Add files using upload-large-folder tool
Browse filesThis view is limited to 50 files because it contains too many changes. See raw diff
- ood-uit_vsfc-bamibert-topic-domain-00-s42-20260827T154343Z-f97f2036/model.safetensors +3 -0
- ood-uit_vsfc-bamibert-topic-domain-00-s43-20260827T160912Z-85e5496f/model.safetensors +3 -0
- ood-uit_vsfc-bamibert-topic-domain-00-s44-20260827T163014Z-b6645c77/model.safetensors +3 -0
- ood-uit_vsfc-bamibert-topic-domain-01-s42-20260827T155348Z-f97f2036/model.safetensors +3 -0
- ood-uit_vsfc-bamibert-topic-domain-01-s43-20260827T161312Z-85e5496f/model.safetensors +3 -0
- ood-uit_vsfc-bamibert-topic-domain-02-s42-20260827T155652Z-f97f2036/model.safetensors +3 -0
- ood-uit_vsfc-bamibert-topic-domain-02-s43-20260827T161535Z-85e5496f/model.safetensors +3 -0
- ood-uit_vsfc-bamibert-topic-domain-03-s42-20260827T160302Z-f97f2036/model.safetensors +3 -0
- ood-uit_vsfc-bamibert-topic-domain-03-s43-20260827T162407Z-85e5496f/model.safetensors +3 -0
- ood-vifactcheck-bamibert-topic-domain-02-s43-20260827T193604Z-3858bb98/trainer/checkpoint-2366/trainer_state.json +844 -0
- ood-vifactcheck-bamibert-topic-domain-03-s43-20260827T194124Z-3858bb98/trainer/checkpoint-1365/special_tokens_map.json +51 -0
- ood-vifactcheck-bamibert-topic-domain-03-s43-20260827T194124Z-3858bb98/trainer/checkpoint-1365/tokenizer.json +0 -0
- ood-vifactcheck-bamibert-topic-domain-03-s43-20260827T194124Z-3858bb98/trainer/checkpoint-1365/tokenizer_config.json +55 -0
- ood-vifactcheck-bamibert-topic-domain-03-s43-20260827T194124Z-3858bb98/trainer/checkpoint-1365/trainer_state.json +498 -0
- ood-vifactcheck-bamibert-topic-domain-03-s43-20260827T194124Z-3858bb98/trainer/checkpoint-1950/special_tokens_map.json +51 -0
- ood-vifactcheck-bamibert-topic-domain-03-s43-20260827T194124Z-3858bb98/trainer/checkpoint-1950/tokenizer.json +0 -0
- ood-vifactcheck-bamibert-topic-domain-03-s43-20260827T194124Z-3858bb98/trainer/checkpoint-1950/tokenizer_config.json +55 -0
- ood-vifactcheck-bamibert-topic-domain-03-s43-20260827T194124Z-3858bb98/trainer/checkpoint-1950/trainer_state.json +699 -0
- ood-vifactcheck-bamibert-topic-domain-04-s43-20260827T194549Z-3858bb98/trainer/checkpoint-1692/special_tokens_map.json +51 -0
- ood-vifactcheck-bamibert-topic-domain-04-s43-20260827T194549Z-3858bb98/trainer/checkpoint-1692/tokenizer.json +0 -0
- ood-vifactcheck-bamibert-topic-domain-04-s43-20260827T194549Z-3858bb98/trainer/checkpoint-1692/tokenizer_config.json +55 -0
- ood-vifactcheck-bamibert-topic-domain-04-s43-20260827T194549Z-3858bb98/trainer/checkpoint-1692/trainer_state.json +611 -0
- ood-vifactcheck-bamibert-topic-domain-04-s43-20260827T194549Z-3858bb98/trainer/checkpoint-2256/special_tokens_map.json +51 -0
- ood-vifactcheck-bamibert-topic-domain-04-s43-20260827T194549Z-3858bb98/trainer/checkpoint-2256/tokenizer.json +0 -0
- ood-vifactcheck-bamibert-topic-domain-04-s43-20260827T194549Z-3858bb98/trainer/checkpoint-2256/tokenizer_config.json +55 -0
- ood-vifactcheck-bamibert-topic-domain-04-s43-20260827T194549Z-3858bb98/trainer/checkpoint-2256/trainer_state.json +805 -0
- ood-vifactcheck-bamibert-topic-domain-05-s43-20260827T195058Z-3858bb98/trainer/checkpoint-1575/special_tokens_map.json +51 -0
- ood-vifactcheck-bamibert-topic-domain-05-s43-20260827T195058Z-3858bb98/trainer/checkpoint-1575/tokenizer.json +0 -0
- ood-vifactcheck-bamibert-topic-domain-05-s43-20260827T195058Z-3858bb98/trainer/checkpoint-1575/tokenizer_config.json +55 -0
- ood-vifactcheck-bamibert-topic-domain-05-s43-20260827T195058Z-3858bb98/trainer/checkpoint-1575/trainer_state.json +583 -0
- ood-vifactcheck-bamibert-topic-domain-05-s43-20260827T195058Z-3858bb98/trainer/checkpoint-2100/special_tokens_map.json +51 -0
- ood-vifactcheck-bamibert-topic-domain-05-s43-20260827T195058Z-3858bb98/trainer/checkpoint-2100/tokenizer.json +0 -0
- ood-vifactcheck-bamibert-topic-domain-05-s43-20260827T195058Z-3858bb98/trainer/checkpoint-2100/tokenizer_config.json +55 -0
- ood-vifactcheck-bamibert-topic-domain-05-s43-20260827T195058Z-3858bb98/trainer/checkpoint-2100/trainer_state.json +763 -0
- ood-vifactcheck-bamibert-topic-domain-06-s43-20260827T195550Z-3858bb98/trainer/checkpoint-1104/special_tokens_map.json +51 -0
- ood-vifactcheck-bamibert-topic-domain-06-s43-20260827T195550Z-3858bb98/trainer/checkpoint-1104/tokenizer.json +0 -0
- ood-vifactcheck-bamibert-topic-domain-06-s43-20260827T195550Z-3858bb98/trainer/checkpoint-1104/tokenizer_config.json +55 -0
- ood-vifactcheck-bamibert-topic-domain-06-s43-20260827T195550Z-3858bb98/trainer/checkpoint-1104/trainer_state.json +417 -0
- ood-vifactcheck-bamibert-topic-domain-06-s43-20260827T195550Z-3858bb98/trainer/checkpoint-552/special_tokens_map.json +51 -0
- ood-vifactcheck-bamibert-topic-domain-06-s43-20260827T195550Z-3858bb98/trainer/checkpoint-552/tokenizer.json +0 -0
- ood-vifactcheck-bamibert-topic-domain-06-s43-20260827T195550Z-3858bb98/trainer/checkpoint-552/tokenizer_config.json +55 -0
- ood-vifactcheck-bamibert-topic-domain-06-s43-20260827T195550Z-3858bb98/trainer/checkpoint-552/trainer_state.json +230 -0
- ood-vifactcheck-bamibert-topic-domain-07-s43-20260827T195828Z-3858bb98/trainer/checkpoint-2167/special_tokens_map.json +51 -0
- ood-vifactcheck-bamibert-topic-domain-07-s43-20260827T195828Z-3858bb98/trainer/checkpoint-2167/tokenizer.json +0 -0
- ood-vifactcheck-bamibert-topic-domain-07-s43-20260827T195828Z-3858bb98/trainer/checkpoint-2167/tokenizer_config.json +55 -0
- ood-vifactcheck-bamibert-topic-domain-07-s43-20260827T195828Z-3858bb98/trainer/checkpoint-2167/trainer_state.json +766 -0
- ood-vifactcheck-bamibert-topic-domain-07-s43-20260827T195828Z-3858bb98/trainer/checkpoint-2758/special_tokens_map.json +51 -0
- ood-vifactcheck-bamibert-topic-domain-07-s43-20260827T195828Z-3858bb98/trainer/checkpoint-2758/tokenizer.json +0 -0
- ood-vifactcheck-bamibert-topic-domain-07-s43-20260827T195828Z-3858bb98/trainer/checkpoint-2758/tokenizer_config.json +55 -0
- ood-vifactcheck-bamibert-topic-domain-07-s43-20260827T195828Z-3858bb98/trainer/checkpoint-2758/trainer_state.json +967 -0
ood-uit_vsfc-bamibert-topic-domain-00-s42-20260827T154343Z-f97f2036/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:d18b68f47f37fff72b4810de2595386570349355627f51af00e797494aa14cb6
|
| 3 |
+
size 411837948
|
ood-uit_vsfc-bamibert-topic-domain-00-s43-20260827T160912Z-85e5496f/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:4c0c200350e5a130fbc9dfee3eff5e9c3b7006f83de7cc0983add56750f3b320
|
| 3 |
+
size 411837948
|
ood-uit_vsfc-bamibert-topic-domain-00-s44-20260827T163014Z-b6645c77/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:459895c6384769ae92d4129645e28694e7f65b7520d51b13e8d66924c5377b9a
|
| 3 |
+
size 411837948
|
ood-uit_vsfc-bamibert-topic-domain-01-s42-20260827T155348Z-f97f2036/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:7366d756026897fe05a7e28ec422e8ce740a146bc80994e2c01ea771dd34a849
|
| 3 |
+
size 411837948
|
ood-uit_vsfc-bamibert-topic-domain-01-s43-20260827T161312Z-85e5496f/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:1c0535e10dd3766ffab5433ad4af3f0e0d19aa7f32426161b8cc29bc75a158d2
|
| 3 |
+
size 411837948
|
ood-uit_vsfc-bamibert-topic-domain-02-s42-20260827T155652Z-f97f2036/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:1c6a83a4c7472f8fdbce737a9ebd31fb63f313f03c88f3935e390688931c3c81
|
| 3 |
+
size 411837948
|
ood-uit_vsfc-bamibert-topic-domain-02-s43-20260827T161535Z-85e5496f/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:6faa227f5764915542c9d9de06237deefd79709b3d03100af3b1bd76bae14af1
|
| 3 |
+
size 411837948
|
ood-uit_vsfc-bamibert-topic-domain-03-s42-20260827T160302Z-f97f2036/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:6ebbad9aa991f3096b62206c9764224f9b5fbc1825a1aeab7c38b0466fbc8f61
|
| 3 |
+
size 411837948
|
ood-uit_vsfc-bamibert-topic-domain-03-s43-20260827T162407Z-85e5496f/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:92800725fa1d649e373cb3909b9db56c939720f423ed70c10d863449bdf6ca2d
|
| 3 |
+
size 411837948
|
ood-vifactcheck-bamibert-topic-domain-02-s43-20260827T193604Z-3858bb98/trainer/checkpoint-2366/trainer_state.json
ADDED
|
@@ -0,0 +1,844 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 1820,
|
| 3 |
+
"best_metric": 0.869309032967074,
|
| 4 |
+
"best_model_checkpoint": "/root/VNEncoder/results/checkpoints/ood-vifactcheck-bamibert-topic-domain-02-s43-20260827T193604Z-3858bb98/trainer/checkpoint-1820",
|
| 5 |
+
"epoch": 13.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 2366,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.13793103448275862,
|
| 14 |
+
"grad_norm": 11.58485221862793,
|
| 15 |
+
"learning_rate": 1.3186813186813187e-06,
|
| 16 |
+
"loss": 4.4314,
|
| 17 |
+
"step": 25
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.27586206896551724,
|
| 21 |
+
"grad_norm": 9.425135612487793,
|
| 22 |
+
"learning_rate": 2.6923076923076923e-06,
|
| 23 |
+
"loss": 4.3978,
|
| 24 |
