984963417053223, + "learning_rate": 7.6238667510014765e-06, + "loss": 0.0354, + "step": 6925 + }, + { + "epoch": 13.188034188034187, + "grad_norm": 2.323141098022461, + "learning_rate": 7.5711574952561674e-06, + "loss": 0.0253, + "step": 6950 + }, + { + "epoch": 13.235517568850902, + "grad_norm": 1.5419995784759521, + "learning_rate": 7.518448239510858e-06, + "loss": 0.0342, + "step": 6975 + }, + { + "epoch": 13.283000949667617, + "grad_norm": 1.283522605895996, + "learning_rate": 7.46573898376555e-06, + "loss": 0.0282, + "step": 7000 + }, + { + "epoch": 13.33048433048433, + "grad_norm": 2.585820198059082, + "learning_rate": 7.413029728020241e-06, + "loss": 0.0289, + "step": 7025 + }, + { + "epoch": 13.377967711301045, + "grad_norm": 1.7829973697662354, + "learning_rate": 7.360320472274932e-06, + "loss": 0.0248, + "step": 7050 + }, + { + "epoch": 13.42545109211776, + "grad_norm": 5.49683141708374, + "learning_rate": 7.307611216529623e-06, + "loss": 0.0264, + "step": 7075 + }, + { + "epoch": 13.472934472934472, + "grad_norm": 16.981597900390625, + "learning_rate": 7.2549019607843145e-06, + "loss": 0.0332, + "step": 7100 + }, + { + "epoch": 13.520417853751187, + "grad_norm": 3.749901294708252, + "learning_rate": 7.202192705039005e-06, + "loss": 0.0457, + "step": 7125 + }, + { + "epoch": 13.567901234567902, + "grad_norm": 2.465292453765869, + "learning_rate": 7.149483449293697e-06, + "loss": 0.0325, + "step": 7150 + }, + { + "epoch": 13.615384615384615, + "grad_norm": 1.3002429008483887, + "learning_rate": 7.096774193548388e-06, + "loss": 0.0242, + "step": 7175 + }, + { + "epoch": 13.66286799620133, + "grad_norm": 2.6421475410461426, + "learning_rate": 7.044064937803079e-06, + "loss": 0.0301, + "step": 7200 + }, + { + "epoch": 13.710351377018045, + "grad_norm": 2.1891891956329346, + "learning_rate": 6.99135568205777e-06, + "loss": 0.0276, + "step": 7225 + }, + { + "epoch": 13.757834757834758, + "grad_norm": 0.8588960766792297, + "learning_rate": 6.938646426312461e-06, + "loss": 0.0233, + "step": 7250 + }, + { + "epoch": 13.805318138651472, + "grad_norm": 4.041521072387695, + "learning_rate": 6.885937170567152e-06, + "loss": 0.0348, + "step": 7275 + }, + { + "epoch": 13.852801519468187, + "grad_norm": 6.659054756164551, + "learning_rate": 6.8332279148218425e-06, + "loss": 0.024, + "step": 7300 + }, + { + "epoch": 13.9002849002849, + "grad_norm": 1.0510109663009644, + "learning_rate": 6.780518659076535e-06, + "loss": 0.0256, + "step": 7325 + }, + { + "epoch": 13.947768281101615, + "grad_norm": 15.406808853149414, + "learning_rate": 6.727809403331226e-06, + "loss": 0.0242, + "step": 7350 + }, + { + "epoch": 13.995251661918328, + "grad_norm": 4.5207719802856445, + "learning_rate": 6.675100147585917e-06, + "loss": 0.0228, + "step": 7375 + }, + { + "epoch": 14.0, + "eval_accuracy": 0.2608929532006455, + "eval_loss": 2.1763834953308105, + "eval_macro_f1": 0.5897944419444678, + "eval_micro_f1": 0.6075230660042583, + "eval_runtime": 5.3544, + "eval_samples_per_second": 347.189, + "eval_steps_per_second": 21.851, + "eval_weighted_f1": 0.6043031140928729, + "step": 7378 + }, + { + "epoch": 14.041785375118709, + "grad_norm": 1.612322211265564, + "learning_rate": 6.622390891840608e-06, + "loss": 0.0173, + "step": 7400 + }, + { + "epoch": 14.089268755935423, + "grad_norm": 1.179650068283081, + "learning_rate": 6.569681636095299e-06, + "loss": 0.0227, + "step": 7425 + }, + { + "epoch": 14.136752136752136, + "grad_norm": 1.2343794107437134, + "learning_rate": 6.5169723803499896e-06, + "loss": 0.0159, + "step": 7450 + }, + { + "epoch": 14.184235517568851, + "grad_norm": 4.791398048400879, + "learning_rate": 6.464263124604681e-06, + "loss": 0.0174, + "step": 7475 + }, + { + "epoch": 14.231718898385566, + "grad_norm": 5.327462673187256, + "learning_rate": 6.411553868859372e-06, + "loss": 0.018, + "step": 7500 + }, + { + "epoch": 14.279202279202279, + "grad_norm": 2.131273031234741, + "learning_rate": 6.358844613114064e-06, + "loss": 0.0195, + "step": 7525 + }, + { + "epoch": 14.326685660018994, + "grad_norm": 2.609938144683838, + "learning_rate": 6.306135357368755e-06, + "loss": 0.0174, + "step": 7550 + }, + { + "epoch": 14.374169040835708, + "grad_norm": 1.8894281387329102, + "learning_rate": 6.253426101623446e-06, + "loss": 0.0189, + "step": 7575 + }, + { + "epoch": 14.421652421652421, + "grad_norm": 0.8751557469367981, + "learning_rate": 6.200716845878137e-06, + "loss": 0.0179, + "step": 7600 + }, + { + "epoch": 14.469135802469136, + "grad_norm": 1.151068091392517, + "learning_rate": 6.1480075901328275e-06, + "loss": 0.0149, + "step": 7625 + }, + { + "epoch": 14.51661918328585, + "grad_norm": 0.7474121451377869, + "learning_rate": 6.095298334387519e-06, + "loss": 0.0153, + "step": 7650 + }, + { + "epoch": 14.564102564102564, + "grad_norm": 0.7535553574562073, + "learning_rate": 6.04258907864221e-06, + "loss": 0.0149, + "step": 7675 + }, + { + "epoch": 14.611585944919279, + "grad_norm": 0.6468724012374878, + "learning_rate": 5.989879822896901e-06, + "loss": 0.018, + "step": 7700 + }, + { + "epoch": 14.659069325735992, + "grad_norm": 4.280311584472656, + "learning_rate": 5.937170567151592e-06, + "loss": 0.024, + "step": 7725 + }, + { + "epoch": 14.706552706552706, + "grad_norm": 1.5876237154006958, + "learning_rate": 5.884461311406283e-06, + "loss": 0.0168, + "step": 7750 + }, + { + "epoch": 14.754036087369421, + "grad_norm": 2.2202248573303223, + "learning_rate": 5.831752055660974e-06, + "loss": 0.0151, + "step": 7775 + }, + { + "epoch": 14.801519468186134, + "grad_norm": 0.5981371402740479, + "learning_rate": 5.779042799915666e-06, + "loss": 0.0233, + "step": 7800 + }, + { + "epoch": 14.849002849002849, + "grad_norm": 1.3764312267303467, + "learning_rate": 5.726333544170357e-06, + "loss": 0.0235, + "step": 7825 + }, + { + "epoch": 14.896486229819564, + "grad_norm": 1.4809439182281494, + "learning_rate": 5.673624288425048e-06, + "loss": 0.0152, + "step": 7850 + }, + { + "epoch": 14.943969610636277, + "grad_norm": 3.3015847206115723, + "learning_rate": 5.620915032679739e-06, + "loss": 0.0201, + "step": 7875 + }, + { + "epoch": 14.991452991452991, + "grad_norm": 2.4564335346221924, + "learning_rate": 5.56820577693443e-06, + "loss": 0.0139, + "step": 7900 + }, + { + "epoch": 15.0, + "eval_accuracy": 0.26358257127487894, + "eval_loss": 2.2441463470458984, + "eval_macro_f1": 0.5907126666859718, + "eval_micro_f1": 0.6064775007165377, + "eval_runtime": 5.8548, + "eval_samples_per_second": 317.519, + "eval_steps_per_second": 19.984, + "eval_weighted_f1": 0.6029898955448771, + "step": 7905 + } + ], + "logging_steps": 25, + "max_steps": 10540, + "num_input_tokens_seen": 0, + "num_train_epochs": 20, + "save_steps": 500, + "stateful_callbacks": { + "EarlyStoppingCallback": { + "args": { + "early_stopping_patience": 3, + "early_stopping_threshold": 0.0 + }, + "attributes": { + "early_stopping_patience_counter": 0 + } + }, + "TrainerControl": { + "args": { + "should_epoch_stop": false, + "should_evaluate": false, + "should_log": false, + "should_save": true, + "should_training_stop": false + }, + "attributes": {} + } + }, + "total_flos": 0.0, + "train_batch_size": 8, + "trial_name": null, + "trial_params": null +} diff --git a/ood-vigoemotions-mmbert-source-source-02-s44-20260905T152208Z-37344fee/trainer/checkpoint-9486/model.safetensors b/ood-vigoemotions-mmbert-source-source-02-s44-20260905T152208Z-37344fee/trainer/checkpoint-9486/model.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..d8c20a726f210442fd53732fb423124daddb74cf --- /dev/null +++ b/ood-vigoemotions-mmbert-source-source-02-s44-20260905T152208Z-37344fee/trainer/checkpoint-9486/model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:ae752a40417c7928572ec4e9e2f0a836dcc84deb12c0606541ba022ea7f3b031 +size 1227859136 diff --git a/ood-vigoemotions-mmbert-source-source-02-s44-20260905T152208Z-37344fee/trainer/checkpoint-9486/optimizer.pt b/ood-vigoemotions-mmbert-source-source-02-s44-20260905T152208Z-37344fee/trainer/checkpoint-9486/optimizer.pt new file mode 100644 index 0000000000000000000000000000000000000000..de5602dbeeaa2b9538913cc36072ca7c5c063738 --- /dev/null +++ b/ood-vigoemotions-mmbert-source-source-02-s44-20260905T152208Z-37344fee/trainer/checkpoint-9486/optimizer.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:eb628f9def673eb253b2976e44b55b65f11b7908f7e2f281c355a457ebe6b040 +size 2455803898 diff --git a/ood-vigoemotions-mmbert-source-source-02-s44-20260905T152208Z-37344fee/trainer/checkpoint-9486/special_tokens_map.json b/ood-vigoemotions-mmbert-source-source-02-s44-20260905T152208Z-37344fee/trainer/checkpoint-9486/special_tokens_map.json new file mode 100644 index 0000000000000000000000000000000000000000..be4ad795f49075ac5e56d8108ef63be314a1e967 --- /dev/null +++ b/ood-vigoemotions-mmbert-source-source-02-s44-20260905T152208Z-37344fee/trainer/checkpoint-9486/special_tokens_map.json @@ -0,0 +1,55 @@ +{ + "additional_special_tokens": [ + "", + "" + ], + "bos_token": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false + }, + "cls_token": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false + }, + "eos_token": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false + }, + "mask_token": { + "content": "", + "lstrip": true, + "normalized": false, + "rstrip": false, + "single_word": false + }, + "pad_token": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false + }, + "sep_token": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false + }, + "unk_token": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false + } +} diff --git a/ood-vigoemotions-mmbert-source-source-02-s44-20260905T152208Z-37344fee/trainer/checkpoint-9486/tokenizer_config.json b/ood-vigoemotions-mmbert-source-source-02-s44-20260905T152208Z-37344fee/trainer/checkpoint-9486/tokenizer_config.json new file mode 100644 index 0000000000000000000000000000000000000000..b002fb8c2960be8ebf9c373d6e5784b4565d7edc --- /dev/null +++ b/ood-vigoemotions-mmbert-source-source-02-s44-20260905T152208Z-37344fee/trainer/checkpoint-9486/tokenizer_config.json @@ -0,0 +1,2018 @@ +{ + "add_bos_token": true, + "added_tokens_decoder": { + "0": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "1": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "2": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "3": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "4": { + "content": "", + "lstrip": true, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "5": { + "content": "<2mass>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "6": { + "content": "[@BOS@]", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "7": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "8": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "9": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "10": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "11": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "12": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "13": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "14": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "15": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "16": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "17": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "18": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "19": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "20": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "21": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "22": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "23": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "24": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "25": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "26": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "27": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "28": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "29": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "30": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "31": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "32": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "33": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "34": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "35": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "36": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "37": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "38": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "39": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "40": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "41": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "42": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "43": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "44": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "45": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "46": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "47": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "48": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "49": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "50": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "51": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "52": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "53": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "54": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "55": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "56": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "57": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "58": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "59": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "60": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "61": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "62": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "63": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "64": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "65": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "66": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "67": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "68": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "69": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "70": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "71": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "72": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "73": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "74": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "75": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "76": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "77": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "78": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "79": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "80": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "81": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "82": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "83": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "84": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "85": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "86": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "87": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "88": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "89": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "90": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "91": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "92": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "93": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "94": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "95": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "96": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "97": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "98": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "99": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "100": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "101": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "102": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "103": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "104": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "105": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "106": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "107": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "108": { + "content": "\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "109": { + "content": "\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "110": { + "content": "\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "111": { + "content": "\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "112": { + "content": "\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "113": { + "content": "\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "114": { + "content": "\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "115": { + "content": "\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "116": { + "content": "\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "117": { + "content": "\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "118": { + "content": "\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "119": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "120": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "121": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "122": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "123": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "124": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "125": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "126": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "127": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "128": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "129": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "130": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "131": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "132": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "133": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "134": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "135": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "136": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "137": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "138": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "139": { + "content": "▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "140": { + "content": "▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "141": { + "content": "▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "142": { + "content": "▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "143": { + "content": "▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "144": { + "content": "▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "145": { + "content": "▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "146": { + "content": "▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "147": { + "content": "▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "148": { + "content": "▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "149": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "150": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "151": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "152": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "153": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "154": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "155": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "156": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "157": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "158": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "159": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "160": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "161": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "162": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "163": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "164": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "165": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "166": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "167": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "168": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "169": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "170": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "172": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "173": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "174": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "175": { + "content": "
", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "171": { + "content": "
", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "176": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "177": { + "content": "
", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "178": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "179": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "180": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "181": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "182": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "183": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "184": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "185": { + "content": "

", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "186": { + "content": "

", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "187": { + "content": "

", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "188": { + "content": "

", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "189": { + "content": "

", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "190": { + "content": "
", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "191": { + "content": "
", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "192": { + "content": "
", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "193": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "194": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "195": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "196": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "197": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "198": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "199": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "200": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "201": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "202": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "203": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "204": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "205": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "206": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "207": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "208": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "209": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "210": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "211": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "212": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "213": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "214": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "215": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "216": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255968": { + "content": "[toxicity=0]", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255969": { + "content": "\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255970": { + "content": "\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255971": { + "content": "\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255972": { + "content": "\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255973": { + "content": "\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255974": { + "content": "\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255975": { + "content": "\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255976": { + "content": "\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255977": { + "content": "\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255978": { + "content": "\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255979": { + "content": "\t\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255980": { + "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255981": { + "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255982": { + "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255983": { + "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255984": { + "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255985": { + "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255986": { + "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255987": { + "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255988": { + "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255989": { + "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255990": { + "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255991": { + "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255992": { + "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255993": { + "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255994": { + "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255995": { + "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255996": { + "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255997": { + "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255998": { + "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "255999": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + } + }, + "additional_special_tokens": [ + "", + "" + ], + "bos_token": "", + "clean_up_tokenization_spaces": false, + "cls_token": "", + "eos_token": "", + "extra_special_tokens": {}, + "mask_token": "", + "model_input_names": [ + "input_ids", + "attention_mask" + ], + "model_max_length": 8192, + "pad_token": "", + "padding_side": "right", + "sep_token": "", + "spaces_between_special_tokens": false, + "tokenizer_class": "PreTrainedTokenizerFast", + "unk_token": "" +} diff --git a/ood-vigoemotions-mmbert-source-source-02-s44-20260905T152208Z-37344fee/trainer/checkpoint-9486/trainer_state.json b/ood-vigoemotions-mmbert-source-source-02-s44-20260905T152208Z-37344fee/trainer/checkpoint-9486/trainer_state.json new file mode 100644 index 0000000000000000000000000000000000000000..49a0f9b50a64cb488f4761ed3f2edb522b89969b --- /dev/null +++ b/ood-vigoemotions-mmbert-source-source-02-s44-20260905T152208Z-37344fee/trainer/checkpoint-9486/trainer_state.json @@ -0,0 +1,2912 @@ +{ + "best_global_step": 7905, + "best_metric": 0.5907126666859718, + "best_model_checkpoint": "/root/VNEncoder/results/checkpoints/ood-vigoemotions-mmbert-source-source-02-s44-20260905T152208Z-37344fee/trainer/checkpoint-7905", + "epoch": 18.0, + "eval_steps": 500, + "global_step": 9486, + "is_hyper_param_search": false, + "is_local_process_zero": true, + "is_world_process_zero": true, + "log_history": [ + { + "epoch": 0.04748338081671415, + "grad_norm": 34.162723541259766, + "learning_rate": 4.554079696394687e-07, + "loss": 6.1444, + "step": 25 + }, + { + "epoch": 0.0949667616334283, + "grad_norm": 36.55830383300781, + "learning_rate": 9.297912713472485e-07, + "loss": 6.2258, + "step": 50 + }, + { + "epoch": 0.14245014245014245, + "grad_norm": 33.597232818603516, + "learning_rate": 1.4041745730550286e-06, + "loss": 6.2708, + "step": 75 + }, + { + "epoch": 0.1899335232668566, + "grad_norm": 39.089359283447266, + "learning_rate": 1.8785578747628086e-06, + "loss": 6.2674, + "step": 100 + }, + { + "epoch": 0.23741690408357075, + "grad_norm": 30.89531135559082, + "learning_rate": 2.3529411764705885e-06, + "loss": 6.2874, + "step": 125 + }, + { + "epoch": 0.2849002849002849, + "grad_norm": 58.9684944152832, + "learning_rate": 2.8273244781783683e-06, + "loss": 6.0477, + "step": 150 + }, + { + "epoch": 0.33238366571699907, + "grad_norm": 142.00091552734375, + "learning_rate": 3.3017077798861485e-06, + "loss": 5.3108, + "step": 175 + }, + { + "epoch": 0.3798670465337132, + "grad_norm": 32.22747039794922, + "learning_rate": 3.7760910815939284e-06, + "loss": 5.1682, + "step": 200 + }, + { + "epoch": 0.42735042735042733, + "grad_norm": 30.62629508972168, + "learning_rate": 4.250474383301708e-06, + "loss": 4.9753, + "step": 225 + }, + { + "epoch": 0.4748338081671415, + "grad_norm": 52.48829650878906, + "learning_rate": 4.724857685009488e-06, + "loss": 4.8218, + "step": 250 + }, + { + "epoch": 0.5223171889838556, + "grad_norm": 48.24330139160156, + "learning_rate": 5.199240986717268e-06, + "loss": 4.5239, + "step": 275 + }, + { + "epoch": 0.5698005698005698, + "grad_norm": 79.2529296875, + "learning_rate": 5.673624288425048e-06, + "loss": 4.2647, + "step": 300 + }, + { + "epoch": 0.6172839506172839, + "grad_norm": 66.08687591552734, + "learning_rate": 6.1480075901328275e-06, + "loss": 4.1624, + "step": 325 + }, + { + "epoch": 0.6647673314339981, + "grad_norm": 66.99076080322266, + "learning_rate": 6.622390891840608e-06, + "loss": 3.8941, + "step": 350 + }, + { + "epoch": 0.7122507122507122, + "grad_norm": 51.146968841552734, + "learning_rate": 7.096774193548388e-06, + "loss": 3.8293, + "step": 375 + }, + { + "epoch": 0.7597340930674265, + "grad_norm": 37.30567932128906, + "learning_rate": 7.5711574952561674e-06, + "loss": 3.7049, + "step": 400 + }, + { + "epoch": 0.8072174738841406, + "grad_norm": 43.886505126953125, + "learning_rate": 8.045540796963948e-06, + "loss": 3.656, + "step": 425 + }, + { + "epoch": 0.8547008547008547, + "grad_norm": 42.3010368347168, + "learning_rate": 8.519924098671729e-06, + "loss": 3.6309, + "step": 450 + }, + { + "epoch": 0.9021842355175689, + "grad_norm": 45.34294509887695, + "learning_rate": 8.994307400379507e-06, + "loss": 3.4031, + "step": 475 + }, + { + "epoch": 0.949667616334283, + "grad_norm": 54.65419387817383, + "learning_rate": 9.468690702087288e-06, + "loss": 3.2897, + "step": 500 + }, + { + "epoch": 0.9971509971509972, + "grad_norm": 55.79121017456055, + "learning_rate": 9.943074003795067e-06, + "loss": 3.3332, + "step": 525 + }, + { + "epoch": 1.0, + "eval_accuracy": 0.0005379236148466917, + "eval_loss": 0.8047798871994019, + "eval_macro_f1": 0.3364472078136643, + "eval_micro_f1": 0.3479386208171566, + "eval_runtime": 5.6709, + "eval_samples_per_second": 327.814, + "eval_steps_per_second": 20.632, + "eval_weighted_f1": 0.40064642675810264, + "step": 527 + }, + { + "epoch": 1.043684710351377, + "grad_norm": 50.92436599731445, + "learning_rate": 1.0417457305502846e-05, + "loss": 3.0012, + "step": 550 + }, + { + "epoch": 1.0911680911680912, + "grad_norm": 51.1168327331543, + "learning_rate": 1.0891840607210626e-05, + "loss": 2.8467, + "step": 575 + }, + { + "epoch": 1.1386514719848053, + "grad_norm": 40.4506950378418, + "learning_rate": 1.1366223908918409e-05, + "loss": 2.8875, + "step": 600 + }, + { + "epoch": 1.1861348528015194, + "grad_norm": 46.12940979003906, + "learning_rate": 1.1840607210626186e-05, + "loss": 2.8581, + "step": 625 + }, + { + "epoch": 1.2336182336182335, + "grad_norm": 57.1807975769043, + "learning_rate": 1.2314990512333966e-05, + "loss": 2.8958, + "step": 650 + }, + { + "epoch": 1.2811016144349479, + "grad_norm": 63.2441520690918, + "learning_rate": 1.2789373814041749e-05, + "loss": 2.9974, + "step": 675 + }, + { + "epoch": 1.328584995251662, + "grad_norm": 35.746822357177734, + "learning_rate": 1.3263757115749526e-05, + "loss": 3.0096, + "step": 700 + }, + { + "epoch": 1.376068376068376, + "grad_norm": 140.2974395751953, + "learning_rate": 1.3738140417457306e-05, + "loss": 2.8375, + "step": 725 + }, + { + "epoch": 1.4235517568850902, + "grad_norm": 86.68020629882812, + "learning_rate": 1.4212523719165085e-05, + "loss": 2.8921, + "step": 750 + }, + { + "epoch": 1.4710351377018043, + "grad_norm": 55.23737335205078, + "learning_rate": 1.4686907020872866e-05, + "loss": 3.0217, + "step": 775 + }, + { + "epoch": 1.5185185185185186, + "grad_norm": 48.04639434814453, + "learning_rate": 1.5161290322580646e-05, + "loss": 2.7802, + "step": 800 + }, + { + "epoch": 1.5660018993352327, + "grad_norm": 31.351240158081055, + "learning_rate": 1.5635673624288425e-05, + "loss": 2.7824, + "step": 825 + }, + { + "epoch": 1.6134852801519468, + "grad_norm": 48.16146469116211, + "learning_rate": 1.6110056925996208e-05, + "loss": 2.7755, + "step": 850 + }, + { + "epoch": 1.660968660968661, + "grad_norm": 59.989097595214844, + "learning_rate": 1.6584440227703987e-05, + "loss": 2.8255, + "step": 875 + }, + { + "epoch": 1.708452041785375, + "grad_norm": 39.0864372253418, + "learning_rate": 1.7058823529411767e-05, + "loss": 2.6891, + "step": 900 + }, + { + "epoch": 1.7559354226020893, + "grad_norm": 35.95862579345703, + "learning_rate": 1.7533206831119546e-05, + "loss": 2.7272, + "step": 925 + }, + { + "epoch": 1.8034188034188035, + "grad_norm": 59.74723815917969, + "learning_rate": 1.8007590132827325e-05, + "loss": 2.6682, + "step": 950 + }, + { + "epoch": 1.8509021842355176, + "grad_norm": 37.617794036865234, + "learning_rate": 1.8481973434535105e-05, + "loss": 2.8645, + "step": 975 + }, + { + "epoch": 1.8983855650522319, + "grad_norm": 33.81266784667969, + "learning_rate": 1.8956356736242888e-05, + "loss": 2.7383, + "step": 1000 + }, + { + "epoch": 1.9458689458689458, + "grad_norm": 75.7305908203125, + "learning_rate": 1.9430740037950667e-05, + "loss": 2.7314, + "step": 1025 + }, + { + "epoch": 1.99335232668566, + "grad_norm": 33.35398864746094, + "learning_rate": 1.9905123339658446e-05, + "loss": 2.5902, + "step": 1050 + }, + { + "epoch": 2.0, + "eval_accuracy": 0.01775147928994083, + "eval_loss": 0.6891770958900452, + "eval_macro_f1": 