+
"step": 50
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.41379310344827586,
|
| 28 |
+
"grad_norm": 8.210010528564453,
|
| 29 |
+
"learning_rate": 4.065934065934066e-06,
|
| 30 |
+
"loss": 4.3829,
|
| 31 |
+
"step": 75
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.5517241379310345,
|
| 35 |
+
"grad_norm": 18.993253707885742,
|
| 36 |
+
"learning_rate": 5.43956043956044e-06,
|
| 37 |
+
"loss": 4.3861,
|
| 38 |
+
"step": 100
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.6896551724137931,
|
| 42 |
+
"grad_norm": 19.984159469604492,
|
| 43 |
+
"learning_rate": 6.813186813186814e-06,
|
| 44 |
+
"loss": 4.3753,
|
| 45 |
+
"step": 125
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.8275862068965517,
|
| 49 |
+
"grad_norm": 18.76668357849121,
|
| 50 |
+
"learning_rate": 8.186813186813188e-06,
|
| 51 |
+
"loss": 4.32,
|
| 52 |
+
"step": 150
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"epoch": 0.9655172413793104,
|
| 56 |
+
"grad_norm": 19.58595085144043,
|
| 57 |
+
"learning_rate": 9.560439560439562e-06,
|
| 58 |
+
"loss": 4.3283,
|
| 59 |
+
"step": 175
|
| 60 |
+
},
|
| 61 |
+
{
|
| 62 |
+
"epoch": 1.0,
|
| 63 |
+
"eval_accuracy": 0.47912524850894633,
|
| 64 |
+
"eval_loss": 1.0147746801376343,
|
| 65 |
+
"eval_macro_f1": 0.4170847419784449,
|
| 66 |
+
"eval_runtime": 0.5825,
|
| 67 |
+
"eval_samples_per_second": 863.45,
|
| 68 |
+
"eval_steps_per_second": 54.931,
|
| 69 |
+
"eval_weighted_f1": 0.4192667321010538,
|
| 70 |
+
"step": 182
|
| 71 |
+
},
|
| 72 |
+
{
|
| 73 |
+
"epoch": 1.0993103448275863,
|
| 74 |
+
"grad_norm": 58.16252517700195,
|
| 75 |
+
"learning_rate": 1.0934065934065933e-05,
|
| 76 |
+
"loss": 3.9772,
|
| 77 |
+
"step": 200
|
| 78 |
+
},
|
| 79 |
+
{
|
| 80 |
+
"epoch": 1.237241379310345,
|
| 81 |
+
"grad_norm": 49.73049545288086,
|
| 82 |
+
"learning_rate": 1.230769230769231e-05,
|
| 83 |
+
"loss": 3.599,
|
| 84 |
+
"step": 225
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 1.3751724137931034,
|
| 88 |
+
"grad_norm": 58.101070404052734,
|
| 89 |
+
"learning_rate": 1.3681318681318683e-05,
|
| 90 |
+
"loss": 3.1096,
|
| 91 |
+
"step": 250
|
| 92 |
+
},
|
| 93 |
+
{
|
| 94 |
+
"epoch": 1.513103448275862,
|
| 95 |
+
"grad_norm": 45.06196212768555,
|
| 96 |
+
"learning_rate": 1.5054945054945056e-05,
|
| 97 |
+
"loss": 3.0102,
|
| 98 |
+
"step": 275
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 1.6510344827586207,
|
| 102 |
+
"grad_norm": 41.06315231323242,
|
| 103 |
+
"learning_rate": 1.642857142857143e-05,
|
| 104 |
+
"loss": 2.8282,
|
| 105 |
+
"step": 300
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 1.7889655172413792,
|
| 109 |
+
"grad_norm": 39.79212951660156,
|
| 110 |
+
"learning_rate": 1.78021978021978e-05,
|
| 111 |
+
"loss": 2.5567,
|
| 112 |
+
"step": 325
|
| 113 |
+
},
|
| 114 |
+
{
|
| 115 |
+
"epoch": 1.926896551724138,
|
| 116 |
+
"grad_norm": 39.06377029418945,
|
| 117 |
+
"learning_rate": 1.9175824175824178e-05,
|
| 118 |
+
"loss": 2.5691,
|
| 119 |
+
"step": 350
|
| 120 |
+
},
|
| 121 |
+
{
|
| 122 |
+
"epoch": 2.0,
|
| 123 |
+
"eval_accuracy": 0.7872763419483101,
|
| 124 |
+
"eval_loss": 0.5018846988677979,
|
| 125 |
+
"eval_macro_f1": 0.7884482306920942,
|
| 126 |
+
"eval_runtime": 0.5852,
|
| 127 |
+
"eval_samples_per_second": 859.475,
|
| 128 |
+
"eval_steps_per_second": 54.678,
|
| 129 |
+
"eval_weighted_f1": 0.788913120430031,
|
| 130 |
+
"step": 364
|
| 131 |
+
},
|
| 132 |
+
{
|
| 133 |
+
"epoch": 2.0606896551724136,
|
| 134 |
+
"grad_norm": 42.192054748535156,
|
| 135 |
+
"learning_rate": 1.993894993894994e-05,
|
| 136 |
+
"loss": 2.1688,
|
| 137 |
+
"step": 375
|
| 138 |
+
},
|
| 139 |
+
{
|
| 140 |
+
"epoch": 2.1986206896551725,
|
| 141 |
+
"grad_norm": 43.01150894165039,
|
| 142 |
+
"learning_rate": 1.9786324786324788e-05,
|
| 143 |
+
"loss": 1.7811,
|
| 144 |
+
"step": 400
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 2.336551724137931,
|
| 148 |
+
"grad_norm": 27.328853607177734,
|
| 149 |
+
"learning_rate": 1.9633699633699634e-05,
|
| 150 |
+
"loss": 1.8751,
|
| 151 |
+
"step": 425
|
| 152 |
+
},
|
| 153 |
+
{
|
| 154 |
+
"epoch": 2.47448275862069,
|
| 155 |
+
"grad_norm": 53.49696350097656,
|
| 156 |
+
"learning_rate": 1.948107448107448e-05,
|
| 157 |
+
"loss": 1.936,
|
| 158 |
+
"step": 450
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 2.6124137931034483,
|
| 162 |
+
"grad_norm": 43.56639099121094,
|
| 163 |
+
"learning_rate": 1.932844932844933e-05,
|
| 164 |
+
"loss": 1.839,
|
| 165 |
+
"step": 475
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 2.750344827586207,
|
| 169 |
+
"grad_norm": 59.44014358520508,
|
| 170 |
+
"learning_rate": 1.9175824175824178e-05,
|
| 171 |
+
"loss": 1.9447,
|
| 172 |
+
"step": 500
|
| 173 |
+
},
|
| 174 |
+
{
|
| 175 |
+
"epoch": 2.8882758620689657,
|
| 176 |
+
"grad_norm": 32.07001876831055,
|
| 177 |
+
"learning_rate": 1.9023199023199025e-05,
|
| 178 |
+
"loss": 1.7301,
|
| 179 |
+
"step": 525
|
| 180 |
+
},
|
| 181 |
+
{
|
| 182 |
+
"epoch": 3.0,
|
| 183 |
+
"eval_accuracy": 0.8330019880715706,
|
| 184 |
+
"eval_loss": 0.4471936821937561,
|
| 185 |
+
"eval_macro_f1": 0.831735853217329,
|
| 186 |
+
"eval_runtime": 0.5916,
|
| 187 |
+
"eval_samples_per_second": 850.291,
|
| 188 |
+
"eval_steps_per_second": 54.094,
|
| 189 |
+
"eval_weighted_f1": 0.8319282926037921,
|
| 190 |
+
"step": 546
|
| 191 |
+
},
|
| 192 |
+
{
|
| 193 |
+
"epoch": 3.0220689655172412,
|
| 194 |
+
"grad_norm": 46.59756088256836,
|
| 195 |
+
"learning_rate": 1.887057387057387e-05,
|
| 196 |
+
"loss": 1.6669,
|
| 197 |
+
"step": 550
|
| 198 |
+
},
|
| 199 |
+
{
|
| 200 |
+
"epoch": 3.16,
|
| 201 |
+
"grad_norm": 66.64631652832031,
|
| 202 |
+
"learning_rate": 1.8717948717948718e-05,
|
| 203 |
+
"loss": 1.2406,
|
| 204 |
+
"step": 575
|
| 205 |
+
},
|
| 206 |
+
{
|
| 207 |
+
"epoch": 3.2979310344827586,
|
| 208 |
+
"grad_norm": 43.43894958496094,
|
| 209 |
+
"learning_rate": 1.8565323565323568e-05,
|
| 210 |
+
"loss": 1.1403,
|
| 211 |
+
"step": 600
|
| 212 |
+
},
|
| 213 |
+
{
|
| 214 |
+
"epoch": 3.435862068965517,
|
| 215 |
+
"grad_norm": 42.17364501953125,
|
| 216 |
+
"learning_rate": 1.8412698412698415e-05,
|
| 217 |
+
"loss": 1.3168,
|
| 218 |
+
"step": 625
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 3.573793103448276,
|
| 222 |
+
"grad_norm": 50.19221878051758,
|
| 223 |
+
"learning_rate": 1.826007326007326e-05,
|
| 224 |
+
"loss": 1.2885,
|
| 225 |
+
"step": 650
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"epoch": 3.7117241379310344,
|
| 229 |
+
"grad_norm": 53.39130783081055,
|
| 230 |
+
"learning_rate": 1.8107448107448108e-05,
|
| 231 |
+
"loss": 1.2697,
|
| 232 |
+
"step": 675
|
| 233 |
+
},
|
| 234 |
+
{
|
| 235 |
+
"epoch": 3.849655172413793,
|
| 236 |
+
"grad_norm": 57.69573211669922,
|
| 237 |
+
"learning_rate": 1.7954822954822958e-05,
|
| 238 |
+
"loss": 1.1977,
|
| 239 |
+
"step": 700
|
| 240 |
+
},
|
| 241 |
+
{
|
| 242 |
+
"epoch": 3.9875862068965517,
|
| 243 |
+
"grad_norm": 62.88957214355469,
|
| 244 |
+
"learning_rate": 1.78021978021978e-05,
|
| 245 |
+
"loss": 1.2081,
|
| 246 |
+
"step": 725
|
| 247 |
+
},
|
| 248 |
+
{
|
| 249 |
+
"epoch": 4.0,
|
| 250 |
+
"eval_accuracy": 0.8469184890656064,
|
| 251 |
+
"eval_loss": 0.4087779223918915,
|
| 252 |
+
"eval_macro_f1": 0.8447465902307477,
|
| 253 |
+
"eval_runtime": 1.5185,
|
| 254 |
+
"eval_samples_per_second": 331.244,
|
| 255 |
+
"eval_steps_per_second": 21.073,
|
| 256 |
+
"eval_weighted_f1": 0.8451006182184996,
|
| 257 |
+
"step": 728
|
| 258 |
+
},
|
| 259 |
+
{
|
| 260 |
+
"epoch": 4.121379310344827,
|
| 261 |
+
"grad_norm": 51.63662338256836,
|
| 262 |
+
"learning_rate": 1.764957264957265e-05,
|
| 263 |
+
"loss": 0.8759,
|
| 264 |
+
"step": 750
|
| 265 |
+
},
|
| 266 |
+
{
|
| 267 |
+
"epoch": 4.259310344827586,
|
| 268 |
+
"grad_norm": 35.93242645263672,
|
| 269 |
+
"learning_rate": 1.74969474969475e-05,
|
| 270 |
+
"loss": 0.8427,
|
| 271 |
+
"step": 775
|
| 272 |
+
},
|
| 273 |
+
{
|
| 274 |
+
"epoch": 4.397241379310345,
|
| 275 |
+
"grad_norm": 54.42876434326172,
|
| 276 |
+
"learning_rate": 1.7344322344322345e-05,
|
| 277 |
+
"loss": 0.7767,
|
| 278 |
+
"step": 800
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 4.535172413793103,
|
| 282 |
+
"grad_norm": 44.173789978027344,
|
| 283 |
+
"learning_rate": 1.7191697191697195e-05,
|
| 284 |
+
"loss": 0.9649,
|
| 285 |
+
"step": 825
|
| 286 |
+
},
|
| 287 |
+
{
|
| 288 |
+
"epoch": 4.673103448275862,
|
| 289 |
+
"grad_norm": 29.70044708251953,
|
| 290 |
+
"learning_rate": 1.703907203907204e-05,
|
| 291 |
+
"loss": 0.6833,
|
| 292 |
+
"step": 850
|
| 293 |
+
},
|
| 294 |
+
{
|
| 295 |
+
"epoch": 4.811034482758621,
|
| 296 |
+
"grad_norm": 82.69610595703125,
|
| 297 |
+
"learning_rate": 1.688644688644689e-05,
|
| 298 |
+
"loss": 1.0218,
|
| 299 |
+
"step": 875
|
| 300 |
+
},
|
| 301 |
+
{
|
| 302 |
+
"epoch": 4.94896551724138,
|
| 303 |
+