0.41180038828201304, + "eval_micro_f1": 0.4208560758402758, + "eval_runtime": 4.9435, + "eval_samples_per_second": 376.051, + "eval_steps_per_second": 23.668, + "eval_weighted_f1": 0.4598716002150326, + "step": 1054 + }, + { + "epoch": 2.03988603988604, + "grad_norm": 71.47340393066406, + "learning_rate": 1.9957832595403753e-05, + "loss": 2.1893, + "step": 1075 + }, + { + "epoch": 2.087369420702754, + "grad_norm": 34.373592376708984, + "learning_rate": 1.9905123339658446e-05, + "loss": 2.1511, + "step": 1100 + }, + { + "epoch": 2.134852801519468, + "grad_norm": 18.752532958984375, + "learning_rate": 1.9852414083913136e-05, + "loss": 1.9717, + "step": 1125 + }, + { + "epoch": 2.1823361823361824, + "grad_norm": 40.68827819824219, + "learning_rate": 1.9799704828167827e-05, + "loss": 2.1369, + "step": 1150 + }, + { + "epoch": 2.2298195631528963, + "grad_norm": 29.010814666748047, + "learning_rate": 1.9746995572422517e-05, + "loss": 2.037, + "step": 1175 + }, + { + "epoch": 2.2773029439696106, + "grad_norm": 33.29954528808594, + "learning_rate": 1.969428631667721e-05, + "loss": 2.053, + "step": 1200 + }, + { + "epoch": 2.324786324786325, + "grad_norm": 32.961490631103516, + "learning_rate": 1.9641577060931903e-05, + "loss": 2.043, + "step": 1225 + }, + { + "epoch": 2.372269705603039, + "grad_norm": 61.129783630371094, + "learning_rate": 1.9588867805186593e-05, + "loss": 2.0912, + "step": 1250 + }, + { + "epoch": 2.419753086419753, + "grad_norm": 49.88624572753906, + "learning_rate": 1.9536158549441284e-05, + "loss": 1.995, + "step": 1275 + }, + { + "epoch": 2.467236467236467, + "grad_norm": 54.424903869628906, + "learning_rate": 1.9483449293695974e-05, + "loss": 1.9819, + "step": 1300 + }, + { + "epoch": 2.5147198480531814, + "grad_norm": 43.5580940246582, + "learning_rate": 1.9430740037950667e-05, + "loss": 2.0496, + "step": 1325 + }, + { + "epoch": 2.5622032288698957, + "grad_norm": 28.404251098632812, + "learning_rate": 1.9378030782205357e-05, + "loss": 2.0856, + "step": 1350 + }, + { + "epoch": 2.6096866096866096, + "grad_norm": 56.033836364746094, + "learning_rate": 1.9325321526460047e-05, + "loss": 2.1025, + "step": 1375 + }, + { + "epoch": 2.657169990503324, + "grad_norm": 30.688547134399414, + "learning_rate": 1.927261227071474e-05, + "loss": 2.0668, + "step": 1400 + }, + { + "epoch": 2.7046533713200382, + "grad_norm": 51.40027618408203, + "learning_rate": 1.921990301496943e-05, + "loss": 2.2163, + "step": 1425 + }, + { + "epoch": 2.752136752136752, + "grad_norm": 44.746917724609375, + "learning_rate": 1.916719375922412e-05, + "loss": 1.9534, + "step": 1450 + }, + { + "epoch": 2.799620132953466, + "grad_norm": 23.693490982055664, + "learning_rate": 1.911448450347881e-05, + "loss": 1.9513, + "step": 1475 + }, + { + "epoch": 2.8471035137701803, + "grad_norm": 40.34550094604492, + "learning_rate": 1.9061775247733504e-05, + "loss": 1.9943, + "step": 1500 + }, + { + "epoch": 2.8945868945868947, + "grad_norm": 43.02825927734375, + "learning_rate": 1.9009065991988194e-05, + "loss": 2.0602, + "step": 1525 + }, + { + "epoch": 2.9420702754036085, + "grad_norm": 42.245784759521484, + "learning_rate": 1.8956356736242888e-05, + "loss": 1.9576, + "step": 1550 + }, + { + "epoch": 2.989553656220323, + "grad_norm": 38.256168365478516, + "learning_rate": 1.8903647480497578e-05, + "loss": 1.9699, + "step": 1575 + }, + { + "epoch": 3.0, + "eval_accuracy": 0.0328133405056482, + "eval_loss": 0.6396704912185669, + "eval_macro_f1": 0.4600545960071464, + "eval_micro_f1": 0.46655478302565334, + "eval_runtime": 4.8346, + "eval_samples_per_second": 384.521, + "eval_steps_per_second": 24.201, + "eval_weighted_f1": 0.5101439509128096, + "step": 1581 + }, + { + "epoch": 3.0360873694207027, + "grad_norm": 20.20168685913086, + "learning_rate": 1.8850938224752268e-05, + "loss": 1.5347, + "step": 1600 + }, + { + "epoch": 3.083570750237417, + "grad_norm": 28.101213455200195, + "learning_rate": 1.879822896900696e-05, + "loss": 1.3689, + "step": 1625 + }, + { + "epoch": 3.131054131054131, + "grad_norm": 33.39235305786133, + "learning_rate": 1.874551971326165e-05, + "loss": 1.4377, + "step": 1650 + }, + { + "epoch": 3.178537511870845, + "grad_norm": 30.524776458740234, + "learning_rate": 1.869281045751634e-05, + "loss": 1.3264, + "step": 1675 + }, + { + "epoch": 3.2260208926875595, + "grad_norm": 26.046913146972656, + "learning_rate": 1.864010120177103e-05, + "loss": 1.4242, + "step": 1700 + }, + { + "epoch": 3.2735042735042734, + "grad_norm": 31.101991653442383, + "learning_rate": 1.8587391946025725e-05, + "loss": 1.3763, + "step": 1725 + }, + { + "epoch": 3.3209876543209877, + "grad_norm": 20.603694915771484, + "learning_rate": 1.8534682690280415e-05, + "loss": 1.3973, + "step": 1750 + }, + { + "epoch": 3.3684710351377016, + "grad_norm": 30.091970443725586, + "learning_rate": 1.8481973434535105e-05, + "loss": 1.3262, + "step": 1775 + }, + { + "epoch": 3.415954415954416, + "grad_norm": 30.08812713623047, + "learning_rate": 1.8429264178789795e-05, + "loss": 1.3596, + "step": 1800 + }, + { + "epoch": 3.4634377967711303, + "grad_norm": 28.665847778320312, + "learning_rate": 1.8376554923044488e-05, + "loss": 1.3754, + "step": 1825 + }, + { + "epoch": 3.510921177587844, + "grad_norm": 23.131080627441406, + "learning_rate": 1.832384566729918e-05, + "loss": 1.4043, + "step": 1850 + }, + { + "epoch": 3.5584045584045585, + "grad_norm": 15.830599784851074, + "learning_rate": 1.8271136411553872e-05, + "loss": 1.3662, + "step": 1875 + }, + { + "epoch": 3.6058879392212724, + "grad_norm": 26.26952362060547, + "learning_rate": 1.8218427155808562e-05, + "loss": 1.3602, + "step": 1900 + }, + { + "epoch": 3.6533713200379867, + "grad_norm": 31.234373092651367, + "learning_rate": 1.8165717900063252e-05, + "loss": 1.3459, + "step": 1925 + }, + { + "epoch": 3.700854700854701, + "grad_norm": 34.2337532043457, + "learning_rate": 1.8113008644317945e-05, + "loss": 1.4278, + "step": 1950 + }, + { + "epoch": 3.748338081671415, + "grad_norm": 36.79152297973633, + "learning_rate": 1.8060299388572635e-05, + "loss": 1.3326, + "step": 1975 + }, + { + "epoch": 3.7958214624881292, + "grad_norm": 40.74213409423828, + "learning_rate": 1.8007590132827325e-05, + "loss": 1.3891, + "step": 2000 + }, + { + "epoch": 3.8433048433048436, + "grad_norm": 100.57337188720703, + "learning_rate": 1.7954880877082015e-05, + "loss": 1.4088, + "step": 2025 + }, + { + "epoch": 3.8907882241215574, + "grad_norm": 23.349496841430664, + "learning_rate": 1.790217162133671e-05, + "loss": 1.3784, + "step": 2050 + }, + { + "epoch": 3.9382716049382713, + "grad_norm": 31.829254150390625, + "learning_rate": 1.78494623655914e-05, + "loss": 1.492, + "step": 2075 + }, + { + "epoch": 3.9857549857549857, + "grad_norm": 25.376209259033203, + "learning_rate": 1.779675310984609e-05, + "loss": 1.3584, + "step": 2100 + }, + { + "epoch": 4.0, + "eval_accuracy": 0.08230231307154384, + "eval_loss": 0.7609887719154358, + "eval_macro_f1": 0.5337177326494273, + "eval_micro_f1": 0.5383452965621458, + "eval_runtime": 4.899, + "eval_samples_per_second": 379.465, + "eval_steps_per_second": 23.882, + "eval_weighted_f1": 0.5553705232275817, + "step": 2108 + }, + { + "epoch": 4.032288698955366, + "grad_norm": 28.834596633911133, + "learning_rate": 1.7744043854100782e-05, + "loss": 1.0238, + "step": 2125 + }, + { + "epoch": 4.07977207977208, + "grad_norm": 27.219921112060547, + "learning_rate": 1.7691334598355472e-05, + "loss": 0.9589, + "step": 2150 + }, + { + "epoch": 4.127255460588794, + "grad_norm": 9.110738754272461, + "learning_rate": 1.7638625342610163e-05, + "loss": 0.94, + "step": 2175 + }, + { + "epoch": 4.174738841405508, + "grad_norm": 32.859806060791016, + "learning_rate": 1.7585916086864856e-05, + "loss": 0.9117, + "step": 2200 + }, + { + "epoch": 4.222222222222222, + "grad_norm": 22.447908401489258, + "learning_rate": 1.7533206831119546e-05, + "loss": 0.9674, + "step": 2225 + }, + { + "epoch": 4.269705603038936, + "grad_norm": 24.79280662536621, + "learning_rate": 1.748049757537424e-05, + "loss": 0.9744, + "step": 2250 + }, + { + "epoch": 4.317188983855651, + "grad_norm": 21.35533905029297, + "learning_rate": 1.742778831962893e-05, + "loss": 0.9193, + "step": 2275 + }, + { + "epoch": 4.364672364672365, + "grad_norm": 25.3253173828125, + "learning_rate": 1.737507906388362e-05, + "loss": 0.9302, + "step": 2300 + }, + { + "epoch": 4.412155745489079, + "grad_norm": 29.509456634521484, + "learning_rate": 1.732236980813831e-05, + "loss": 0.9541, + "step": 2325 + }, + { + "epoch": 4.459639126305793, + "grad_norm": 26.883155822753906, + "learning_rate": 1.7269660552393003e-05, + "loss": 0.9354, + "step": 2350 + }, + { + "epoch": 4.507122507122507, + "grad_norm": 44.8468132019043, + "learning_rate": 1.7216951296647693e-05, + "loss": 0.9985, + "step": 2375 + }, + { + "epoch": 4.554605887939221, + "grad_norm": 22.781490325927734, + "learning_rate": 1.7164242040902383e-05, + "loss": 1.0022, + "step": 2400 + }, + { + "epoch": 4.602089268755935, + "grad_norm": 29.72208023071289, + "learning_rate": 1.7111532785157073e-05, + "loss": 1.0015, + "step": 2425 + }, + { + "epoch": 4.64957264957265, + "grad_norm": 27.31267738342285, + "learning_rate": 1.7058823529411767e-05, + "loss": 0.8904, + "step": 2450 + }, + { + "epoch": 4.697056030389364, + "grad_norm": 23.065595626831055, + "learning_rate": 1.7006114273666457e-05, + "loss": 0.8698, + "step": 2475 + }, + { + "epoch": 4.744539411206078, + "grad_norm": 39.216705322265625, + "learning_rate": 1.6953405017921147e-05, + "loss": 0.9714, + "step": 2500 + }, + { + "epoch": 4.792022792022792, + "grad_norm": 45.64983367919922, + "learning_rate": 1.690069576217584e-05, + "loss": 0.8886, + "step": 2525 + }, + { + "epoch": 4.839506172839506, + "grad_norm": 24.30289649963379, + "learning_rate": 1.684798650643053e-05, + "loss": 0.9215, + "step": 2550 + }, + { + "epoch": 4.88698955365622, + "grad_norm": 55.504310607910156, + "learning_rate": 1.6795277250685224e-05, + "loss": 0.9372, + "step": 2575 + }, + { + "epoch": 4.934472934472934, + "grad_norm": 25.460933685302734, + "learning_rate": 1.6742567994939914e-05, + "loss": 0.9582, + "step": 2600 + }, + { + "epoch": 4.981956315289649, + "grad_norm": 24.881301879882812, + "learning_rate": 1.6689858739194604e-05, + "loss": 0.9422, + "step": 2625 + }, + { + "epoch": 5.0, + "eval_accuracy": 0.12641204948897256, + "eval_loss": 0.8357337713241577, + "eval_macro_f1": 0.5648626881399819, + "eval_micro_f1": 0.5698924731182796, + "eval_runtime": 4.9764, + "eval_samples_per_second": 373.561, + "eval_steps_per_second": 23.511, + "eval_weighted_f1": 0.584282614713785, + "step": 2635 + }, + { + "epoch": 5.028490028490029, + "grad_norm": 27.813451766967773, + "learning_rate": 1.6637149483449294e-05, + "loss": 0.7289, + "step": 2650 + }, + { + "epoch": 5.075973409306743, + "grad_norm": 21.82946014404297, + "learning_rate": 1.6584440227703987e-05, + "loss": 0.7127, + "step": 2675 + }, + { + "epoch": 5.1234567901234565, + "grad_norm": 19.29052734375, + "learning_rate": 1.6531730971958677e-05, + "loss": 0.6762, + "step": 2700 + }, + { + "epoch": 5.170940170940171, + "grad_norm": 42.652442932128906, + "learning_rate": 1.6479021716213367e-05, + "loss": 0.6371, + "step": 2725 + }, + { + "epoch": 5.218423551756885, + "grad_norm": 18.059337615966797, + "learning_rate": 