"grad_norm": 120.68234252929688,
|
| 304 |
+
"learning_rate": 1.6733821733821735e-05,
|
| 305 |
+
"loss": 0.7451,
|
| 306 |
+
"step": 900
|
| 307 |
+
},
|
| 308 |
+
{
|
| 309 |
+
"epoch": 5.0,
|
| 310 |
+
"eval_accuracy": 0.8250497017892644,
|
| 311 |
+
"eval_loss": 0.4958866238594055,
|
| 312 |
+
"eval_macro_f1": 0.8235745820260516,
|
| 313 |
+
"eval_runtime": 0.5765,
|
| 314 |
+
"eval_samples_per_second": 872.527,
|
| 315 |
+
"eval_steps_per_second": 55.509,
|
| 316 |
+
"eval_weighted_f1": 0.8239097745772974,
|
| 317 |
+
"step": 910
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 5.082758620689655,
|
| 321 |
+
"grad_norm": 36.33766174316406,
|
| 322 |
+
"learning_rate": 1.6581196581196585e-05,
|
| 323 |
+
"loss": 0.605,
|
| 324 |
+
"step": 925
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 5.220689655172414,
|
| 328 |
+
"grad_norm": 141.90615844726562,
|
| 329 |
+
"learning_rate": 1.642857142857143e-05,
|
| 330 |
+
"loss": 0.5546,
|
| 331 |
+
"step": 950
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 5.358620689655172,
|
| 335 |
+
"grad_norm": 32.26487350463867,
|
| 336 |
+
"learning_rate": 1.627594627594628e-05,
|
| 337 |
+
"loss": 0.5659,
|
| 338 |
+
"step": 975
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 5.496551724137931,
|
| 342 |
+
"grad_norm": 42.44941711425781,
|
| 343 |
+
"learning_rate": 1.6123321123321125e-05,
|
| 344 |
+
"loss": 0.5046,
|
| 345 |
+
"step": 1000
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 5.63448275862069,
|
| 349 |
+
"grad_norm": 57.650203704833984,
|
| 350 |
+
"learning_rate": 1.5970695970695972e-05,
|
| 351 |
+
"loss": 0.5452,
|
| 352 |
+
"step": 1025
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 5.772413793103448,
|
| 356 |
+
"grad_norm": 15.052305221557617,
|
| 357 |
+
"learning_rate": 1.581807081807082e-05,
|
| 358 |
+
"loss": 0.5173,
|
| 359 |
+
"step": 1050
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 5.910344827586207,
|
| 363 |
+
"grad_norm": 46.666446685791016,
|
| 364 |
+
"learning_rate": 1.5665445665445665e-05,
|
| 365 |
+
"loss": 0.6587,
|
| 366 |
+
"step": 1075
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 6.0,
|
| 370 |
+
"eval_accuracy": 0.8508946322067594,
|
| 371 |
+
"eval_loss": 0.5746742486953735,
|
| 372 |
+
"eval_macro_f1": 0.8519946935182702,
|
| 373 |
+
"eval_runtime": 0.5866,
|
| 374 |
+
"eval_samples_per_second": 857.472,
|
| 375 |
+
"eval_steps_per_second": 54.551,
|
| 376 |
+
"eval_weighted_f1": 0.8521578610698961,
|
| 377 |
+
"step": 1092
|
| 378 |
+
},
|
| 379 |
+
{
|
| 380 |
+
"epoch": 6.0441379310344825,
|
| 381 |
+
"grad_norm": 81.79331970214844,
|
| 382 |
+
"learning_rate": 1.5512820512820516e-05,
|
| 383 |
+
"loss": 0.5335,
|
| 384 |
+
"step": 1100
|
| 385 |
+
},
|
| 386 |
+
{
|
| 387 |
+
"epoch": 6.182068965517241,
|
| 388 |
+
"grad_norm": 114.01929473876953,
|
| 389 |
+
"learning_rate": 1.5360195360195362e-05,
|
| 390 |
+
"loss": 0.4932,
|
| 391 |
+
"step": 1125
|
| 392 |
+
},
|
| 393 |
+
{
|
| 394 |
+
"epoch": 6.32,
|
| 395 |
+
"grad_norm": 67.68749237060547,
|
| 396 |
+
"learning_rate": 1.5207570207570207e-05,
|
| 397 |
+
"loss": 0.3986,
|
| 398 |
+
"step": 1150
|
| 399 |
+
},
|
| 400 |
+
{
|
| 401 |
+
"epoch": 6.457931034482758,
|
| 402 |
+
"grad_norm": 44.281253814697266,
|
| 403 |
+
"learning_rate": 1.5054945054945056e-05,
|
| 404 |
+
"loss": 0.3298,
|
| 405 |
+
"step": 1175
|
| 406 |
+
},
|
| 407 |
+
{
|
| 408 |
+
"epoch": 6.595862068965517,
|
| 409 |
+
"grad_norm": 270.51141357421875,
|
| 410 |
+
"learning_rate": 1.4902319902319904e-05,
|
| 411 |
+
"loss": 0.4579,
|
| 412 |
+
"step": 1200
|
| 413 |
+
},
|
| 414 |
+
{
|
| 415 |
+
"epoch": 6.733793103448276,
|
| 416 |
+
"grad_norm": 40.47040557861328,
|
| 417 |
+
"learning_rate": 1.4749694749694752e-05,
|
| 418 |
+
"loss": 0.4534,
|
| 419 |
+
"step": 1225
|
| 420 |
+
},
|
| 421 |
+
{
|
| 422 |
+
"epoch": 6.871724137931034,
|
| 423 |
+
"grad_norm": 25.67839813232422,
|
| 424 |
+
"learning_rate": 1.4597069597069597e-05,
|
| 425 |
+
"loss": 0.4817,
|
| 426 |
+
"step": 1250
|
| 427 |
+
},
|
| 428 |
+
{
|
| 429 |
+
"epoch": 7.0,
|
| 430 |
+
"eval_accuracy": 0.8429423459244533,
|
| 431 |
+
"eval_loss": 0.679874837398529,
|
| 432 |
+
"eval_macro_f1": 0.8419964374428378,
|
| 433 |
+
"eval_runtime": 0.577,
|
| 434 |
+
"eval_samples_per_second": 871.8,
|
| 435 |
+
"eval_steps_per_second": 55.462,
|
| 436 |
+
"eval_weighted_f1": 0.8422470469909669,
|
| 437 |
+
"step": 1274
|
| 438 |
+
},
|
| 439 |
+
{
|
| 440 |
+
"epoch": 7.0055172413793105,
|
| 441 |
+
"grad_norm": 0.925383448600769,
|
| 442 |
+
"learning_rate": 1.4444444444444446e-05,
|
| 443 |
+
"loss": 0.3014,
|
| 444 |
+
"step": 1275
|
| 445 |
+
},
|
| 446 |
+
{
|
| 447 |
+
"epoch": 7.1434482758620685,
|
| 448 |
+
"grad_norm": 93.09014892578125,
|
| 449 |
+
"learning_rate": 1.4291819291819293e-05,
|
| 450 |
+
"loss": 0.3997,
|
| 451 |
+
"step": 1300
|
| 452 |
+
},
|
| 453 |
+
{
|
| 454 |
+
"epoch": 7.281379310344827,
|
| 455 |
+
"grad_norm": 0.23526118695735931,
|
| 456 |
+
"learning_rate": 1.4139194139194141e-05,
|
| 457 |
+
"loss": 0.37,
|
| 458 |
+
"step": 1325
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 7.419310344827586,
|
| 462 |
+
"grad_norm": 129.131591796875,
|
| 463 |
+
"learning_rate": 1.3986568986568988e-05,
|
| 464 |
+
"loss": 0.3375,
|
| 465 |
+
"step": 1350
|
| 466 |
+
},
|
| 467 |
+
{
|
| 468 |
+
"epoch": 7.557241379310344,
|
| 469 |
+
"grad_norm": 1.749083161354065,
|
| 470 |
+
"learning_rate": 1.3833943833943834e-05,
|
| 471 |
+
"loss": 0.3286,
|
| 472 |
+
"step": 1375
|
| 473 |
+
},
|
| 474 |
+
{
|
| 475 |
+
"epoch": 7.695172413793103,
|
| 476 |
+
"grad_norm": 94.52139282226562,
|
| 477 |
+
"learning_rate": 1.3681318681318683e-05,
|
| 478 |
+
"loss": 0.3205,
|
| 479 |
+
"step": 1400
|
| 480 |
+
},
|
| 481 |
+
{
|
| 482 |
+
"epoch": 7.833103448275862,
|
| 483 |
+
"grad_norm": 178.98956298828125,
|
| 484 |
+
"learning_rate": 1.3528693528693531e-05,
|
| 485 |
+
"loss": 0.3422,
|
| 486 |
+
"step": 1425
|
| 487 |
+
},
|
| 488 |
+
{
|
| 489 |
+
"epoch": 7.971034482758621,
|
| 490 |
+
"grad_norm": 1.7286487817764282,
|
| 491 |
+
"learning_rate": 1.3376068376068376e-05,
|
| 492 |
+
"loss": 0.6482,
|
| 493 |
+
"step": 1450
|
| 494 |
+
},
|
| 495 |
+
{
|
| 496 |
+
"epoch": 8.0,
|
| 497 |
+
"eval_accuracy": 0.8548707753479126,
|
| 498 |
+
"eval_loss": 0.7450631260871887,
|
| 499 |
+
"eval_macro_f1": 0.8554842491486295,
|
| 500 |
+
"eval_runtime": 0.591,
|
| 501 |
+
"eval_samples_per_second": 851.136,
|
| 502 |
+
"eval_steps_per_second": 54.148,
|
| 503 |
+
"eval_weighted_f1": 0.8557116333768144,
|
| 504 |
+
"step": 1456
|
| 505 |
+
},
|
| 506 |
+
{
|
| 507 |
+
"epoch": 8.104827586206897,
|
| 508 |
+
"grad_norm": 125.99210357666016,
|
| 509 |
+
"learning_rate": 1.3223443223443224e-05,
|
| 510 |
+
"loss": 0.3478,
|
| 511 |
+
"step": 1475
|
| 512 |
+
},
|
| 513 |
+
{
|
| 514 |
+
"epoch": 8.242758620689655,
|
| 515 |
+
"grad_norm": 65.56297302246094,
|
| 516 |
+
"learning_rate": 1.3070818070818073e-05,
|
| 517 |
+
"loss": 0.1924,
|
| 518 |
+
"step": 1500
|
| 519 |
+
},
|
| 520 |
+
{
|
| 521 |
+
"epoch": 8.380689655172414,
|
| 522 |
+
"grad_norm": 16.62967872619629,
|
| 523 |
+
"learning_rate": 1.291819291819292e-05,
|
| 524 |
+
"loss": 0.2844,
|
| 525 |
+
"step": 1525
|
| 526 |
+
},
|
| 527 |
+
{
|
| 528 |
+
"epoch": 8.518620689655172,
|
| 529 |
+
"grad_norm": 0.2591530680656433,
|
| 530 |
+
"learning_rate": 1.2765567765567766e-05,
|
| 531 |
+
"loss": 0.3596,
|
| 532 |
+
"step": 1550
|
| 533 |
+
},
|
| 534 |
+
{
|
| 535 |
+
"epoch": 8.65655172413793,
|
| 536 |
+
"grad_norm": 155.3113250732422,
|
| 537 |
+
"learning_rate": 1.2612942612942613e-05,
|
| 538 |
+
"loss": 0.3918,
|
| 539 |
+
"step": 1575
|
| 540 |
+
},
|
| 541 |
+
{
|
| 542 |
+
"epoch": 8.79448275862069,
|
| 543 |
+
"grad_norm": 10.316084861755371,
|
| 544 |
+
"learning_rate": 1.2460317460317461e-05,
|
| 545 |
+
"loss": 0.351,
|
| 546 |
+
"step": 1600
|
| 547 |
+
},
|
| 548 |
+
{
|
| 549 |
+
"epoch": 8.932413793103448,
|
| 550 |
+
"grad_norm": 0.09080405533313751,
|
| 551 |
+
"learning_rate": 1.230769230769231e-05,
|
| 552 |
+
"loss": 0.1491,
|
| 553 |
+
"step": 1625
|
| 554 |
+
},
|
| 555 |
+
{
|
| 556 |
+
"epoch": 9.0,
|
| 557 |
+
"eval_accuracy": 0.8608349900596421,
|
| 558 |
+
"eval_loss": 0.7955670952796936,
|
| 559 |
+
"eval_macro_f1": 0.8604545280496444,
|
| 560 |
+
"eval_runtime": 0.5744,
|
| 561 |
+
"eval_samples_per_second": 875.751,
|
| 562 |
+
"eval_steps_per_second": 55.714,
|
| 563 |
+
"eval_weighted_f1": 0.8606573163234017,
|
| 564 |
+
"step": 1638
|
| 565 |
+
},
|
| 566 |
+
{
|
| 567 |
+
"epoch": 9.066206896551725,
|
| 568 |
+
"grad_norm": 35.8708610534668,
|
| 569 |
+
"learning_rate": 1.2155067155067155e-05,
|
| 570 |
+
"loss": 0.1611,
|
| 571 |
+
"step": 1650
|
| 572 |
+
},
|
| 573 |
+
{
|
| 574 |
+
"epoch": 9.204137931034483,
|
| 575 |
+
"grad_norm": 2.7454748153686523,
|
| 576 |
+
"learning_rate": 1.2002442002442003e-05,
|