1.642631246046806e-05, + "loss": 0.6415, + "step": 2750 + }, + { + "epoch": 5.265906932573599, + "grad_norm": 14.32496452331543, + "learning_rate": 1.637360320472275e-05, + "loss": 0.6583, + "step": 2775 + }, + { + "epoch": 5.313390313390314, + "grad_norm": 41.976932525634766, + "learning_rate": 1.632089394897744e-05, + "loss": 0.678, + "step": 2800 + }, + { + "epoch": 5.360873694207028, + "grad_norm": 38.39088439941406, + "learning_rate": 1.626818469323213e-05, + "loss": 0.6329, + "step": 2825 + }, + { + "epoch": 5.4083570750237415, + "grad_norm": 21.257617950439453, + "learning_rate": 1.6215475437486824e-05, + "loss": 0.6527, + "step": 2850 + }, + { + "epoch": 5.455840455840455, + "grad_norm": 15.757341384887695, + "learning_rate": 1.6162766181741514e-05, + "loss": 0.6476, + "step": 2875 + }, + { + "epoch": 5.50332383665717, + "grad_norm": 11.70419979095459, + "learning_rate": 1.6110056925996208e-05, + "loss": 0.6269, + "step": 2900 + }, + { + "epoch": 5.550807217473884, + "grad_norm": 44.25235366821289, + "learning_rate": 1.6057347670250898e-05, + "loss": 0.6537, + "step": 2925 + }, + { + "epoch": 5.598290598290598, + "grad_norm": 25.33641242980957, + "learning_rate": 1.6004638414505588e-05, + "loss": 0.6865, + "step": 2950 + }, + { + "epoch": 5.645773979107313, + "grad_norm": 16.503864288330078, + "learning_rate": 1.595192915876028e-05, + "loss": 0.6372, + "step": 2975 + }, + { + "epoch": 5.693257359924027, + "grad_norm": 21.774084091186523, + "learning_rate": 1.589921990301497e-05, + "loss": 0.6898, + "step": 3000 + }, + { + "epoch": 5.7407407407407405, + "grad_norm": 39.14348220825195, + "learning_rate": 1.584651064726966e-05, + "loss": 0.6671, + "step": 3025 + }, + { + "epoch": 5.788224121557455, + "grad_norm": 28.4456787109375, + "learning_rate": 1.579380139152435e-05, + "loss": 0.6059, + "step": 3050 + }, + { + "epoch": 5.835707502374169, + "grad_norm": 25.33370018005371, + "learning_rate": 1.5741092135779045e-05, + "loss": 0.6202, + "step": 3075 + }, + { + "epoch": 5.883190883190883, + "grad_norm": 30.26043128967285, + "learning_rate": 1.5688382880033735e-05, + "loss": 0.6552, + "step": 3100 + }, + { + "epoch": 5.930674264007598, + "grad_norm": 18.209794998168945, + "learning_rate": 1.5635673624288425e-05, + "loss": 0.6824, + "step": 3125 + }, + { + "epoch": 5.978157644824312, + "grad_norm": 38.43061828613281, + "learning_rate": 1.5582964368543115e-05, + "loss": 0.7002, + "step": 3150 + }, + { + "epoch": 6.0, + "eval_accuracy": 0.15438407746100055, + "eval_loss": 0.9505584836006165, + "eval_macro_f1": 0.5611208876450059, + "eval_micro_f1": 0.574826560951437, + "eval_runtime": 4.8418, + "eval_samples_per_second": 383.951, + "eval_steps_per_second": 24.165, + "eval_weighted_f1": 0.5844921772307327, + "step": 3162 + }, + { + "epoch": 6.0246913580246915, + "grad_norm": 18.99205780029297, + "learning_rate": 1.553025511279781e-05, + "loss": 0.5742, + "step": 3175 + }, + { + "epoch": 6.072174738841405, + "grad_norm": 9.61410903930664, + "learning_rate": 1.5477545857052502e-05, + "loss": 0.481, + "step": 3200 + }, + { + "epoch": 6.119658119658119, + "grad_norm": 16.063552856445312, + "learning_rate": 1.5424836601307192e-05, + "loss": 0.4474, + "step": 3225 + }, + { + "epoch": 6.167141500474834, + "grad_norm": 17.643842697143555, + "learning_rate": 1.5372127345561882e-05, + "loss": 0.4736, + "step": 3250 + }, + { + "epoch": 6.214624881291548, + "grad_norm": 41.78548049926758, + "learning_rate": 1.5319418089816572e-05, + "loss": 0.4752, + "step": 3275 + }, + { + "epoch": 6.262108262108262, + "grad_norm": 28.454172134399414, + "learning_rate": 1.5266708834071265e-05, + "loss": 0.4468, + "step": 3300 + }, + { + "epoch": 6.3095916429249765, + "grad_norm": 11.904889106750488, + "learning_rate": 1.5213999578325956e-05, + "loss": 0.4544, + "step": 3325 + }, + { + "epoch": 6.35707502374169, + "grad_norm": 16.409883499145508, + "learning_rate": 1.5161290322580646e-05, + "loss": 0.4544, + "step": 3350 + }, + { + "epoch": 6.404558404558404, + "grad_norm": 9.785941123962402, + "learning_rate": 1.5108581066835337e-05, + "loss": 0.4191, + "step": 3375 + }, + { + "epoch": 6.452041785375119, + "grad_norm": 32.8699836730957, + "learning_rate": 1.5055871811090027e-05, + "loss": 0.425, + "step": 3400 + }, + { + "epoch": 6.499525166191833, + "grad_norm": 32.70186233520508, + "learning_rate": 1.5003162555344719e-05, + "loss": 0.4595, + "step": 3425 + }, + { + "epoch": 6.547008547008547, + "grad_norm": 35.721683502197266, + "learning_rate": 1.4950453299599411e-05, + "loss": 0.4743, + "step": 3450 + }, + { + "epoch": 6.594491927825262, + "grad_norm": 12.760897636413574, + "learning_rate": 1.4897744043854101e-05, + "loss": 0.4509, + "step": 3475 + }, + { + "epoch": 6.6419753086419755, + "grad_norm": 10.61234188079834, + "learning_rate": 1.4845034788108794e-05, + "loss": 0.4619, + "step": 3500 + }, + { + "epoch": 6.689458689458689, + "grad_norm": 22.560327529907227, + "learning_rate": 1.4792325532363484e-05, + "loss": 0.4604, + "step": 3525 + }, + { + "epoch": 6.736942070275403, + "grad_norm": 15.491840362548828, + "learning_rate": 1.4739616276618176e-05, + "loss": 0.4701, + "step": 3550 + }, + { + "epoch": 6.784425451092118, + "grad_norm": 37.8037223815918, + "learning_rate": 1.4686907020872866e-05, + "loss": 0.4648, + "step": 3575 + }, + { + "epoch": 6.831908831908832, + "grad_norm": 16.956642150878906, + "learning_rate": 1.4634197765127558e-05, + "loss": 0.4579, + "step": 3600 + }, + { + "epoch": 6.879392212725546, + "grad_norm": 14.967536926269531, + "learning_rate": 1.458148850938225e-05, + "loss": 0.4411, + "step": 3625 + }, + { + "epoch": 6.926875593542261, + "grad_norm": 9.597579956054688, + "learning_rate": 1.452877925363694e-05, + "loss": 0.445, + "step": 3650 + }, + { + "epoch": 6.9743589743589745, + "grad_norm": 27.15288734436035, + "learning_rate": 1.4476069997891631e-05, + "loss": 0.4283, + "step": 3675 + }, + { + "epoch": 7.0, + "eval_accuracy": 0.18450779989241528, + "eval_loss": 1.1455060243606567, + "eval_macro_f1": 0.5782723196884059, + "eval_micro_f1": 0.5810215501845458, + "eval_runtime": 4.9967, + "eval_samples_per_second": 372.048, + "eval_steps_per_second": 23.416, + "eval_weighted_f1": 0.5939735287884271, + "step": 3689 + }, + { + "epoch": 7.020892687559354, + "grad_norm": 15.282934188842773, + "learning_rate": 1.4423360742146321e-05, + "loss": 0.3855, + "step": 3700 + }, + { + "epoch": 7.068376068376068, + "grad_norm": 11.879530906677246, + "learning_rate": 1.4370651486401013e-05, + "loss": 0.3347, + "step": 3725 + }, + { + "epoch": 7.115859449192783, + "grad_norm": 9.561882019042969, + "learning_rate": 1.4317942230655703e-05, + "loss": 0.3209, + "step": 3750 + }, + { + "epoch": 7.163342830009497, + "grad_norm": 13.189828872680664, + "learning_rate": 1.4265232974910395e-05, + "loss": 0.3162, + "step": 3775 + }, + { + "epoch": 7.210826210826211, + "grad_norm": 7.225050926208496, + "learning_rate": 1.4212523719165085e-05, + "loss": 0.3047, + "step": 3800 + }, + { + "epoch": 7.258309591642925, + "grad_norm": 9.511372566223145, + "learning_rate": 1.4159814463419779e-05, + "loss": 0.2889, + "step": 3825 + }, + { + "epoch": 7.305792972459639, + "grad_norm": 9.01153564453125, + "learning_rate": 1.410710520767447e-05, + "loss": 0.2788, + "step": 3850 + }, + { + "epoch": 7.353276353276353, + "grad_norm": 16.266386032104492, + "learning_rate": 1.405439595192916e-05, + "loss": 0.3126, + "step": 3875 + }, + { + "epoch": 7.400759734093067, + "grad_norm": 17.959672927856445, + "learning_rate": 1.4001686696183852e-05, + "loss": 0.2955, + "step": 3900 + }, + { + "epoch": 7.448243114909782, + "grad_norm": 18.825002670288086, + "learning_rate": 1.3948977440438542e-05, + "loss": 0.2955, + "step": 3925 + }, + { + "epoch": 7.495726495726496, + "grad_norm": 16.141721725463867, + "learning_rate": 1.3896268184693234e-05, + "loss": 0.2851, + "step": 3950 + }, + { + "epoch": 7.54320987654321, + "grad_norm": 19.40777587890625, + "learning_rate": 1.3843558928947924e-05, + "loss": 0.2869, + "step": 3975 + }, + { + "epoch": 7.590693257359924, + "grad_norm": 7.993305206298828, + "learning_rate": 1.3790849673202616e-05, + "loss": 0.3241, + "step": 4000 + }, + { + "epoch": 7.638176638176638, + "grad_norm": 24.663816452026367, + "learning_rate": 1.3738140417457306e-05, + "loss": 0.2923, + "step": 4025 + }, + { + "epoch": 7.685660018993352, + "grad_norm": 33.08135986328125, + "learning_rate": 1.3685431161711997e-05, + "loss": 0.2951, + "step": 4050 + }, + { + "epoch": 7.733143399810066, + "grad_norm": 11.239387512207031, + "learning_rate": 1.3632721905966687e-05, + "loss": 0.3102, + "step": 4075 + }, + { + "epoch": 7.780626780626781, + "grad_norm": 29.755128860473633, + "learning_rate": 1.358001265022138e-05, + "loss": 0.3141, + "step": 4100 + }, + { + "epoch": 7.828110161443495, + "grad_norm": 19.09785270690918, + "learning_rate": 1.3527303394476071e-05, + "loss": 0.3105, + "step": 4125 + }, + { + "epoch": 7.875593542260209, + "grad_norm": 18.331241607666016, + "learning_rate": 1.3474594138730763e-05, + "loss": 0.3177, + "step": 4150 + }, + { + "epoch": 7.923076923076923, + "grad_norm": 12.62179183959961, + "learning_rate": 1.3421884882985454e-05, + "loss": 0.2848, + "step": 4175 + }, + { + "epoch": 7.970560303893637, + "grad_norm": 45.56703186035156, + "learning_rate": 1.3369175627240144e-05, + "loss": 0.2963, + "step": 4200 + }, + { + "epoch": 8.0, + "eval_accuracy": 0.21732114039806347, + "eval_loss": 1.3072589635849, + "eval_macro_f1": 0.5765370791626004, + "eval_micro_f1": 0.5921717171717171, + "eval_runtime": 5.0047, + "eval_samples_per_second": 371.448, + "eval_steps_per_second": 23.378, + "eval_weighted_f1": 0.5926505086354816, + "step": 4216 + }, + { + "epoch": 8.017094017094017, + "grad_norm": 9.57081127166748, + "learning_rate": 1.3316466371494836e-05, + "loss": 0.2372, + "step": 4225 + }, + { + "epoch": 8.064577397910732, + "grad_norm": 6.422524452209473, + "learning_rate": 1.3263757115749526e-05, + "loss": 0.2301, + "step": 4250 + }, + { + "epoch": 8.112060778727445, + "grad_norm": 43.28707504272461, + "learning_rate": 1.3211047860004218e-05, + "loss": 0.1943, + "step": 4275 + }, + { + "epoch": 8.15954415954416, + "grad_norm": 13.590718269348145, + "learning_rate": 1.315833860425891e-05, + "loss": 0.2152, + "step": 4300 + }, + { + "epoch": 8.207027540360874, + "grad_norm": 15.272283554077148, + "learning_rate": 1.31056293485136e-05, + "loss": 0.2347, + "step": 4325 + }, + { + "epoch": 8.254510921177587, + "grad_norm": 7.845763206481934, + "learning_rate": 1.3052920092768292e-05, + "loss": 0.184, + "step": 4350 + }, + { + "epoch": 8.301994301994302, + "grad_norm": 6.99273681640625, + "learning_rate": 1.3000210837022982e-05, + "loss": 0.1949, + "step": 4375 + }, + { + "epoch": 8.349477682811017, + "grad_norm": 9.851614952087402, + "learning_rate": 1.2947501581277673e-05, + "loss": 0.2184, + "step": 4400 + }, + { + "epoch": 8.39696106362773, + "grad_norm": 21.555150985717773, + "learning_rate": 1.2894792325532363e-05, + "loss": 0.1959, + "step": 4425 + }, + { + "epoch": 8.444444444444445, + "grad_norm": 5.740550518035889, + "learning_rate": 1.2842083069787055e-05, + "loss": 0.1973, + "step": 4450 + }, + { + "epoch": 8.49192782526116, + "grad_norm": 51.750831604003906, + "learning_rate": 