| 577 |
+
"loss": 0.279,
|
| 578 |
+
"step": 1675
|
| 579 |
+
},
|
| 580 |
+
{
|
| 581 |
+
"epoch": 9.34206896551724,
|
| 582 |
+
"grad_norm": 267.3585205078125,
|
| 583 |
+
"learning_rate": 1.1849816849816852e-05,
|
| 584 |
+
"loss": 0.1466,
|
| 585 |
+
"step": 1700
|
| 586 |
+
},
|
| 587 |
+
{
|
| 588 |
+
"epoch": 9.48,
|
| 589 |
+
"grad_norm": 195.13316345214844,
|
| 590 |
+
"learning_rate": 1.1697191697191698e-05,
|
| 591 |
+
"loss": 0.3403,
|
| 592 |
+
"step": 1725
|
| 593 |
+
},
|
| 594 |
+
{
|
| 595 |
+
"epoch": 9.617931034482758,
|
| 596 |
+
"grad_norm": 0.22231760621070862,
|
| 597 |
+
"learning_rate": 1.1544566544566545e-05,
|
| 598 |
+
"loss": 0.1618,
|
| 599 |
+
"step": 1750
|
| 600 |
+
},
|
| 601 |
+
{
|
| 602 |
+
"epoch": 9.755862068965516,
|
| 603 |
+
"grad_norm": 0.14777538180351257,
|
| 604 |
+
"learning_rate": 1.1391941391941392e-05,
|
| 605 |
+
"loss": 0.2352,
|
| 606 |
+
"step": 1775
|
| 607 |
+
},
|
| 608 |
+
{
|
| 609 |
+
"epoch": 9.893793103448276,
|
| 610 |
+
"grad_norm": 27.933271408081055,
|
| 611 |
+
"learning_rate": 1.123931623931624e-05,
|
| 612 |
+
"loss": 0.2912,
|
| 613 |
+
"step": 1800
|
| 614 |
+
},
|
| 615 |
+
{
|
| 616 |
+
"epoch": 10.0,
|
| 617 |
+
"eval_accuracy": 0.8707753479125249,
|
| 618 |
+
"eval_loss": 0.8268889784812927,
|
| 619 |
+
"eval_macro_f1": 0.869309032967074,
|
| 620 |
+
"eval_runtime": 0.5821,
|
| 621 |
+
"eval_samples_per_second": 864.083,
|
| 622 |
+
"eval_steps_per_second": 54.972,
|
| 623 |
+
"eval_weighted_f1": 0.8693453848174799,
|
| 624 |
+
"step": 1820
|
| 625 |
+
},
|
| 626 |
+
{
|
| 627 |
+
"epoch": 10.027586206896551,
|
| 628 |
+
"grad_norm": 0.05576696991920471,
|
| 629 |
+
"learning_rate": 1.1086691086691088e-05,
|
| 630 |
+
"loss": 0.0176,
|
| 631 |
+
"step": 1825
|
| 632 |
+
},
|
| 633 |
+
{
|
| 634 |
+
"epoch": 10.16551724137931,
|
| 635 |
+
"grad_norm": 0.09126320481300354,
|
| 636 |
+
"learning_rate": 1.0934065934065933e-05,
|
| 637 |
+
"loss": 0.154,
|
| 638 |
+
"step": 1850
|
| 639 |
+
},
|
| 640 |
+
{
|
| 641 |
+
"epoch": 10.303448275862069,
|
| 642 |
+
"grad_norm": 0.05391448363661766,
|
| 643 |
+
"learning_rate": 1.0781440781440782e-05,
|
| 644 |
+
"loss": 0.1238,
|
| 645 |
+
"step": 1875
|
| 646 |
+
},
|
| 647 |
+
{
|
| 648 |
+
"epoch": 10.441379310344828,
|
| 649 |
+
"grad_norm": 38.219852447509766,
|
| 650 |
+
"learning_rate": 1.062881562881563e-05,
|
| 651 |
+
"loss": 0.2173,
|
| 652 |
+
"step": 1900
|
| 653 |
+
},
|
| 654 |
+
{
|
| 655 |
+
"epoch": 10.579310344827586,
|
| 656 |
+
"grad_norm": 0.13331815600395203,
|
| 657 |
+
"learning_rate": 1.0476190476190477e-05,
|
| 658 |
+
"loss": 0.2177,
|
| 659 |
+
"step": 1925
|
| 660 |
+
},
|
| 661 |
+
{
|
| 662 |
+
"epoch": 10.717241379310344,
|
| 663 |
+
"grad_norm": 0.04378219693899155,
|
| 664 |
+
"learning_rate": 1.0323565323565325e-05,
|
| 665 |
+
"loss": 0.0169,
|
| 666 |
+
"step": 1950
|
| 667 |
+
},
|
| 668 |
+
{
|
| 669 |
+
"epoch": 10.855172413793104,
|
| 670 |
+
"grad_norm": 0.05484893172979355,
|
| 671 |
+
"learning_rate": 1.0170940170940172e-05,
|
| 672 |
+
"loss": 0.1468,
|
| 673 |
+
"step": 1975
|
| 674 |
+
},
|
| 675 |
+
{
|
| 676 |
+
"epoch": 10.993103448275862,
|
| 677 |
+
"grad_norm": 41.13323211669922,
|
| 678 |
+
"learning_rate": 1.0018315018315019e-05,
|
| 679 |
+
"loss": 0.147,
|
| 680 |
+
"step": 2000
|
| 681 |
+
},
|
| 682 |
+
{
|
| 683 |
+
"epoch": 11.0,
|
| 684 |
+
"eval_accuracy": 0.852882703777336,
|
| 685 |
+
"eval_loss": 0.9407215118408203,
|
| 686 |
+
"eval_macro_f1": 0.8529817844817731,
|
| 687 |
+
"eval_runtime": 0.7279,
|
| 688 |
+
"eval_samples_per_second": 691.051,
|
| 689 |
+
"eval_steps_per_second": 43.963,
|
| 690 |
+
"eval_weighted_f1": 0.8530347155842798,
|
| 691 |
+
"step": 2002
|
| 692 |
+
},
|
| 693 |
+
{
|
| 694 |
+
"epoch": 11.126896551724139,
|
| 695 |
+
"grad_norm": 186.65052795410156,
|
| 696 |
+
"learning_rate": 9.865689865689867e-06,
|
| 697 |
+
"loss": 0.1559,
|
| 698 |
+
"step": 2025
|
| 699 |
+
},
|
| 700 |
+
{
|
| 701 |
+
"epoch": 11.264827586206897,
|
| 702 |
+
"grad_norm": 3.6744449138641357,
|
| 703 |
+
"learning_rate": 9.713064713064714e-06,
|
| 704 |
+
"loss": 0.1353,
|
| 705 |
+
"step": 2050
|
| 706 |
+
},
|
| 707 |
+
{
|
| 708 |
+
"epoch": 11.402758620689655,
|
| 709 |
+
"grad_norm": 0.29433631896972656,
|
| 710 |
+
"learning_rate": 9.560439560439562e-06,
|
| 711 |
+
"loss": 0.1238,
|
| 712 |
+
"step": 2075
|
| 713 |
+
},
|
| 714 |
+
{
|
| 715 |
+
"epoch": 11.540689655172415,
|
| 716 |
+
"grad_norm": 10.499589920043945,
|
| 717 |
+
"learning_rate": 9.407814407814409e-06,
|
| 718 |
+
"loss": 0.0302,
|
| 719 |
+
"step": 2100
|
| 720 |
+
},
|
| 721 |
+
{
|
| 722 |
+
"epoch": 11.678620689655173,
|
| 723 |
+
"grad_norm": 0.3906395435333252,
|
| 724 |
+
"learning_rate": 9.255189255189256e-06,
|
| 725 |
+
"loss": 0.2127,
|
| 726 |
+
"step": 2125
|
| 727 |
+
},
|
| 728 |
+
{
|
| 729 |
+
"epoch": 11.81655172413793,
|
| 730 |
+
"grad_norm": 0.05349710211157799,
|
| 731 |
+
"learning_rate": 9.102564102564104e-06,
|
| 732 |
+
"loss": 0.0853,
|
| 733 |
+
"step": 2150
|
| 734 |
+
},
|
| 735 |
+
{
|
| 736 |
+
"epoch": 11.95448275862069,
|
| 737 |
+
"grad_norm": 0.6620715260505676,
|
| 738 |
+
"learning_rate": 8.94993894993895e-06,
|
| 739 |
+
"loss": 0.2229,
|
| 740 |
+
"step": 2175
|
| 741 |
+
},
|
| 742 |
+
{
|
| 743 |
+
"epoch": 12.0,
|
| 744 |
+
"eval_accuracy": 0.8648111332007953,
|
| 745 |
+
"eval_loss": 0.8919137716293335,
|
| 746 |
+
"eval_macro_f1": 0.8627538493489771,
|
| 747 |
+
"eval_runtime": 0.5828,
|
| 748 |
+
"eval_samples_per_second": 863.102,
|
| 749 |
+
"eval_steps_per_second": 54.909,
|
| 750 |
+
"eval_weighted_f1": 0.8630339956507285,
|
| 751 |
+
"step": 2184
|
| 752 |
+
},
|
| 753 |
+
{
|
| 754 |
+
"epoch": 12.088275862068965,
|
| 755 |
+
"grad_norm": 0.24117650091648102,
|
| 756 |
+
"learning_rate": 8.797313797313797e-06,
|
| 757 |
+
"loss": 0.1231,
|
| 758 |
+
"step": 2200
|
| 759 |
+
},
|
| 760 |
+
{
|
| 761 |
+
"epoch": 12.226206896551725,
|
| 762 |
+
"grad_norm": 0.5760621428489685,
|
| 763 |
+
"learning_rate": 8.644688644688646e-06,
|
| 764 |
+
"loss": 0.0774,
|
| 765 |
+
"step": 2225
|
| 766 |
+
},
|
| 767 |
+
{
|
| 768 |
+
"epoch": 12.364137931034483,
|
| 769 |
+
"grad_norm": 0.44422709941864014,
|
| 770 |
+
"learning_rate": 8.492063492063492e-06,
|
| 771 |
+
"loss": 0.2168,
|
| 772 |
+
"step": 2250
|
| 773 |
+
},
|
| 774 |
+
{
|
| 775 |
+
"epoch": 12.50206896551724,
|
| 776 |
+
"grad_norm": 0.11377622932195663,
|
| 777 |
+
"learning_rate": 8.33943833943834e-06,
|
| 778 |
+
"loss": 0.008,
|
| 779 |
+
"step": 2275
|
| 780 |
+
},
|
| 781 |
+
{
|
| 782 |
+
"epoch": 12.64,
|
| 783 |
+
"grad_norm": 0.04294457659125328,
|
| 784 |
+
"learning_rate": 8.186813186813188e-06,
|
| 785 |
+
"loss": 0.1057,
|
| 786 |
+
"step": 2300
|
| 787 |
+
},
|
| 788 |
+
{
|
| 789 |
+
"epoch": 12.777931034482759,
|
| 790 |
+
"grad_norm": 0.3207155466079712,
|
| 791 |
+
"learning_rate": 8.034188034188036e-06,
|
| 792 |
+
"loss": 0.0884,
|
| 793 |
+
"step": 2325
|
| 794 |
+
},
|
| 795 |
+
{
|
| 796 |
+
"epoch": 12.915862068965517,
|
| 797 |
+
"grad_norm": 0.06088980287313461,
|
| 798 |
+
"learning_rate": 7.881562881562883e-06,
|
| 799 |
+
"loss": 0.0636,
|
| 800 |
+
"step": 2350
|
| 801 |
+
},
|
| 802 |
+
{
|
| 803 |
+
"epoch": 13.0,
|
| 804 |
+
"eval_accuracy": 0.8687872763419483,
|
| 805 |
+
"eval_loss": 0.9100527167320251,
|
| 806 |
+
"eval_macro_f1": 0.8680771243408606,
|
| 807 |
+
"eval_runtime": 0.5824,
|
| 808 |
+
"eval_samples_per_second": 863.732,
|
| 809 |
+
"eval_steps_per_second": 54.949,
|
| 810 |
+
"eval_weighted_f1": 0.8683255136491761,
|
| 811 |
+
"step": 2366
|
| 812 |
+
}
|
| 813 |
+
],
|
| 814 |
+
"logging_steps": 25,
|
| 815 |
+
"max_steps": 3640,
|
| 816 |
+
"num_input_tokens_seen": 0,
|
| 817 |
+
"num_train_epochs": 20,
|
| 818 |
+
"save_steps": 500,
|
| 819 |
+
"stateful_callbacks": {
|
| 820 |
+
"EarlyStoppingCallback": {
|
| 821 |
+
"args": {
|
| 822 |
+
"early_stopping_patience": 3,
|
| 823 |
+
"early_stopping_threshold": 0.0
|
| 824 |
+
},
|
| 825 |
+
"attributes": {
|
| 826 |
+
"early_stopping_patience_counter": 3
|
| 827 |
+
}
|
| 828 |
+
},
|
| 829 |
+
"TrainerControl": {
|
| 830 |
+
"args": {
|
| 831 |
+
"should_epoch_stop": false,
|
| 832 |
+
"should_evaluate": false,
|
| 833 |
+
"should_log": false,
|
| 834 |
+
"should_save": true,
|
| 835 |
+
"should_training_stop": true
|
| 836 |
+
},
|
| 837 |
+
"attributes": {}
|
| 838 |
+
}
|
| 839 |
+
},
|
| 840 |
+
"total_flos": 0.0,
|
| 841 |
+
"train_batch_size": 8,
|
| 842 |
+
"trial_name": null,
|
| 843 |
+
"trial_params": null
|
| 844 |
+
}
|
ood-vifactcheck-bamibert-topic-domain-03-s43-20260827T194124Z-3858bb98/trainer/checkpoint-1365/special_tokens_map.json
ADDED
|
@@ -0,0 +1,51 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": {
|
| 3 |
+
"content": "<s>",
|
| 4 |
+
"lstrip": false,
|
| 5 |
+
"normalized": false,
|
| 6 |
+
"rstrip": false,
|
| 7 |
+
"single_word": false
|
| 8 |