1.2789373814041749e-05, + "loss": 0.2322, + "step": 4475 + }, + { + "epoch": 8.539411206077872, + "grad_norm": 18.474130630493164, + "learning_rate": 1.2736664558296439e-05, + "loss": 0.2058, + "step": 4500 + }, + { + "epoch": 8.586894586894587, + "grad_norm": 14.638848304748535, + "learning_rate": 1.268395530255113e-05, + "loss": 0.1931, + "step": 4525 + }, + { + "epoch": 8.634377967711302, + "grad_norm": 20.490144729614258, + "learning_rate": 1.263124604680582e-05, + "loss": 0.2042, + "step": 4550 + }, + { + "epoch": 8.681861348528015, + "grad_norm": 10.135281562805176, + "learning_rate": 1.2578536791060512e-05, + "loss": 0.2113, + "step": 4575 + }, + { + "epoch": 8.72934472934473, + "grad_norm": 15.780980110168457, + "learning_rate": 1.2525827535315202e-05, + "loss": 0.2051, + "step": 4600 + }, + { + "epoch": 8.776828110161443, + "grad_norm": 36.281375885009766, + "learning_rate": 1.2473118279569894e-05, + "loss": 0.2284, + "step": 4625 + }, + { + "epoch": 8.824311490978157, + "grad_norm": 10.127740859985352, + "learning_rate": 1.2420409023824584e-05, + "loss": 0.1895, + "step": 4650 + }, + { + "epoch": 8.871794871794872, + "grad_norm": 24.48247718811035, + "learning_rate": 1.2367699768079276e-05, + "loss": 0.2153, + "step": 4675 + }, + { + "epoch": 8.919278252611585, + "grad_norm": 28.687623977661133, + "learning_rate": 1.2314990512333966e-05, + "loss": 0.2224, + "step": 4700 + }, + { + "epoch": 8.9667616334283, + "grad_norm": 23.58778190612793, + "learning_rate": 1.2262281256588657e-05, + "loss": 0.2009, + "step": 4725 + }, + { + "epoch": 9.0, + "eval_accuracy": 0.23991393222162452, + "eval_loss": 1.4571115970611572, + "eval_macro_f1": 0.5877494809891864, + "eval_micro_f1": 0.600444967936134, + "eval_runtime": 4.8234, + "eval_samples_per_second": 385.412, + "eval_steps_per_second": 24.257, + "eval_weighted_f1": 0.6004586928382261, + "step": 4743 + }, + { + "epoch": 9.01329534662868, + "grad_norm": 42.445255279541016, + "learning_rate": 1.2209572000843348e-05, + "loss": 0.1844, + "step": 4750 + }, + { + "epoch": 9.060778727445394, + "grad_norm": 10.499454498291016, + "learning_rate": 1.215686274509804e-05, + "loss": 0.1227, + "step": 4775 + }, + { + "epoch": 9.108262108262108, + "grad_norm": 74.01554107666016, + "learning_rate": 1.2104153489352733e-05, + "loss": 0.1515, + "step": 4800 + }, + { + "epoch": 9.155745489078823, + "grad_norm": 31.23480987548828, + "learning_rate": 1.2051444233607423e-05, + "loss": 0.1447, + "step": 4825 + }, + { + "epoch": 9.203228869895536, + "grad_norm": 7.121490478515625, + "learning_rate": 1.1998734977862115e-05, + "loss": 0.1555, + "step": 4850 + }, + { + "epoch": 9.250712250712251, + "grad_norm": 9.050018310546875, + "learning_rate": 1.1946025722116805e-05, + "loss": 0.1354, + "step": 4875 + }, + { + "epoch": 9.298195631528966, + "grad_norm": 7.522151947021484, + "learning_rate": 1.1893316466371496e-05, + "loss": 0.1349, + "step": 4900 + }, + { + "epoch": 9.345679012345679, + "grad_norm": 19.035409927368164, + "learning_rate": 1.1840607210626186e-05, + "loss": 0.1239, + "step": 4925 + }, + { + "epoch": 9.393162393162394, + "grad_norm": 7.271290302276611, + "learning_rate": 1.1787897954880878e-05, + "loss": 0.1152, + "step": 4950 + }, + { + "epoch": 9.440645773979107, + "grad_norm": 11.508015632629395, + "learning_rate": 1.173518869913557e-05, + "loss": 0.1185, + "step": 4975 + }, + { + "epoch": 9.488129154795821, + "grad_norm": 21.420225143432617, + "learning_rate": 1.168247944339026e-05, + "loss": 0.131, + "step": 5000 + }, + { + "epoch": 9.535612535612536, + "grad_norm": 21.15029525756836, + "learning_rate": 1.1629770187644952e-05, + "loss": 0.1384, + "step": 5025 + }, + { + "epoch": 9.583095916429249, + "grad_norm": 9.246801376342773, + "learning_rate": 1.1577060931899642e-05, + "loss": 0.1332, + "step": 5050 + }, + { + "epoch": 9.630579297245964, + "grad_norm": 5.654879093170166, + "learning_rate": 1.1524351676154333e-05, + "loss": 0.1607, + "step": 5075 + }, + { + "epoch": 9.678062678062679, + "grad_norm": 13.765711784362793, + "learning_rate": 1.1471642420409023e-05, + "loss": 0.1269, + "step": 5100 + }, + { + "epoch": 9.725546058879392, + "grad_norm": 11.771772384643555, + "learning_rate": 1.1418933164663717e-05, + "loss": 0.1225, + "step": 5125 + }, + { + "epoch": 9.773029439696106, + "grad_norm": 9.324225425720215, + "learning_rate": 1.1366223908918409e-05, + "loss": 0.1261, + "step": 5150 + }, + { + "epoch": 9.820512820512821, + "grad_norm": 12.03601360321045, + "learning_rate": 1.1313514653173099e-05, + "loss": 0.1357, + "step": 5175 + }, + { + "epoch": 9.867996201329534, + "grad_norm": 10.868319511413574, + "learning_rate": 1.126080539742779e-05, + "loss": 0.1235, + "step": 5200 + }, + { + "epoch": 9.915479582146249, + "grad_norm": 10.426804542541504, + "learning_rate": 1.120809614168248e-05, + "loss": 0.1277, + "step": 5225 + }, + { + "epoch": 9.962962962962964, + "grad_norm": 6.7754225730896, + "learning_rate": 1.1155386885937172e-05, + "loss": 0.1195, + "step": 5250 + }, + { + "epoch": 10.0, + "eval_accuracy": 0.2490586336740183, + "eval_loss": 1.572389841079712, + "eval_macro_f1": 0.5855009716659698, + "eval_micro_f1": 0.6038342941413058, + "eval_runtime": 4.9291, + "eval_samples_per_second": 377.152, + "eval_steps_per_second": 23.737, + "eval_weighted_f1": 0.6024200138375176, + "step": 5270 + }, + { + "epoch": 10.009496676163343, + "grad_norm": 4.07381010055542, + "learning_rate": 1.1102677630191862e-05, + "loss": 0.1133, + "step": 5275 + }, + { + "epoch": 10.056980056980057, + "grad_norm": 4.855506420135498, + "learning_rate": 1.1049968374446554e-05, + "loss": 0.0899, + "step": 5300 + }, + { + "epoch": 10.10446343779677, + "grad_norm": 6.390931129455566, + "learning_rate": 1.0997259118701244e-05, + "loss": 0.1037, + "step": 5325 + }, + { + "epoch": 10.151946818613485, + "grad_norm": 19.132612228393555, + "learning_rate": 1.0944549862955936e-05, + "loss": 0.0888, + "step": 5350 + }, + { + "epoch": 10.1994301994302, + "grad_norm": 7.906110763549805, + "learning_rate": 1.0891840607210626e-05, + "loss": 0.0909, + "step": 5375 + }, + { + "epoch": 10.246913580246913, + "grad_norm": 7.56940221786499, + "learning_rate": 1.0839131351465318e-05, + "loss": 0.1005, + "step": 5400 + }, + { + "epoch": 10.294396961063628, + "grad_norm": 17.0421142578125, + "learning_rate": 1.0786422095720008e-05, + "loss": 0.1084, + "step": 5425 + }, + { + "epoch": 10.341880341880342, + "grad_norm": 11.954485893249512, + "learning_rate": 1.0733712839974701e-05, + "loss": 0.0945, + "step": 5450 + }, + { + "epoch": 10.389363722697055, + "grad_norm": 5.614569187164307, + "learning_rate": 1.0681003584229393e-05, + "loss": 0.0911, + "step": 5475 + }, + { + "epoch": 10.43684710351377, + "grad_norm": 6.112778663635254, + "learning_rate": 1.0628294328484083e-05, + "loss": 0.0974, + "step": 5500 + }, + { + "epoch": 10.484330484330485, + "grad_norm": 24.94161033630371, + "learning_rate": 1.0575585072738775e-05, + "loss": 0.1222, + "step": 5525 + }, + { + "epoch": 10.531813865147198, + "grad_norm": 1.951548457145691, + "learning_rate": 1.0522875816993465e-05, + "loss": 0.1037, + "step": 5550 + }, + { + "epoch": 10.579297245963913, + "grad_norm": 4.477534294128418, + "learning_rate": 1.0470166561248156e-05, + "loss": 0.0811, + "step": 5575 + }, + { + "epoch": 10.626780626780628, + "grad_norm": 9.41921329498291, + "learning_rate": 1.0417457305502846e-05, + "loss": 0.0802, + "step": 5600 + }, + { + "epoch": 10.67426400759734, + "grad_norm": 13.954259872436523, + "learning_rate": 1.0364748049757538e-05, + "loss": 0.1071, + "step": 5625 + }, + { + "epoch": 10.721747388414055, + "grad_norm": 5.14282751083374, + "learning_rate": 1.031203879401223e-05, + "loss": 0.0902, + "step": 5650 + }, + { + "epoch": 10.76923076923077, + "grad_norm": 6.13380765914917, + "learning_rate": 1.025932953826692e-05, + "loss": 0.0817, + "step": 5675 + }, + { + "epoch": 10.816714150047483, + "grad_norm": 4.080665111541748, + "learning_rate": 1.0206620282521612e-05, + "loss": 0.0872, + "step": 5700 + }, + { + "epoch": 10.864197530864198, + "grad_norm": 12.922618865966797, + "learning_rate": 1.0153911026776302e-05, + "loss": 0.0816, + "step": 5725 + }, + { + "epoch": 10.91168091168091, + "grad_norm": 24.40860939025879, + "learning_rate": 1.0101201771030993e-05, + "loss": 0.0875, + "step": 5750 + }, + { + "epoch": 10.959164292497626, + "grad_norm": 8.776957511901855, + "learning_rate": 1.0048492515285685e-05, + "loss": 0.0789, + "step": 5775 + }, + { + "epoch": 11.0, + "eval_accuracy": 0.2528240989779451, + "eval_loss": 1.7437480688095093, + "eval_macro_f1": 0.5863809123209947, + "eval_micro_f1": 0.6008548186957121, + "eval_runtime": 4.8242, + "eval_samples_per_second": 385.349, + "eval_steps_per_second": 24.253, + "eval_weighted_f1": 0.599604351453644, + "step": 5797 + }, + { + "epoch": 11.005698005698006, + "grad_norm": 9.598529815673828, + "learning_rate": 9.995783259540375e-06, + "loss": 0.1007, + "step": 5800 + }, + { + "epoch": 11.05318138651472, + "grad_norm": 17.230466842651367, + "learning_rate": 9.943074003795067e-06, + "loss": 0.0626, + "step": 5825 + }, + { + "epoch": 11.100664767331434, + "grad_norm": 4.226415634155273, + "learning_rate": 9.890364748049759e-06, + "loss": 0.0668, + "step": 5850 + }, + { + "epoch": 11.148148148148149, + "grad_norm": 4.840086460113525, + "learning_rate": 9.83765549230445e-06, + "loss": 0.0613, + "step": 5875 + }, + { + "epoch": 11.195631528964862, + "grad_norm": 7.025901794433594, + "learning_rate": 9.78494623655914e-06, + "loss": 0.0671, + "step": 5900 + }, + { + "epoch": 11.243114909781577, + "grad_norm": 4.434010982513428, + "learning_rate": 9.732236980813832e-06, + "loss": 0.0698, + "step": 5925 + }, + { + "epoch": 11.290598290598291, + "grad_norm": 54.92803955078125, + "learning_rate": 9.679527725068522e-06, + "loss": 0.08, + "step": 5950 + }, + { + "epoch": 11.338081671415004, + "grad_norm": 6.1089630126953125, + "learning_rate": 9.626818469323214e-06, + "loss": 0.0519, + "step": 5975 + }, + { + "epoch": 11.385565052231719, + "grad_norm": 2.2659831047058105, + "learning_rate": 9.574109213577904e-06, + "loss": 0.0487, + "step": 6000 + }, + { + "epoch": 11.433048433048434, + "grad_norm": 3.403841257095337, + "learning_rate": 9.521399957832596e-06, + "loss": 0.0663, + "step": 6025 + }, + { + "epoch": 11.480531813865147, + "grad_norm": 3.4662554264068604, + "learning_rate": 9.468690702087288e-06, + "loss": 0.0663, + "step": 6050 + }, + { + "epoch": 11.528015194681862, + "grad_norm": 4.2505035400390625, + "learning_rate": 9.41598144634198e-06, + "loss": 0.0489, + "step": 6075 + }, + { + "epoch": 11.575498575498575, + "grad_norm": 11.110034942626953, + "learning_rate": 9.36327219059667e-06, + "loss": 0.0618, + "step": 6100 + }, + { + "epoch": 11.62298195631529, + "grad_norm": 4.673676490783691, + "learning_rate": 9.310562934851361e-06, + "loss": 0.0645, + "step": 6125 + }, + { + "epoch": 11.670465337132004, + "grad_norm": 4.6900553703308105, + "learning_rate": 9.257853679106051e-06, + "loss": 0.0614, + "step": 6150 + }, + { + "epoch": 11.717948717948717, + "grad_norm": 3.712474822998047, + "learning_rate": 9.205144423360743e-06, + "loss": 0.0643, + "step": 6175 + }, + { + "epoch": 11.765432098765432, + "grad_norm": 9.34663200378418, + "learning_rate": 