+
},
|
| 9 |
+
"cls_token": {
|
| 10 |
+
"content": "<s>",
|
| 11 |
+
"lstrip": false,
|
| 12 |
+
"normalized": false,
|
| 13 |
+
"rstrip": false,
|
| 14 |
+
"single_word": false
|
| 15 |
+
},
|
| 16 |
+
"eos_token": {
|
| 17 |
+
"content": "</s>",
|
| 18 |
+
"lstrip": false,
|
| 19 |
+
"normalized": false,
|
| 20 |
+
"rstrip": false,
|
| 21 |
+
"single_word": false
|
| 22 |
+
},
|
| 23 |
+
"mask_token": {
|
| 24 |
+
"content": "<mask>",
|
| 25 |
+
"lstrip": false,
|
| 26 |
+
"normalized": false,
|
| 27 |
+
"rstrip": false,
|
| 28 |
+
"single_word": false
|
| 29 |
+
},
|
| 30 |
+
"pad_token": {
|
| 31 |
+
"content": "<pad>",
|
| 32 |
+
"lstrip": false,
|
| 33 |
+
"normalized": false,
|
| 34 |
+
"rstrip": false,
|
| 35 |
+
"single_word": false
|
| 36 |
+
},
|
| 37 |
+
"sep_token": {
|
| 38 |
+
"content": "</s>",
|
| 39 |
+
"lstrip": false,
|
| 40 |
+
"normalized": false,
|
| 41 |
+
"rstrip": false,
|
| 42 |
+
"single_word": false
|
| 43 |
+
},
|
| 44 |
+
"unk_token": {
|
| 45 |
+
"content": "<unk>",
|
| 46 |
+
"lstrip": false,
|
| 47 |
+
"normalized": false,
|
| 48 |
+
"rstrip": false,
|
| 49 |
+
"single_word": false
|
| 50 |
+
}
|
| 51 |
+
}
|
ood-vifactcheck-bamibert-topic-domain-03-s43-20260827T194124Z-3858bb98/trainer/checkpoint-1365/tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-bamibert-topic-domain-03-s43-20260827T194124Z-3858bb98/trainer/checkpoint-1365/tokenizer_config.json
ADDED
|
@@ -0,0 +1,55 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"added_tokens_decoder": {
|
| 3 |
+
"0": {
|
| 4 |
+
"content": "<s>",
|
| 5 |
+
"lstrip": false,
|
| 6 |
+
"normalized": false,
|
| 7 |
+
"rstrip": false,
|
| 8 |
+
"single_word": false,
|
| 9 |
+
"special": true
|
| 10 |
+
},
|
| 11 |
+
"1": {
|
| 12 |
+
"content": "<pad>",
|
| 13 |
+
"lstrip": false,
|
| 14 |
+
"normalized": false,
|
| 15 |
+
"rstrip": false,
|
| 16 |
+
"single_word": false,
|
| 17 |
+
"special": true
|
| 18 |
+
},
|
| 19 |
+
"2": {
|
| 20 |
+
"content": "</s>",
|
| 21 |
+
"lstrip": false,
|
| 22 |
+
"normalized": false,
|
| 23 |
+
"rstrip": false,
|
| 24 |
+
"single_word": false,
|
| 25 |
+
"special": true
|
| 26 |
+
},
|
| 27 |
+
"3": {
|
| 28 |
+
"content": "<unk>",
|
| 29 |
+
"lstrip": false,
|
| 30 |
+
"normalized": false,
|
| 31 |
+
"rstrip": false,
|
| 32 |
+
"single_word": false,
|
| 33 |
+
"special": true
|
| 34 |
+
},
|
| 35 |
+
"20480": {
|
| 36 |
+
"content": "<mask>",
|
| 37 |
+
"lstrip": false,
|
| 38 |
+
"normalized": false,
|
| 39 |
+
"rstrip": false,
|
| 40 |
+
"single_word": false,
|
| 41 |
+
"special": true
|
| 42 |
+
}
|
| 43 |
+
},
|
| 44 |
+
"bos_token": "<s>",
|
| 45 |
+
"clean_up_tokenization_spaces": false,
|
| 46 |
+
"cls_token": "<s>",
|
| 47 |
+
"eos_token": "</s>",
|
| 48 |
+
"extra_special_tokens": {},
|
| 49 |
+
"mask_token": "<mask>",
|
| 50 |
+
"model_max_length": 2048,
|
| 51 |
+
"pad_token": "<pad>",
|
| 52 |
+
"sep_token": "</s>",
|
| 53 |
+
"tokenizer_class": "XLMRobertaTokenizerFast",
|
| 54 |
+
"unk_token": "<unk>"
|
| 55 |
+
}
|
ood-vifactcheck-bamibert-topic-domain-03-s43-20260827T194124Z-3858bb98/trainer/checkpoint-1365/trainer_state.json
ADDED
|
@@ -0,0 +1,498 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 1365,
|
| 3 |
+
"best_metric": 0.8413026357896819,
|
| 4 |
+
"best_model_checkpoint": "/root/VNEncoder/results/checkpoints/ood-vifactcheck-bamibert-topic-domain-03-s43-20260827T194124Z-3858bb98/trainer/checkpoint-1365",
|
| 5 |
+
"epoch": 7.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 1365,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.12836970474967907,
|
| 14 |
+
"grad_norm": 125.39506530761719,
|
| 15 |
+
"learning_rate": 1.230769230769231e-06,
|
| 16 |
+
"loss": 4.4301,
|
| 17 |
+
"step": 25
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.25673940949935814,
|
| 21 |
+
"grad_norm": 21.763351440429688,
|
| 22 |
+
"learning_rate": 2.512820512820513e-06,
|
| 23 |
+
"loss": 4.4124,
|
| 24 |
+
"step": 50
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.3851091142490372,
|
| 28 |
+
"grad_norm": 26.893400192260742,
|
| 29 |
+
"learning_rate": 3.794871794871795e-06,
|
| 30 |
+
"loss": 4.4034,
|
| 31 |
+
"step": 75
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.5134788189987163,
|
| 35 |
+
"grad_norm": 11.034686088562012,
|
| 36 |
+
"learning_rate": 5.076923076923077e-06,
|
| 37 |
+
"loss": 4.3708,
|
| 38 |
+
"step": 100
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.6418485237483954,
|
| 42 |
+
"grad_norm": 20.18111801147461,
|
| 43 |
+
"learning_rate": 6.358974358974359e-06,
|
| 44 |
+
"loss": 4.3665,
|
| 45 |
+
"step": 125
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.7702182284980744,
|
| 49 |
+
"grad_norm": 12.325210571289062,
|
| 50 |
+
"learning_rate": 7.641025641025641e-06,
|
| 51 |
+
"loss": 4.3211,
|
| 52 |
+
"step": 150
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"epoch": 0.8985879332477535,
|
| 56 |
+
"grad_norm": 13.29734992980957,
|
| 57 |
+
"learning_rate": 8.923076923076925e-06,
|
| 58 |
+
"loss": 4.2091,
|
| 59 |
+
"step": 175
|
| 60 |
+
},
|
| 61 |
+
{
|
| 62 |
+
"epoch": 1.0,
|
| 63 |
+
"eval_accuracy": 0.5449826989619377,
|
| 64 |
+
"eval_loss": 0.967367947101593,
|
| 65 |
+
"eval_macro_f1": 0.5398821587253875,
|
| 66 |
+
"eval_runtime": 0.689,
|
| 67 |
+
"eval_samples_per_second": 838.955,
|
| 68 |
+
"eval_steps_per_second": 53.705,
|
| 69 |
+
"eval_weighted_f1": 0.5417829452457378,
|
| 70 |
+
"step": 195
|
| 71 |
+
},
|
| 72 |
+
{
|
| 73 |
+
"epoch": 1.0256739409499358,
|
| 74 |
+
"grad_norm": 34.1318359375,
|
| 75 |
+
"learning_rate": 1.0205128205128205e-05,
|
| 76 |
+
"loss": 4.0754,
|
| 77 |
+
"step": 200
|
| 78 |
+
},
|
| 79 |
+
{
|
| 80 |
+
"epoch": 1.1540436456996148,
|
| 81 |
+
"grad_norm": 26.616811752319336,
|
| 82 |
+
"learning_rate": 1.1487179487179487e-05,
|
| 83 |
+
"loss": 3.7534,
|
| 84 |
+
"step": 225
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 1.282413350449294,
|
| 88 |
+
"grad_norm": 56.07647705078125,
|
| 89 |
+
"learning_rate": 1.2769230769230769e-05,
|
| 90 |
+
"loss": 3.5805,
|
| 91 |
+
"step": 250
|
| 92 |
+
},
|
| 93 |
+
{
|
| 94 |
+
"epoch": 1.410783055198973,
|
| 95 |
+
"grad_norm": 41.29022979736328,
|
| 96 |
+
"learning_rate": 1.405128205128205e-05,
|
| 97 |
+
"loss": 3.2171,
|
| 98 |
+
"step": 275
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 1.539152759948652,
|
| 102 |
+
"grad_norm": 58.27585983276367,
|
| 103 |
+
"learning_rate": 1.5333333333333334e-05,
|
| 104 |
+
"loss": 2.8051,
|
| 105 |
+
"step": 300
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 1.6675224646983313,
|
| 109 |
+
"grad_norm": 41.664772033691406,
|
| 110 |
+
"learning_rate": 1.6615384615384618e-05,
|
| 111 |
+
"loss": 2.6895,
|
| 112 |
+
"step": 325
|
| 113 |
+
},
|
| 114 |
+
{
|
| 115 |
+
"epoch": 1.7958921694480101,
|
| 116 |
+
"grad_norm": 38.359046936035156,
|
| 117 |
+
"learning_rate": 1.78974358974359e-05,
|
| 118 |
+
"loss": 2.5902,
|
| 119 |
+
"step": 350
|
| 120 |
+
},
|
| 121 |
+
{
|
| 122 |
+
"epoch": 1.9242618741976893,
|
| 123 |
+
"grad_norm": 73.98174285888672,
|
| 124 |
+
"learning_rate": 1.9179487179487182e-05,
|
| 125 |
+
"loss": 2.5097,
|
| 126 |
+
"step": 375
|
| 127 |
+
},
|
| 128 |
+
{
|
| 129 |
+
"epoch": 2.0,
|
| 130 |
+
"eval_accuracy": 0.7906574394463668,
|
| 131 |
+
"eval_loss": 0.48849034309387207,
|
| 132 |
+
"eval_macro_f1": 0.7940493835954587,
|
| 133 |
+
"eval_runtime": 0.7563,
|
| 134 |
+
"eval_samples_per_second": 764.272,
|
| 135 |
+
"eval_steps_per_second": 48.924,
|
| 136 |
+
"eval_weighted_f1": 0.7943996562007074,
|
| 137 |
+
"step": 390
|
| 138 |
+
},
|
| 139 |
+
{
|
| 140 |
+
"epoch": 2.0513478818998716,
|
| 141 |
+
"grad_norm": 68.99273681640625,
|
| 142 |
+
"learning_rate": 1.994871794871795e-05,
|
| 143 |
+
"loss": 2.2107,
|
| 144 |
+
"step": 400
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 2.179717586649551,
|
| 148 |
+
"grad_norm": 34.362003326416016,
|
| 149 |
+
"learning_rate": 1.9806267806267806e-05,
|
| 150 |
+
"loss": 2.002,
|
| 151 |
+
"step": 425
|
| 152 |
+
},
|
| 153 |
+
{
|
| 154 |
+
"epoch": 2.3080872913992296,
|
| 155 |
+
"grad_norm": 40.58723068237305,
|
| 156 |
+
"learning_rate": 1.9663817663817667e-05,
|
| 157 |
+
"loss": 2.0017,
|
| 158 |
+
"step": 450
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 2.436456996148909,
|
| 162 |
+
"grad_norm": 40.8438720703125,
|
| 163 |
+
"learning_rate": 1.952136752136752e-05,
|
| 164 |
+
"loss": 2.1637,
|
| 165 |
+
"step": 475
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 2.564826700898588,
|
| 169 |
+
"grad_norm": 52.28580856323242,
|
| 170 |
+
"learning_rate": 1.9378917378917382e-05,
|
| 171 |
+
"loss": 2.0534,
|
| 172 |
+
"step": 500
|
| 173 |
+
},
|
| 174 |
+
{
|
| 175 |
+
"epoch": 2.693196405648267,
|
| 176 |
+
"grad_norm": 53.85476303100586,
|
| 177 |
+
"learning_rate": 1.923646723646724e-05,
|
| 178 |
+
"loss": 2.3043,
|
| 179 |
+
"step": 525
|
| 180 |
+
},
|
| 181 |
+
{
|
| 182 |