9.152435167615435e-06, + "loss": 0.0638, + "step": 6200 + }, + { + "epoch": 11.812915479582147, + "grad_norm": 5.74763298034668, + "learning_rate": 9.099725911870125e-06, + "loss": 0.0519, + "step": 6225 + }, + { + "epoch": 11.86039886039886, + "grad_norm": 10.277482032775879, + "learning_rate": 9.047016656124816e-06, + "loss": 0.0488, + "step": 6250 + }, + { + "epoch": 11.907882241215574, + "grad_norm": 3.5318892002105713, + "learning_rate": 8.994307400379507e-06, + "loss": 0.053, + "step": 6275 + }, + { + "epoch": 11.95536562203229, + "grad_norm": 8.122844696044922, + "learning_rate": 8.941598144634198e-06, + "loss": 0.0692, + "step": 6300 + }, + { + "epoch": 12.0, + "eval_accuracy": 0.24529316837009144, + "eval_loss": 1.8085319995880127, + "eval_macro_f1": 0.5892817059465624, + "eval_micro_f1": 0.6006609749380336, + "eval_runtime": 4.9571, + "eval_samples_per_second": 375.019, + "eval_steps_per_second": 23.603, + "eval_weighted_f1": 0.6017616341149384, + "step": 6324 + }, + { + "epoch": 12.001899335232668, + "grad_norm": 1.5765128135681152, + "learning_rate": 8.888888888888888e-06, + "loss": 0.0597, + "step": 6325 + }, + { + "epoch": 12.049382716049383, + "grad_norm": 14.969993591308594, + "learning_rate": 8.836179633143582e-06, + "loss": 0.0483, + "step": 6350 + }, + { + "epoch": 12.096866096866098, + "grad_norm": 21.624101638793945, + "learning_rate": 8.783470377398272e-06, + "loss": 0.0435, + "step": 6375 + }, + { + "epoch": 12.14434947768281, + "grad_norm": 2.6236233711242676, + "learning_rate": 8.730761121652964e-06, + "loss": 0.0409, + "step": 6400 + }, + { + "epoch": 12.191832858499525, + "grad_norm": 2.5968286991119385, + "learning_rate": 8.678051865907654e-06, + "loss": 0.0423, + "step": 6425 + }, + { + "epoch": 12.239316239316238, + "grad_norm": 1.7756624221801758, + "learning_rate": 8.625342610162345e-06, + "loss": 0.0474, + "step": 6450 + }, + { + "epoch": 12.286799620132953, + "grad_norm": 3.3645403385162354, + "learning_rate": 8.572633354417035e-06, + "loss": 0.0349, + "step": 6475 + }, + { + "epoch": 12.334283000949668, + "grad_norm": 1.835996150970459, + "learning_rate": 8.519924098671729e-06, + "loss": 0.0396, + "step": 6500 + }, + { + "epoch": 12.381766381766381, + "grad_norm": 2.93648624420166, + "learning_rate": 8.467214842926419e-06, + "loss": 0.0472, + "step": 6525 + }, + { + "epoch": 12.429249762583096, + "grad_norm": 23.154329299926758, + "learning_rate": 8.41450558718111e-06, + "loss": 0.0563, + "step": 6550 + }, + { + "epoch": 12.47673314339981, + "grad_norm": 49.2928352355957, + "learning_rate": 8.3617963314358e-06, + "loss": 0.0407, + "step": 6575 + }, + { + "epoch": 12.524216524216524, + "grad_norm": 2.346856117248535, + "learning_rate": 8.309087075690492e-06, + "loss": 0.0343, + "step": 6600 + }, + { + "epoch": 12.571699905033238, + "grad_norm": 19.43263816833496, + "learning_rate": 8.256377819945182e-06, + "loss": 0.0479, + "step": 6625 + }, + { + "epoch": 12.619183285849953, + "grad_norm": 9.567797660827637, + "learning_rate": 8.203668564199874e-06, + "loss": 0.0333, + "step": 6650 + }, + { + "epoch": 12.666666666666666, + "grad_norm": 3.1313788890838623, + "learning_rate": 8.150959308454566e-06, + "loss": 0.036, + "step": 6675 + }, + { + "epoch": 12.71415004748338, + "grad_norm": 2.256866455078125, + "learning_rate": 8.098250052709256e-06, + "loss": 0.0401, + "step": 6700 + }, + { + "epoch": 12.761633428300096, + "grad_norm": 4.29447603225708, + "learning_rate": 8.045540796963948e-06, + "loss": 0.0387, + "step": 6725 + }, + { + "epoch": 12.809116809116809, + "grad_norm": 9.302247047424316, + "learning_rate": 7.992831541218638e-06, + "loss": 0.0336, + "step": 6750 + }, + { + "epoch": 12.856600189933523, + "grad_norm": 1.325979232788086, + "learning_rate": 7.94012228547333e-06, + "loss": 0.0354, + "step": 6775 + }, + { + "epoch": 12.904083570750238, + "grad_norm": 6.64365816116333, + "learning_rate": 7.887413029728021e-06, + "loss": 0.0313, + "step": 6800 + }, + { + "epoch": 12.951566951566951, + "grad_norm": 12.782024383544922, + "learning_rate": 7.834703773982713e-06, + "loss": 0.0392, + "step": 6825 + }, + { + "epoch": 12.999050332383666, + "grad_norm": 4.69736385345459, + "learning_rate": 7.781994518237403e-06, + "loss": 0.0347, + "step": 6850 + }, + { + "epoch": 13.0, + "eval_accuracy": 0.2528240989779451, + "eval_loss": 1.9986724853515625, + "eval_macro_f1": 0.5857716436175565, + "eval_micro_f1": 0.6015121814617754, + "eval_runtime": 4.8425, + "eval_samples_per_second": 383.894, + "eval_steps_per_second": 24.161, + "eval_weighted_f1": 0.6002283261093951, + "step": 6851 + }, + { + "epoch": 13.045584045584045, + "grad_norm": 0.9540898203849792, + "learning_rate": 7.729285262492095e-06, + "loss": 0.033, + "step": 6875 + }, + { + "epoch": 13.09306742640076, + "grad_norm": 2.6230499744415283, + "learning_rate": 7.676576006746785e-06, + "loss": 0.0225, + "step": 6900 + }, + { + "epoch": 13.140550807217474, + "grad_norm": 12.984963417053223, + "learning_rate": 7.6238667510014765e-06, + "loss": 0.0354, + "step": 6925 + }, + { + "epoch": 13.188034188034187, + "grad_norm": 2.323141098022461, + "learning_rate": 7.5711574952561674e-06, + "loss": 0.0253, + "step": 6950 + }, + { + "epoch": 13.235517568850902, + "grad_norm": 1.5419995784759521, + "learning_rate": 7.518448239510858e-06, + "loss": 0.0342, + "step": 6975 + }, + { + "epoch": 13.283000949667617, + "grad_norm": 1.283522605895996, + "learning_rate": 7.46573898376555e-06, + "loss": 0.0282, + "step": 7000 + }, + { + "epoch": 13.33048433048433, + "grad_norm": 2.585820198059082, + "learning_rate": 7.413029728020241e-06, + "loss": 0.0289, + "step": 7025 + }, + { + "epoch": 13.377967711301045, + "grad_norm": 1.7829973697662354, + "learning_rate": 7.360320472274932e-06, + "loss": 0.0248, + "step": 7050 + }, + { + "epoch": 13.42545109211776, + "grad_norm": 5.49683141708374, + "learning_rate": 7.307611216529623e-06, + "loss": 0.0264, + "step": 7075 + }, + { + "epoch": 13.472934472934472, + "grad_norm": 16.981597900390625, + "learning_rate": 7.2549019607843145e-06, + "loss": 0.0332, + "step": 7100 + }, + { + "epoch": 13.520417853751187, + "grad_norm": 3.749901294708252, + "learning_rate": 7.202192705039005e-06, + "loss": 0.0457, + "step": 7125 + }, + { + "epoch": 13.567901234567902, + "grad_norm": 2.465292453765869, + "learning_rate": 7.149483449293697e-06, + "loss": 0.0325, + "step": 7150 + }, + { + "epoch": 13.615384615384615, + "grad_norm": 1.3002429008483887, + "learning_rate": 7.096774193548388e-06, + "loss": 0.0242, + "step": 7175 + }, + { + "epoch": 13.66286799620133, + "grad_norm": 2.6421475410461426, + "learning_rate": 7.044064937803079e-06, + "loss": 0.0301, + "step": 7200 + }, + { + "epoch": 13.710351377018045, + "grad_norm": 2.1891891956329346, + "learning_rate": 6.99135568205777e-06, + "loss": 0.0276, + "step": 7225 + }, + { + "epoch": 13.757834757834758, + "grad_norm": 0.8588960766792297, + "learning_rate": 6.938646426312461e-06, + "loss": 0.0233, + "step": 7250 + }, + { + "epoch": 13.805318138651472, + "grad_norm": 4.041521072387695, + "learning_rate": 6.885937170567152e-06, + "loss": 0.0348, + "step": 7275 + }, + { + "epoch": 13.852801519468187, + "grad_norm": 6.659054756164551, + "learning_rate": 6.8332279148218425e-06, + "loss": 0.024, + "step": 7300 + }, + { + "epoch": 13.9002849002849, + "grad_norm": 1.0510109663009644, + "learning_rate": 6.780518659076535e-06, + "loss": 0.0256, + "step": 7325 + }, + { + "epoch": 13.947768281101615, + "grad_norm": 15.406808853149414, + "learning_rate": 6.727809403331226e-06, + "loss": 0.0242, + "step": 7350 + }, + { + "epoch": 13.995251661918328, + "grad_norm": 4.5207719802856445, + "learning_rate": 6.675100147585917e-06, + "loss": 0.0228, + "step": 7375 + }, + { + "epoch": 14.0, + "eval_accuracy": 0.2608929532006455, + "eval_loss": 2.1763834953308105, + "eval_macro_f1": 0.5897944419444678, + "eval_micro_f1": 0.6075230660042583, + "eval_runtime": 5.3544, + "eval_samples_per_second": 347.189, + "eval_steps_per_second": 21.851, + "eval_weighted_f1": 0.6043031140928729, + "step": 7378 + }, + { + "epoch": 14.041785375118709, + "grad_norm": 1.612322211265564, + "learning_rate": 6.622390891840608e-06, + "loss": 0.0173, + "step": 7400 + }, + { + "epoch": 14.089268755935423, + "grad_norm": 1.179650068283081, + "learning_rate": 6.569681636095299e-06, + "loss": 0.0227, + "step": 7425 + }, + { + "epoch": 14.136752136752136, + "grad_norm": 1.2343794107437134, + "learning_rate": 6.5169723803499896e-06, + "loss": 0.0159, + "step": 7450 + }, + { + "epoch": 14.184235517568851, + "grad_norm": 4.791398048400879, + "learning_rate": 6.464263124604681e-06, + "loss": 0.0174, + "step": 7475 + }, + { + "epoch": 14.231718898385566, + "grad_norm": 5.327462673187256, + "learning_rate": 6.411553868859372e-06, + "loss": 0.018, + "step": 7500 + }, + { + "epoch": 14.279202279202279, + "grad_norm": 2.131273031234741, + "learning_rate": 6.358844613114064e-06, + "loss": 0.0195, + "step": 7525 + }, + { + "epoch": 14.326685660018994, + "grad_norm": 2.609938144683838, + "learning_rate": 6.306135357368755e-06, + "loss": 0.0174, + "step": 7550 + }, + { + "epoch": 14.374169040835708, + "grad_norm": 1.8894281387329102, + "learning_rate": 6.253426101623446e-06, + "loss": 0.0189, + "step": 7575 + }, + { + "epoch": 14.421652421652421, + "grad_norm": 0.8751557469367981, + "learning_rate": 6.200716845878137e-06, + "loss": 0.0179, + "step": 7600 + }, + { + "epoch": 14.469135802469136, + "grad_norm": 1.151068091392517, + "learning_rate": 6.1480075901328275e-06, + "loss": 0.0149, + "step": 7625 + }, + { + "epoch": 14.51661918328585, + "grad_norm": 0.7474121451377869, + "learning_rate": 6.095298334387519e-06, + "loss": 0.0153, + "step": 7650 + }, + { + "epoch": 14.564102564102564, + "grad_norm": 0.7535553574562073, + "learning_rate": 6.04258907864221e-06, + "loss": 0.0149, + "step": 7675 + }, + { + "epoch": 14.611585944919279, + "grad_norm": 0.6468724012374878, + "learning_rate": 5.989879822896901e-06, + "loss": 0.018, + "step": 7700 + }, + { + "epoch": 14.659069325735992, + "grad_norm": 4.280311584472656, + "learning_rate": 5.937170567151592e-06, + "loss": 0.024, + "step": 7725 + }, + { + "epoch": 14.706552706552706, + "grad_norm": 1.5876237154006958, + "learning_rate": 5.884461311406283e-06, + "loss": 0.0168, + "step": 7750 + }, + { + "epoch": 14.754036087369421, + "grad_norm": 2.2202248573303223, + "learning_rate": 5.831752055660974e-06, + "loss": 0.0151, + "step": 7775 + }, + { + "epoch": 14.801519468186134, + "grad_norm": 0.5981371402740479, + "learning_rate": 5.779042799915666e-06, + "loss": 0.0233, + "step": 7800 + }, + { + "epoch": 14.849002849002849, + "grad_norm": 1.3764312267303467, + "learning_rate": 5.726333544170357e-06, + "loss": 0.0235, + "step": 7825 + }, + { + "epoch": 14.896486229819564, + "grad_norm": 1.4809439182281494, + "learning_rate": 5.673624288425048e-06, + "loss": 0.0152, + "step": 7850 + }, + { + "epoch": 14.943969610636277, + "grad_norm": 3.3015847206115723, + "learning_rate": 5.620915032679739e-06, + "loss": 0.0201, + "step": 7875 + }, + { + "epoch": 14.991452991452991, + "grad_norm": 2.4564335346221924, + "learning_rate": 5.56820577693443e-06, + "loss": 0.0139, + "step": 7900 + }, + { + "epoch": 15.0, + "eval_accuracy": 0.26358257127487894, + "eval_loss": 2.2441463470458984, + "eval_macro_f1": 0.5907126666859718, + "eval_micro_f1": 0.6064775007165377, + "eval_runtime": 5.8548, + "eval_samples_per_second": 317.519, + "eval_steps_per_second": 19.984, + "eval_weighted_f1": 0.6029898955448771, + "step": 7905 + }, + { + "epoch": 15.037986704653372, + "grad_norm": 0.6661067008972168, + "learning_rate": 5.515496521189121e-06, + "loss": 0.0106, + "step": 7925 + }, + { + "epoch": 15.085470085470085, + "grad_norm": 11.312368392944336, + "learning_rate": 5.462787265443812e-06, + "loss": 0.0174, + "step": 7950 + }, + { + "epoch": 15.1329534662868, + "grad_norm": 3.180299758911133, + "learning_rate": 5.410078009698504e-06, + "loss": 0.0137, + "step": 7975 + }, + { + "epoch": 15.180436847103513, + "grad_norm": 2.811321258544922, + "learning_rate": 5.357368753953195e-06, + "loss": 0.0149, + "step": 8000 + }, + { + "epoch": 15.227920227920228, + "grad_norm": 0.7089502215385437, + "learning_rate": 5.304659498207886e-06, + "loss": 0.01, + "step": 8025 + }, + { + "epoch": 15.275403608736942, + "grad_norm": 1.0107699632644653, + "learning_rate": 5.251950242462577e-06, + "loss": 0.0104, + "step": 8050 + }, + { + "epoch": 15.322886989553655, + "grad_norm": 0.368387907743454, + "learning_rate": 5.199240986717268e-06, + "loss": 0.0133, + "step": 8075 + }, + { + "epoch": 15.37037037037037, + "grad_norm": 2.0068130493164062, + "learning_rate": 5.146531730971959e-06, + "loss": 0.011, + "step": 8100 + }, + { + "epoch": 15.417853751187085, + "grad_norm": 0.8005654215812683, + "learning_rate": 5.0938224752266505e-06, + "loss": 0.0086, + "step": 8125 + }, + { + "epoch": 15.465337132003798, + "grad_norm": 0.40726298093795776, + "learning_rate": 5.041113219481341e-06, + "loss": 0.0108, + "step": 8150 + }, + { + "epoch": 15.512820512820513, + "grad_norm": 0.28840142488479614, + "learning_rate": 4.988403963736032e-06, + "loss": 0.0118, + "step": 8175 + }, + { + "epoch": 15.560303893637228, + "grad_norm": 0.8915543556213379, + "learning_rate": 4.935694707990723e-06, + "loss": 0.0104, + "step": 8200 + }, + { + "epoch": 15.60778727445394, + "grad_norm": 0.6993778944015503, + "learning_rate": 4.882985452245415e-06, + "loss": 0.0132, + "step": 8225 + }, + { + "epoch": 15.655270655270655, + "grad_norm": 1.1701513528823853, + "learning_rate": 4.830276196500106e-06, + "loss": 0.0096, + "step": 8250 + }, + { + "epoch": 15.70275403608737, + "grad_norm": 0.5855883955955505, + "learning_rate": 4.777566940754797e-06, + "loss": 0.0085, + "step": 8275 + }, + { + "epoch": 15.750237416904083, + "grad_norm": 0.9998419880867004, + "learning_rate": 4.724857685009488e-06, + "loss": 0.0126, + "step": 8300 + }, + { + "epoch": 15.797720797720798, + "grad_norm": 0.3291891813278198, + "learning_rate": 4.672148429264179e-06, + "loss": 0.01, + "step": 8325 + }, + { + "epoch": 15.845204178537513, + "grad_norm": 1.0975167751312256, + "learning_rate": 4.61943917351887e-06, + "loss": 0.0091, + "step": 8350 + }, + { + "epoch": 15.892687559354226, + "grad_norm": 1.5337963104248047, + "learning_rate": 4.566729917773561e-06, + "loss": 0.0116, + "step": 8375 + }, + { + "epoch": 15.94017094017094, + "grad_norm": 0.47165215015411377, + "learning_rate": 4.514020662028252e-06, + "loss": 0.0091, + "step": 8400 + }, + { + "epoch": 15.987654320987655, + "grad_norm": 3.7654190063476562, + "learning_rate": 4.461311406282944e-06, + "loss": 0.0126, + "step": 8425 + }, + { + "epoch": 16.0, + "eval_accuracy": 0.2608929532006455, + "eval_loss": 2.3537466526031494, + "eval_macro_f1": 0.5901704096408106, + "eval_micro_f1": 0.6061388410786002, + "eval_runtime": 4.8504, + "eval_samples_per_second": 383.264, + "eval_steps_per_second": 24.122, + "eval_weighted_f1": 0.6016353753076193, + "step": 8432 + }, + { + "epoch": 16.034188034188034, + "grad_norm": 36.62592697143555, + "learning_rate": 4.408602150537635e-06, + "loss": 0.0122, + "step": 8450 + }, + { + "epoch": 16.08167141500475, + "grad_norm": 0.5512651801109314, + "learning_rate": 4.3558928947923256e-06, + "loss": 0.0075, + "step": 8475 + }, + { + "epoch": 16.129154795821464, + "grad_norm": 0.8293737769126892, + "learning_rate": 4.303183639047017e-06, + "loss": 0.0078, + "step": 8500 + }, + { + "epoch": 16.17663817663818, + "grad_norm": 0.7829784154891968, + "learning_rate": 4.250474383301708e-06, + "loss": 0.0098, + "step": 8525 + }, + { + "epoch": 16.22412155745489, + "grad_norm": 0.523930549621582, + "learning_rate": 4.197765127556399e-06, + "loss": 0.0118, + "step": 8550 + }, + { + "epoch": 16.271604938271604, + "grad_norm": 0.3021624982357025, + "learning_rate": 4.145055871811091e-06, + "loss": 0.0074, + "step": 8575 + }, + { + "epoch": 16.31908831908832, + "grad_norm": 1.662912368774414, + "learning_rate": 4.092346616065782e-06, + "loss": 0.0063, + "step": 8600 + }, + { + "epoch": 16.366571699905034, + "grad_norm": 1.29483163356781, + "learning_rate": 4.039637360320473e-06, + "loss": 0.0109, + "step": 8625 + }, + { + "epoch": 16.41405508072175, + "grad_norm": 0.34829622507095337, + "learning_rate": 3.986928104575164e-06, + "loss": 0.0067, + "step": 8650 + }, + { + "epoch": 16.46153846153846, + "grad_norm": 0.32038912177085876, + "learning_rate": 3.934218848829855e-06, + "loss": 0.0071, + "step": 8675 + }, + { + "epoch": 16.509021842355175, + "grad_norm": 1.0431098937988281, + "learning_rate": 3.881509593084546e-06, + "loss": 0.0118, + "step": 8700 + }, + { + "epoch": 16.55650522317189, + "grad_norm": 0.41909337043762207, + "learning_rate": 3.828800337339237e-06, + "loss": 0.006, + "step": 8725 + }, + { + "epoch": 16.603988603988604, + "grad_norm": 0.24123843014240265, + "learning_rate": 3.7760910815939284e-06, + "loss": 0.0065, + "step": 8750 + }, + { + "epoch": 16.65147198480532, + "grad_norm": 0.23615187406539917, + "learning_rate": 3.7233818258486193e-06, + "loss": 0.0073, + "step": 8775 + }, + { + "epoch": 16.698955365622034, + "grad_norm": 0.7171230316162109, + "learning_rate": 3.67067257010331e-06, + "loss": 0.0057, + "step": 8800 + }, + { + "epoch": 16.746438746438745, + "grad_norm": 0.27123910188674927, + "learning_rate": 3.617963314358002e-06, + "loss": 0.0062, + "step": 8825 + }, + { + "epoch": 16.79392212725546, + "grad_norm": 0.3976919949054718, + "learning_rate": 3.565254058612693e-06, + "loss": 0.0063, + "step": 8850 + }, + { + "epoch": 16.841405508072175, + "grad_norm": 0.5465272665023804, + "learning_rate": 3.5125448028673837e-06, + "loss": 0.006, + "step": 8875 + }, + { + "epoch": 16.88888888888889, + "grad_norm": 0.19512887299060822, + "learning_rate": 3.459835547122075e-06, + "loss": 0.007, + "step": 8900 + }, + { + "epoch": 16.936372269705604, + "grad_norm": 0.3607795536518097, + "learning_rate": 3.407126291376766e-06, + "loss": 0.006, + "step": 8925 + }, + { + "epoch": 16.98385565052232, + "grad_norm": 1.1325352191925049, + "learning_rate": 3.3544170356314572e-06, + "loss": 0.0066, + "step": 8950 + }, + { + "epoch": 17.0, + "eval_accuracy": 0.2571274878967187, + "eval_loss": 2.4635748863220215, + "eval_macro_f1": 0.5841512510317065, + "eval_micro_f1": 0.60182900275802, + "eval_runtime": 4.978, + "eval_samples_per_second": 373.441, + "eval_steps_per_second": 23.503, + "eval_weighted_f1": 0.5982886457519079, + "step": 8959 + }, + { + "epoch": 17.030389363722698, + "grad_norm": 0.1882496327161789, + "learning_rate": 3.3017077798861485e-06, + "loss": 0.0056, + "step": 8975 + }, + { + "epoch": 17.077872744539413, + "grad_norm": 0.22095094621181488, + "learning_rate": 3.2489985241408394e-06, + "loss": 0.0051, + "step": 9000 + }, + { + "epoch": 17.125356125356124, + "grad_norm": 0.22288943827152252, + "learning_rate": 3.1962892683955303e-06, + "loss": 0.0061, + "step": 9025 + }, + { + "epoch": 17.17283950617284, + "grad_norm": 0.11666292697191238, + "learning_rate": 3.1435800126502212e-06, + "loss": 0.0042, + "step": 9050 + }, + { + "epoch": 17.220322886989553, + "grad_norm": 0.1373267024755478, + "learning_rate": 3.090870756904913e-06, + "loss": 0.0046, + "step": 9075 + }, + { + "epoch": 17.267806267806268, + "grad_norm": 0.22875097393989563, + "learning_rate": 3.038161501159604e-06, + "loss": 0.0047, + "step": 9100 + }, + { + "epoch": 17.315289648622983, + "grad_norm": 0.24153967201709747, + "learning_rate": 2.9854522454142948e-06, + "loss": 0.0045, + "step": 9125 + }, + { + "epoch": 17.362773029439698, + "grad_norm": 0.29302600026130676, + "learning_rate": 2.9327429896689865e-06, + "loss": 0.0048, + "step": 9150 + }, + { + "epoch": 17.41025641025641, + "grad_norm": 0.8335040807723999, + "learning_rate": 2.8800337339236774e-06, + "loss": 0.005, + "step": 9175 + }, + { + "epoch": 17.457739791073124, + "grad_norm": 0.2276637852191925, + "learning_rate": 2.8273244781783683e-06, + "loss": 0.0049, + "step": 9200 + }, + { + "epoch": 17.50522317188984, + "grad_norm": 0.2108197659254074, + "learning_rate": 2.7746152224330596e-06, + "loss": 0.0069, + "step": 9225 + }, + { + "epoch": 17.552706552706553, + "grad_norm": 0.27526208758354187, + "learning_rate": 2.7219059666877505e-06, + "loss": 0.0102, + "step": 9250 + }, + { + "epoch": 17.600189933523268, + "grad_norm": 0.24892586469650269, + "learning_rate": 2.6691967109424414e-06, + "loss": 0.0044, + "step": 9275 + }, + { + "epoch": 17.647673314339983, + "grad_norm": 0.24519450962543488, + "learning_rate": 2.616487455197133e-06, + "loss": 0.0048, + "step": 9300 + }, + { + "epoch": 17.695156695156694, + "grad_norm": 0.33007410168647766, + "learning_rate": 2.563778199451824e-06, + "loss": 0.0059, + "step": 9325 + }, + { + "epoch": 17.74264007597341, + "grad_norm": 0.1790347844362259, + "learning_rate": 2.511068943706515e-06, + "loss": 0.0048, + "step": 9350 + }, + { + "epoch": 17.790123456790123, + "grad_norm": 0.9636605381965637, + "learning_rate": 2.4583596879612062e-06, + "loss": 0.0045, + "step": 9375 + }, + { + "epoch": 17.837606837606838, + "grad_norm": 0.4977332353591919, + "learning_rate": 2.4056504322158976e-06, + "loss": 0.0047, + "step": 9400 + }, + { + "epoch": 17.885090218423553, + "grad_norm": 0.18379653990268707, + "learning_rate": 2.3529411764705885e-06, + "loss": 0.0059, + "step": 9425 + }, + { + "epoch": 17.932573599240264, + "grad_norm": 0.24678801000118256, + "learning_rate": 2.3002319207252798e-06, + "loss": 0.0055, + "step": 9450 + }, + { + "epoch": 17.98005698005698, + "grad_norm": 1.1211447715759277, + "learning_rate": 2.2475226649799707e-06, + "loss": 0.0046, + "step": 9475 + }, + { + "epoch": 18.0, + "eval_accuracy": 0.2625067240451856, + "eval_loss": 2.5587069988250732, + "eval_macro_f1": 0.5867033947654595, + "eval_micro_f1": 0.6049038320364117, + "eval_runtime": 5.8467, + "eval_samples_per_second": 317.958, + "eval_steps_per_second": 20.011, + "eval_weighted_f1": 0.5999985279947887, + "step": 9486 + } + ], + "logging_steps": 25, + "max_steps": 10540, + "num_input_tokens_seen": 0, + "num_train_epochs": 20, + "save_steps": 500, + "stateful_callbacks": { + "EarlyStoppingCallback": { + "args": { + "early_stopping_patience": 3, + "early_stopping_threshold": 0.0 + }, + "attributes": { + "early_stopping_patience_counter": 3 + } + }, + "TrainerControl": { + "args": { + "should_epoch_stop": false, + "should_evaluate": false, + "should_log": false, + "should_save": true, + "should_training_stop": true + }, + "attributes": {} + } + }, + "total_flos": 0.0, + "train_batch_size": 8, + "trial_name": null, + "trial_params": null +}