+
"epoch": 2.821566110397946,
|
| 183 |
+
"grad_norm": 46.694725036621094,
|
| 184 |
+
"learning_rate": 1.9094017094017097e-05,
|
| 185 |
+
"loss": 2.0307,
|
| 186 |
+
"step": 550
|
| 187 |
+
},
|
| 188 |
+
{
|
| 189 |
+
"epoch": 2.949935815147625,
|
| 190 |
+
"grad_norm": 43.3126335144043,
|
| 191 |
+
"learning_rate": 1.8951566951566954e-05,
|
| 192 |
+
"loss": 1.9716,
|
| 193 |
+
"step": 575
|
| 194 |
+
},
|
| 195 |
+
{
|
| 196 |
+
"epoch": 3.0,
|
| 197 |
+
"eval_accuracy": 0.8235294117647058,
|
| 198 |
+
"eval_loss": 0.48839518427848816,
|
| 199 |
+
"eval_macro_f1": 0.8205261134241102,
|
| 200 |
+
"eval_runtime": 0.6894,
|
| 201 |
+
"eval_samples_per_second": 838.461,
|
| 202 |
+
"eval_steps_per_second": 53.673,
|
| 203 |
+
"eval_weighted_f1": 0.8206248219819976,
|
| 204 |
+
"step": 585
|
| 205 |
+
},
|
| 206 |
+
{
|
| 207 |
+
"epoch": 3.0770218228498076,
|
| 208 |
+
"grad_norm": 29.892906188964844,
|
| 209 |
+
"learning_rate": 1.880911680911681e-05,
|
| 210 |
+
"loss": 1.585,
|
| 211 |
+
"step": 600
|
| 212 |
+
},
|
| 213 |
+
{
|
| 214 |
+
"epoch": 3.2053915275994864,
|
| 215 |
+
"grad_norm": 73.68291473388672,
|
| 216 |
+
"learning_rate": 1.866666666666667e-05,
|
| 217 |
+
"loss": 1.2214,
|
| 218 |
+
"step": 625
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 3.3337612323491657,
|
| 222 |
+
"grad_norm": 62.524200439453125,
|
| 223 |
+
"learning_rate": 1.8524216524216526e-05,
|
| 224 |
+
"loss": 1.4244,
|
| 225 |
+
"step": 650
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"epoch": 3.462130937098845,
|
| 229 |
+
"grad_norm": 28.509963989257812,
|
| 230 |
+
"learning_rate": 1.8381766381766383e-05,
|
| 231 |
+
"loss": 1.7582,
|
| 232 |
+
"step": 675
|
| 233 |
+
},
|
| 234 |
+
{
|
| 235 |
+
"epoch": 3.5905006418485237,
|
| 236 |
+
"grad_norm": 31.75885009765625,
|
| 237 |
+
"learning_rate": 1.823931623931624e-05,
|
| 238 |
+
"loss": 1.4261,
|
| 239 |
+
"step": 700
|
| 240 |
+
},
|
| 241 |
+
{
|
| 242 |
+
"epoch": 3.718870346598203,
|
| 243 |
+
"grad_norm": 69.422119140625,
|
| 244 |
+
"learning_rate": 1.8096866096866098e-05,
|
| 245 |
+
"loss": 1.353,
|
| 246 |
+
"step": 725
|
| 247 |
+
},
|
| 248 |
+
{
|
| 249 |
+
"epoch": 3.8472400513478817,
|
| 250 |
+
"grad_norm": 64.89217376708984,
|
| 251 |
+
"learning_rate": 1.7954415954415955e-05,
|
| 252 |
+
"loss": 1.7283,
|
| 253 |
+
"step": 750
|
| 254 |
+
},
|
| 255 |
+
{
|
| 256 |
+
"epoch": 3.975609756097561,
|
| 257 |
+
"grad_norm": 36.044368743896484,
|
| 258 |
+
"learning_rate": 1.7811965811965813e-05,
|
| 259 |
+
"loss": 1.5472,
|
| 260 |
+
"step": 775
|
| 261 |
+
},
|
| 262 |
+
{
|
| 263 |
+
"epoch": 4.0,
|
| 264 |
+
"eval_accuracy": 0.8235294117647058,
|
| 265 |
+
"eval_loss": 0.5028790831565857,
|
| 266 |
+
"eval_macro_f1": 0.8240687789818887,
|
| 267 |
+
"eval_runtime": 0.6855,
|
| 268 |
+
"eval_samples_per_second": 843.119,
|
| 269 |
+
"eval_steps_per_second": 53.971,
|
| 270 |
+
"eval_weighted_f1": 0.824345881112668,
|
| 271 |
+
"step": 780
|
| 272 |
+
},
|
| 273 |
+
{
|
| 274 |
+
"epoch": 4.102695763799743,
|
| 275 |
+
"grad_norm": 23.045373916625977,
|
| 276 |
+
"learning_rate": 1.766951566951567e-05,
|
| 277 |
+
"loss": 1.1473,
|
| 278 |
+
"step": 800
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 4.2310654685494224,
|
| 282 |
+
"grad_norm": 32.24570083618164,
|
| 283 |
+
"learning_rate": 1.7527065527065528e-05,
|
| 284 |
+
"loss": 1.0577,
|
| 285 |
+
"step": 825
|
| 286 |
+
},
|
| 287 |
+
{
|
| 288 |
+
"epoch": 4.359435173299102,
|
| 289 |
+
"grad_norm": 66.68640899658203,
|
| 290 |
+
"learning_rate": 1.7384615384615385e-05,
|
| 291 |
+
"loss": 1.0787,
|
| 292 |
+
"step": 850
|
| 293 |
+
},
|
| 294 |
+
{
|
| 295 |
+
"epoch": 4.487804878048781,
|
| 296 |
+
"grad_norm": 27.690940856933594,
|
| 297 |
+
"learning_rate": 1.7242165242165242e-05,
|
| 298 |
+
"loss": 1.0185,
|
| 299 |
+
"step": 875
|
| 300 |
+
},
|
| 301 |
+
{
|
| 302 |
+
"epoch": 4.616174582798459,
|
| 303 |
+
"grad_norm": 10.199621200561523,
|
| 304 |
+
"learning_rate": 1.7099715099715103e-05,
|
| 305 |
+
"loss": 1.0083,
|
| 306 |
+
"step": 900
|
| 307 |
+
},
|
| 308 |
+
{
|
| 309 |
+
"epoch": 4.7445442875481385,
|
| 310 |
+
"grad_norm": 75.68169403076172,
|
| 311 |
+
"learning_rate": 1.6957264957264957e-05,
|
| 312 |
+
"loss": 0.9427,
|
| 313 |
+
"step": 925
|
| 314 |
+
},
|
| 315 |
+
{
|
| 316 |
+
"epoch": 4.872913992297818,
|
| 317 |
+
"grad_norm": 36.596832275390625,
|
| 318 |
+
"learning_rate": 1.6814814814814818e-05,
|
| 319 |
+
"loss": 1.0547,
|
| 320 |
+
"step": 950
|
| 321 |
+
},
|
| 322 |
+
{
|
| 323 |
+
"epoch": 5.0,
|
| 324 |
+
"grad_norm": 22.902475357055664,
|
| 325 |
+
"learning_rate": 1.6672364672364675e-05,
|
| 326 |
+
"loss": 0.9542,
|
| 327 |
+
"step": 975
|
| 328 |
+
},
|
| 329 |
+
{
|
| 330 |
+
"epoch": 5.0,
|
| 331 |
+
"eval_accuracy": 0.8321799307958477,
|
| 332 |
+
"eval_loss": 0.5185052156448364,
|
| 333 |
+
"eval_macro_f1": 0.8312759321908921,
|
| 334 |
+
"eval_runtime": 0.6828,
|
| 335 |
+
"eval_samples_per_second": 846.485,
|
| 336 |
+
"eval_steps_per_second": 54.187,
|
| 337 |
+
"eval_weighted_f1": 0.8313969451901295,
|
| 338 |
+
"step": 975
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 5.128369704749679,
|
| 342 |
+
"grad_norm": 39.282318115234375,
|
| 343 |
+
"learning_rate": 1.652991452991453e-05,
|
| 344 |
+
"loss": 0.72,
|
| 345 |
+
"step": 1000
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 5.2567394094993585,
|
| 349 |
+
"grad_norm": 44.60416030883789,
|
| 350 |
+
"learning_rate": 1.638746438746439e-05,
|
| 351 |
+
"loss": 0.6888,
|
| 352 |
+
"step": 1025
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 5.385109114249037,
|
| 356 |
+
"grad_norm": 108.50936889648438,
|
| 357 |
+
"learning_rate": 1.6245014245014247e-05,
|
| 358 |
+
"loss": 0.6149,
|
| 359 |
+
"step": 1050
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 5.513478818998716,
|
| 363 |
+
"grad_norm": 22.086883544921875,
|
| 364 |
+
"learning_rate": 1.6102564102564105e-05,
|
| 365 |
+
"loss": 0.818,
|
| 366 |
+
"step": 1075
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 5.641848523748395,
|
| 370 |
+
"grad_norm": 69.6677474975586,
|
| 371 |
+
"learning_rate": 1.5960113960113962e-05,
|
| 372 |
+
"loss": 0.8463,
|
| 373 |
+
"step": 1100
|
| 374 |
+
},
|
| 375 |
+
{
|
| 376 |
+
"epoch": 5.7702182284980745,
|
| 377 |
+
"grad_norm": 21.006175994873047,
|
| 378 |
+
"learning_rate": 1.581766381766382e-05,
|
| 379 |
+
"loss": 0.7456,
|
| 380 |
+
"step": 1125
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 5.898587933247754,
|
| 384 |
+
"grad_norm": 149.01759338378906,
|
| 385 |
+
"learning_rate": 1.5675213675213677e-05,
|
| 386 |
+
"loss": 0.7288,
|
| 387 |
+
"step": 1150
|
| 388 |
+
},
|
| 389 |
+
{
|
| 390 |
+
"epoch": 6.0,
|
| 391 |
+
"eval_accuracy": 0.8252595155709342,
|
| 392 |
+
"eval_loss": 0.6358432769775391,
|
| 393 |
+
"eval_macro_f1": 0.8255422616715045,
|
| 394 |
+
"eval_runtime": 0.6992,
|
| 395 |
+
"eval_samples_per_second": 826.661,
|
| 396 |
+
"eval_steps_per_second": 52.918,
|
| 397 |
+
"eval_weighted_f1": 0.825789478675633,
|
| 398 |
+
"step": 1170
|
| 399 |
+
},
|
| 400 |
+
{
|
| 401 |
+
"epoch": 6.025673940949936,
|
| 402 |
+
"grad_norm": 38.22592544555664,
|
| 403 |
+
"learning_rate": 1.5532763532763534e-05,
|
| 404 |
+
"loss": 0.8363,
|
| 405 |
+
"step": 1175
|
| 406 |
+
},
|
| 407 |
+
{
|
| 408 |
+
"epoch": 6.154043645699615,
|
| 409 |
+
"grad_norm": 23.07004165649414,
|
| 410 |
+
"learning_rate": 1.539031339031339e-05,
|
| 411 |
+
"loss": 0.6124,
|
| 412 |
+
"step": 1200
|
| 413 |
+
},
|
| 414 |
+
{
|
| 415 |
+
"epoch": 6.282413350449294,
|
| 416 |
+
"grad_norm": 28.19152069091797,
|
| 417 |
+
"learning_rate": 1.5247863247863249e-05,
|
| 418 |
+
"loss": 0.6765,
|
| 419 |
+
"step": 1225
|
| 420 |
+
},
|
| 421 |
+
{
|
| 422 |
+
"epoch": 6.410783055198973,
|
| 423 |
+
"grad_norm": 9.752038955688477,
|
| 424 |
+
"learning_rate": 1.5105413105413106e-05,
|
| 425 |
+
"loss": 0.3463,
|
| 426 |
+
"step": 1250
|
| 427 |
+
},
|
| 428 |
+
{
|
| 429 |
+
"epoch": 6.539152759948652,
|
| 430 |
+
"grad_norm": 74.94021606445312,
|
| 431 |
+
"learning_rate": 1.4962962962962964e-05,
|
| 432 |
+
"loss": 0.4683,
|
| 433 |
+
"step": 1275
|
| 434 |
+
},
|
| 435 |
+
{
|
| 436 |
+
"epoch": 6.667522464698331,
|
| 437 |
+
"grad_norm": 76.32896423339844,
|
| 438 |
+
"learning_rate": 1.4820512820512823e-05,
|
| 439 |
+
"loss": 0.6158,
|
| 440 |
+
"step": 1300
|
| 441 |
+
},
|
| 442 |
+
{
|
| 443 |
+
"epoch": 6.7958921694480106,
|
| 444 |
+
"grad_norm": 11.797883987426758,
|
| 445 |
+
"learning_rate": 1.4678062678062678e-05,
|
| 446 |
+
"loss": 0.6217,
|
| 447 |
+
"step": 1325
|
| 448 |
+
},
|
| 449 |
+
{
|
| 450 |
+
"epoch": 6.92426187419769,
|
| 451 |
+
"grad_norm": 20.600454330444336,
|
| 452 |
+
"learning_rate": 1.4535612535612537e-05,
|
| 453 |
+
"loss": 0.5526,
|
| 454 |
+
"step": 1350
|
| 455 |
+
},
|
| 456 |
+
{
|
| 457 |
+
"epoch": 7.0,
|
| 458 |
+
"eval_accuracy": 0.842560553633218,
|
| 459 |
+
"eval_loss": 0.7795780897140503,
|
| 460 |
+
"eval_macro_f1": 0.8413026357896819,
|
| 461 |
+
"eval_runtime": 0.6988,
|
| 462 |
+
"eval_samples_per_second": 827.136,
|
| 463 |
+
"eval_steps_per_second": 52.948,
|
| 464 |
+
"eval_weighted_f1": 0.8413283564596132,
|
| 465 |
+
"step": 1365
|
| 466 |
+
}
|
| 467 |
+
],
|
| 468 |
+
"logging_steps": 25,
|
| 469 |
+
"max_steps": 3900,
|
| 470 |
+
"num_input_tokens_seen": 0,
|
| 471 |
+
"num_train_epochs": 20,
|
| 472 |
+
"save_steps": 500,
|
| 473 |
+
"stateful_callbacks": {
|
| 474 |
+
"EarlyStoppingCallback": {
|
| 475 |
+
"args": {
|
| 476 |
+
"early_stopping_patience": 3,
|
| 477 |
+
"early_stopping_threshold": 0.0
|
| 478 |
+
},
|
| 479 |
+
"attributes": {
|
| 480 |
+
"early_stopping_patience_counter": 0
|
| 481 |
+
}
|
| 482 |
+
},
|
| 483 |
+
"TrainerControl": {
|
| 484 |
+
"args": {
|
| 485 |
+
"should_epoch_stop": false,
|
| 486 |
+
"should_evaluate": false,
|
| 487 |
+
"should_log": false,
|
| 488 |
+
"should_save": true,
|
| 489 |
+
"should_training_stop": false
|
| 490 |
+
},
|
| 491 |
+
"attributes": {}
|
| 492 |
+
}
|
| 493 |
+
},
|
| 494 |
+
"total_flos": 0.0,
|
| 495 |
+
"train_batch_size": 8,
|
| 496 |
+
"trial_name": null,
|
| 497 |
+
"trial_params": null
|
| 498 |
+
}
|
ood-vifactcheck-bamibert-topic-domain-03-s43-20260827T194124Z-3858bb98/trainer/checkpoint-1950/special_tokens_map.json
ADDED
|
@@ -0,0 +1,51 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": {
|
| 3 |
+
"content": "<s>",
|
| 4 |
+
"lstrip": false,
|
| 5 |
+
"normalized": false,
|
| 6 |
+
"rstrip": false,
|
| 7 |
+
"single_word": false
|
| 8 |
+
},
|
| 9 |
+
"cls_token": {
|
| 10 |
+
"content": "<s>",
|
| 11 |
+
"lstrip": false,
|
| 12 |
+
"normalized": false,
|
| 13 |
+
"rstrip": false,
|
| 14 |
+
"single_word": false
|
| 15 |
+
},
|
| 16 |
+
"eos_token": {
|
| 17 |
+
"content": "</s>",
|
| 18 |
+
"lstrip": false,
|
| 19 |
+
"normalized": false,
|
| 20 |
+
"rstrip": false,
|
| 21 |
+
"single_word": false
|
| 22 |
+
},
|
| 23 |
+
"mask_token": {
|
| 24 |
+
"content": "<mask>",
|
| 25 |
+
"lstrip": false,
|
| 26 |
+
"normalized": false,
|
| 27 |
+
"rstrip": false,
|
| 28 |
+
"single_word": false
|
| 29 |
+
},
|
| 30 |
+
"pad_token": {
|
| 31 |
+
"content": "<pad>",
|
| 32 |
+
"lstrip": false,
|
| 33 |
+
"normalized": false,
|
| 34 |
+
"rstrip": false,
|
| 35 |
+
"single_word": false
|
| 36 |
+
},
|
| 37 |
+
"sep_token": {
|
| 38 |
+
"content": "</s>",
|
| 39 |
+
"lstrip": false,
|
| 40 |
+
"normalized": false,
|
| 41 |
+
"rstrip": false,
|
| 42 |
+
"single_word": false
|
| 43 |
+
},
|
| 44 |
+
"unk_token": {
|
| 45 |
+
"content": "<unk>",
|
| 46 |
+
"lstrip": false,
|
| 47 |
+
"normalized": false,
|
| 48 |
+
"rstrip": false,
|
| 49 |
+
"single_word": false
|
| 50 |
+
}
|
| 51 |
+
}
|
ood-vifactcheck-bamibert-topic-domain-03-s43-20260827T194124Z-3858bb98/trainer/checkpoint-1950/tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-bamibert-topic-domain-03-s43-20260827T194124Z-3858bb98/trainer/checkpoint-1950/tokenizer_config.json
ADDED
|
@@ -0,0 +1,55 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"added_tokens_decoder": {
|
| 3 |
+
"0": {
|
| 4 |
+
"content": "<s>",
|
| 5 |
+
"lstrip": false,
|
| 6 |
+
"normalized": false,
|
| 7 |
+
"rstrip": false,
|
| 8 |
+
"single_word": false,
|
| 9 |
+
"special": true
|
| 10 |
+
},
|
| 11 |
+
"1": {
|
| 12 |
+
"content": "<pad>",
|
| 13 |
+
"lstrip": false,
|
| 14 |
+
"normalized": false,
|
| 15 |
+
"rstrip": false,
|
| 16 |
+
"single_word": false,
|
| 17 |
+
"special": true
|
| 18 |
+
},
|
| 19 |
+
"2": {
|
| 20 |
+
"content": "</s>",
|
| 21 |
+
"lstrip": false,
|
| 22 |
+
"normalized": false,
|
| 23 |
+
"rstrip": false,
|
| 24 |
+
"single_word": false,
|
| 25 |
+
"special": true
|
| 26 |
+
},
|
| 27 |
+
"3": {
|
| 28 |
+
"content": "<unk>",
|
| 29 |
+
"lstrip": false,
|
| 30 |
+
"normalized": false,
|
| 31 |
+
"rstrip": false,
|
| 32 |
+
"single_word": false,
|
| 33 |
+
"special": true
|
| 34 |
+
},
|
| 35 |
+
"20480": {
|
| 36 |
+
"content": "<mask>",
|
| 37 |
+
"lstrip": false,
|
| 38 |
+
"normalized": false,
|
| 39 |
+
"rstrip": false,
|
| 40 |
+
"single_word": false,
|
| 41 |
+
"special": true
|
| 42 |
+
}
|
| 43 |
+
},
|
| 44 |
+
"bos_token": "<s>",
|
| 45 |
+
"clean_up_tokenization_spaces": false,
|
| 46 |
+
"cls_token": "<s>",
|
| 47 |
+
"eos_token": "</s>",
|
| 48 |
+
"extra_special_tokens": {},
|
| 49 |
+
"mask_token": "<mask>",
|
| 50 |
+
"model_max_length": 2048,
|
| 51 |
+
"pad_token": "<pad>",
|
| 52 |
+
"sep_token": "</s>",
|
| 53 |
+
"tokenizer_class": "XLMRobertaTokenizerFast",
|
| 54 |
+
"unk_token": "<unk>"
|
| 55 |
+
}
|
ood-vifactcheck-bamibert-topic-domain-03-s43-20260827T194124Z-3858bb98/trainer/checkpoint-1950/trainer_state.json
ADDED
|
@@ -0,0 +1,699 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 1365,
|
| 3 |
+
"best_metric": 0.8413026357896819,
|
| 4 |
+
"best_model_checkpoint": "/root/VNEncoder/results/checkpoints/ood-vifactcheck-bamibert-topic-domain-03-s43-20260827T194124Z-3858bb98/trainer/checkpoint-1365",
|
| 5 |
+
"epoch": 10.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 1950,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.12836970474967907,
|
| 14 |
+
"grad_norm": 125.39506530761719,
|
| 15 |
+
"learning_rate": 1.230769230769231e-06,
|
| 16 |
+
"loss": 4.4301,
|
| 17 |
+
"step": 25
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.25673940949935814,
|
| 21 |
+
"grad_norm": 21.763351440429688,
|
| 22 |
+
"learning_rate": 2.512820512820513e-06,
|
| 23 |
+
"loss": 4.4124,
|
| 24 |
+
"step": 50
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.3851091142490372,
|
| 28 |
+
"grad_norm": 26.893400192260742,
|
| 29 |
+
"learning_rate": 3.794871794871795e-06,
|
| 30 |
+
"loss": 4.4034,
|
| 31 |
+
"step": 75
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.5134788189987163,
|
| 35 |
+
"grad_norm": 11.034686088562012,
|
| 36 |
+
"learning_rate": 5.076923076923077e-06,
|
| 37 |
+
"loss": 4.3708,
|
| 38 |
+
"step": 100
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.6418485237483954,
|
| 42 |
+
"grad_norm": 20.18111801147461,
|
| 43 |
+
"learning_rate": 6.358974358974359e-06,
|
| 44 |
+
"loss": 4.3665,
|
| 45 |
+
"step": 125
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.7702182284980744,
|
| 49 |
+
"grad_norm": 12.325210571289062,
|
| 50 |
+
"learning_rate": 7.641025641025641e-06,
|
| 51 |
+
"loss": 4.3211,
|
| 52 |
+
"step": 150
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"epoch": 0.8985879332477535,
|
| 56 |
+
"grad_norm": 13.29734992980957,
|
| 57 |
+
"learning_rate": 8.923076923076925e-06,
|
| 58 |
+
"loss": 4.2091,
|
| 59 |
+
"step": 175
|
| 60 |
+
},
|
| 61 |
+
{
|
| 62 |
+
"epoch": 1.0,
|
| 63 |
+
"eval_accuracy": 0.5449826989619377,
|
| 64 |
+
"eval_loss": 0.967367947101593,
|
| 65 |
+
"eval_macro_f1": 0.5398821587253875,
|
| 66 |
+
"eval_runtime": 0.689,
|
| 67 |
+
"eval_samples_per_second": 838.955,
|
| 68 |
+
"eval_steps_per_second": 53.705,
|
| 69 |
+
"eval_weighted_f1": 0.5417829452457378,
|
| 70 |
+
"step": 195
|
| 71 |
+
},
|
| 72 |
+
{
|
| 73 |
+
"epoch": 1.0256739409499358,
|
| 74 |
+
"grad_norm": 34.1318359375,
|
| 75 |
+
"learning_rate": 1.0205128205128205e-05,
|
| 76 |
+
"loss": 4.0754,
|
| 77 |
+
"step": 200
|
| 78 |
+
},
|
| 79 |
+
{
|
| 80 |
+
"epoch": 1.1540436456996148,
|
| 81 |
+
"grad_norm": 26.616811752319336,
|
| 82 |
+
"learning_rate": 1.1487179487179487e-05,
|
| 83 |
+
"loss": 3.7534,
|
| 84 |
+
"step": 225
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 1.282413350449294,
|
| 88 |
+
"grad_norm": 56.07647705078125,
|
| 89 |
+
"learning_rate": 1.2769230769230769e-05,
|
| 90 |
+
"loss": 3.5805,
|
| 91 |
+
"step": 250
|
| 92 |
+
},
|
| 93 |
+
{
|
| 94 |
+
"epoch": 1.410783055198973,
|
| 95 |
+
"grad_norm": 41.29022979736328,
|
| 96 |
+
"learning_rate": 1.405128205128205e-05,
|
| 97 |
+
"loss": 3.2171,
|
| 98 |
+
"step": 275
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 1.539152759948652,
|
| 102 |
+
"grad_norm": 58.27585983276367,
|
| 103 |
+
"learning_rate": 1.5333333333333334e-05,
|
| 104 |
+
"loss": 2.8051,
|
| 105 |
+
"step": 300
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 1.6675224646983313,
|
| 109 |
+
"grad_norm": 41.664772033691406,
|
| 110 |
+
"learning_rate": 1.6615384615384618e-05,
|
| 111 |
+
"loss": 2.6895,
|
| 112 |
+
"step": 325
|
| 113 |
+
},
|
| 114 |
+
{
|
| 115 |
+
"epoch": 1.7958921694480101,
|
| 116 |
+
"grad_norm": 38.359046936035156,
|
| 117 |
+
"learning_rate": 1.78974358974359e-05,
|
| 118 |
+
"loss": 2.5902,
|
| 119 |
+
"step": 350
|
| 120 |
+
},
|
| 121 |
+
{
|
| 122 |
+
"epoch": 1.9242618741976893,
|
| 123 |
+
"grad_norm": 73.98174285888672,
|
| 124 |
+
"learning_rate": 1.9179487179487182e-05,
|
| 125 |
+
"loss": 2.5097,
|
| 126 |
+
"step": 375
|
| 127 |
+
},
|
| 128 |
+
{
|
| 129 |
+
"epoch": 2.0,
|
| 130 |
+
"eval_accuracy": 0.7906574394463668,
|
| 131 |
+
"eval_loss": 0.48849034309387207,
|
| 132 |
+
"eval_macro_f1": 0.7940493835954587,
|
| 133 |
+
"eval_runtime": 0.7563,
|
| 134 |
+
"eval_samples_per_second": 764.272,
|
| 135 |
+
"eval_steps_per_second": 48.924,
|
| 136 |
+
"eval_weighted_f1": 0.7943996562007074,
|
| 137 |
+
"step": 390
|
| 138 |
+
},
|
| 139 |
+
{
|
| 140 |
+
"epoch": 2.0513478818998716,
|
| 141 |
+
"grad_norm": 68.99273681640625,
|
| 142 |
+
"learning_rate": 1.994871794871795e-05,
|
| 143 |
+
"loss": 2.2107,
|
| 144 |
+
"step": 400
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 2.179717586649551,
|
| 148 |
+
"grad_norm": 34.362003326416016,
|
| 149 |
+
"learning_rate": 1.9806267806267806e-05,
|
| 150 |
+
"loss": 2.002,
|
| 151 |
+
"step": 425
|
| 152 |
+
},
|
| 153 |
+
{
|
| 154 |
+
"epoch": 2.3080872913992296,
|
| 155 |
+
"grad_norm": 40.58723068237305,
|
| 156 |
+
"learning_rate": 1.9663817663817667e-05,
|
| 157 |
+
"loss": 2.0017,
|
| 158 |
+
"step": 450
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 2.436456996148909,
|
| 162 |
+
"grad_norm": 40.8438720703125,
|
| 163 |
+
"learning_rate": 1.952136752136752e-05,
|
| 164 |
+
"loss": 2.1637,
|
| 165 |
+
"step": 475
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 2.564826700898588,
|
| 169 |
+
"grad_norm": 52.28580856323242,
|
| 170 |
+
"learning_rate": 1.9378917378917382e-05,
|
| 171 |
+
"loss": 2.0534,
|
| 172 |
+
"step": 500
|
| 173 |
+
},
|
| 174 |
+
{
|
| 175 |
+
"epoch": 2.693196405648267,
|
| 176 |
+
"grad_norm": 53.85476303100586,
|
| 177 |
+
"learning_rate": 1.923646723646724e-05,
|
| 178 |
+
"loss": 2.3043,
|
| 179 |
+
"step": 525
|
| 180 |
+
},
|
| 181 |
+
{
|
| 182 |
+
"epoch": 2.821566110397946,
|
| 183 |
+
"grad_norm": 46.694725036621094,
|
| 184 |
+
"learning_rate": 1.9094017094017097e-05,
|
| 185 |
+
"loss": 2.0307,
|
| 186 |
+
"step": 550
|
| 187 |
+
},
|
| 188 |
+
{
|
| 189 |
+
"epoch": 2.949935815147625,
|
| 190 |
+
"grad_norm": 43.3126335144043,
|
| 191 |
+
"learning_rate": 1.8951566951566954e-05,
|
| 192 |
+
"loss": 1.9716,
|
| 193 |
+
"step": 575
|
| 194 |
+
},
|
| 195 |
+
{
|
| 196 |
+
"epoch": 3.0,
|
| 197 |
+
"eval_accuracy": 0.8235294117647058,
|
| 198 |
+
"eval_loss": 0.48839518427848816,
|
| 199 |
+
"eval_macro_f1": 0.8205261134241102,
|
| 200 |
+
"eval_runtime": 0.6894,
|
| 201 |
+
"eval_samples_per_second": 838.461,
|
| 202 |
+
"eval_steps_per_second": 53.673,
|
| 203 |
+
"eval_weighted_f1": 0.8206248219819976,
|
| 204 |
+
"step": 585
|
| 205 |
+
},
|
| 206 |
+
{
|
| 207 |
+
"epoch": 3.0770218228498076,
|
| 208 |
+
"grad_norm": 29.892906188964844,
|
| 209 |
+
"learning_rate": 1.880911680911681e-05,
|
| 210 |
+
"loss": 1.585,
|
| 211 |
+
"step": 600
|
| 212 |
+
},
|
| 213 |
+
{
|
| 214 |
+
"epoch": 3.2053915275994864,
|
| 215 |
+
"grad_norm": 73.68291473388672,
|
| 216 |
+
"learning_rate": 1.866666666666667e-05,
|
| 217 |
+
"loss": 1.2214,
|
| 218 |
+
"step": 625
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 3.3337612323491657,
|
| 222 |
+
"grad_norm": 62.524200439453125,
|
| 223 |
+
"learning_rate": 1.8524216524216526e-05,
|
| 224 |
+
"loss": 1.4244,
|
| 225 |
+
"step": 650
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"epoch": 3.462130937098845,
|
| 229 |
+
"grad_norm": 28.509963989257812,
|
| 230 |
+
"learning_rate": 1.8381766381766383e-05,
|
| 231 |
+
"loss": 1.7582,
|
| 232 |
+
"step": 675
|
| 233 |
+
},
|
| 234 |
+
{
|
| 235 |
+
"epoch": 3.5905006418485237,
|
| 236 |
+
"grad_norm": 31.75885009765625,
|
| 237 |
+
"learning_rate": 1.823931623931624e-05,
|
| 238 |
+
"loss": 1.4261,
|
| 239 |
+
"step": 700
|
| 240 |
+
},
|
| 241 |
+
{
|
| 242 |
+
"epoch": 3.718870346598203,
|
| 243 |
+
"grad_norm": 69.422119140625,
|
| 244 |
+
"learning_rate": 1.8096866096866098e-05,
|
| 245 |
+
"loss": 1.353,
|
| 246 |
+
"step": 725
|
| 247 |
+
},
|
| 248 |
+
{
|
| 249 |
+
"epoch": 3.8472400513478817,
|
| 250 |
+
"grad_norm": 64.89217376708984,
|
| 251 |
+
"learning_rate": 1.7954415954415955e-05,
|
| 252 |
+
"loss": 1.7283,
|
| 253 |
+
"step": 750
|
| 254 |
+
},
|
| 255 |
+
{
|
| 256 |
+
"epoch": 3.975609756097561,
|
| 257 |
+
"grad_norm": 36.044368743896484,
|
| 258 |
+
"learning_rate": 1.7811965811965813e-05,
|
| 259 |
+
"loss": 1.5472,
|
| 260 |
+
"step": 775
|
| 261 |
+
},
|
| 262 |
+
{
|
| 263 |
+
"epoch": 4.0,
|
| 264 |
+
"eval_accuracy": 0.8235294117647058,
|
| 265 |
+
"eval_loss": 0.5028790831565857,
|
| 266 |
+
"eval_macro_f1": 0.8240687789818887,
|
| 267 |
+
"eval_runtime": 0.6855,
|
| 268 |
+
"eval_samples_per_second": 843.119,
|
| 269 |
+
"eval_steps_per_second": 53.971,
|
| 270 |
+
"eval_weighted_f1": 0.824345881112668,
|
| 271 |
+
"step": 780
|
| 272 |
+
},
|
| 273 |
+
{
|
| 274 |
+
"epoch": 4.102695763799743,
|
| 275 |
+
"grad_norm": 23.045373916625977,
|
| 276 |
+
"learning_rate": 1.766951566951567e-05,
|
| 277 |
+
"loss": 1.1473,
|
| 278 |
+
"step": 800
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 4.2310654685494224,
|
| 282 |
+
"grad_norm": 32.24570083618164,
|
| 283 |
+
"learning_rate": 1.7527065527065528e-05,
|
| 284 |
+
"loss": 1.0577,
|
| 285 |
+
"step": 825
|
| 286 |
+
},
|
| 287 |
+
{
|
| 288 |
+
"epoch": 4.359435173299102,
|
| 289 |
+
"grad_norm": 66.68640899658203,
|
| 290 |
+
"learning_rate": 1.7384615384615385e-05,
|
| 291 |
+
"loss": 1.0787,
|
| 292 |
+
"step": 850
|
| 293 |
+
},
|
| 294 |
+
{
|
| 295 |
+
"epoch": 4.487804878048781,
|
| 296 |
+
"grad_norm": 27.690940856933594,
|
| 297 |
+
"learning_rate": 1.7242165242165242e-05,
|
| 298 |
+
"loss": 1.0185,
|
| 299 |
+
"step": 875
|
| 300 |
+
},
|
| 301 |
+
{
|
| 302 |
+
"epoch": 4.616174582798459,
|
| 303 |
+
"grad_norm": 10.199621200561523,
|
| 304 |
+
"learning_rate": 1.7099715099715103e-05,
|
| 305 |
+
"loss": 1.0083,
|
| 306 |
+
"step": 900
|
| 307 |
+
},
|
| 308 |
+
{
|
| 309 |
+
"epoch": 4.7445442875481385,
|
| 310 |
+
"grad_norm": 75.68169403076172,
|
| 311 |
+
"learning_rate": 1.6957264957264957e-05,
|
| 312 |
+
"loss": 0.9427,
|
| 313 |
+
"step": 925
|
| 314 |
+
},
|
| 315 |
+
{
|
| 316 |
+
"epoch": 4.872913992297818,
|
| 317 |
+
"grad_norm": 36.596832275390625,
|
| 318 |
+
"learning_rate": 1.6814814814814818e-05,
|
| 319 |
+
"loss": 1.0547,
|
| 320 |
+
"step": 950
|
| 321 |
+
},
|
| 322 |
+
{
|
| 323 |
+
"epoch": 5.0,
|
| 324 |
+
"grad_norm": 22.902475357055664,
|
| 325 |
+
"learning_rate": 1.6672364672364675e-05,
|
| 326 |
+
"loss": 0.9542,
|
| 327 |
+
"step": 975
|
| 328 |
+
},
|
| 329 |
+
{
|
| 330 |
+
"epoch": 5.0,
|
| 331 |
+
"eval_accuracy": 0.8321799307958477,
|
| 332 |
+
"eval_loss": 0.5185052156448364,
|
| 333 |
+
"eval_macro_f1": 0.8312759321908921,
|
| 334 |
+
"eval_runtime": 0.6828,
|
| 335 |
+
"eval_samples_per_second": 846.485,
|
| 336 |
+
"eval_steps_per_second": 54.187,
|
| 337 |
+
"eval_weighted_f1": 0.8313969451901295,
|
| 338 |
+
"step": 975
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 5.128369704749679,
|
| 342 |
+
"grad_norm": 39.282318115234375,
|
| 343 |
+
"learning_rate": 1.652991452991453e-05,
|
| 344 |
+
"loss": 0.72,
|
| 345 |
+
"step": 1000
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 5.2567394094993585,
|
| 349 |
+
"grad_norm": 44.60416030883789,
|
| 350 |
+
"learning_rate": 1.638746438746439e-05,
|
| 351 |
+
"loss": 0.6888,
|
| 352 |
+
"step": 1025
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 5.385109114249037,
|
| 356 |
+
"grad_norm": 108.50936889648438,
|
| 357 |
+
"learning_rate": 1.6245014245014247e-05,
|
| 358 |
+
"loss": 0.6149,
|
| 359 |
+
"step": 1050
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 5.513478818998716,
|
| 363 |
+
"grad_norm": 22.086883544921875,
|
| 364 |
+
"learning_rate": 1.6102564102564105e-05,
|
| 365 |
+
"loss": 0.818,
|
| 366 |
+
"step": 1075
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 5.641848523748395,
|
| 370 |
+
"grad_norm": 69.6677474975586,
|
| 371 |
+
"learning_rate": 1.5960113960113962e-05,
|
| 372 |
+
"loss": 0.8463,
|
| 373 |
+
"step": 1100
|
| 374 |
+
},
|
| 375 |
+
{
|
| 376 |
+
"epoch": 5.7702182284980745,
|
| 377 |
+
"grad_norm": 21.006175994873047,
|
| 378 |
+
"learning_rate": 1.581766381766382e-05,
|
| 379 |
+
"loss": 0.7456,
|
| 380 |
+
"step": 1125
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 5.898587933247754,
|
| 384 |
+
"grad_norm": 149.01759338378906,
|
| 385 |
+
"learning_rate": 1.5675213675213677e-05,
|
| 386 |
+
"loss": 0.7288,
|
| 387 |
+
"step": 1150
|
| 388 |
+
},
|
| 389 |
+
{
|
| 390 |
+
"epoch": 6.0,
|
| 391 |
+
"eval_accuracy": 0.8252595155709342,
|
| 392 |
+
"eval_loss": 0.6358432769775391,
|
| 393 |
+
"eval_macro_f1": 0.8255422616715045,
|
| 394 |
+
"eval_runtime": 0.6992,
|
| 395 |
+
"eval_samples_per_second": 826.661,
|
| 396 |
+
"eval_steps_per_second": 52.918,
|
| 397 |
+
"eval_weighted_f1": 0.825789478675633,
|
| 398 |
+
"step": 1170
|
| 399 |
+
},
|
| 400 |
+
{
|
| 401 |
+
"epoch": 6.025673940949936,
|
| 402 |
+
"grad_norm": 38.22592544555664,
|
| 403 |
+
"learning_rate": 1.5532763532763534e-05,
|
| 404 |
+
"loss": 0.8363,
|
| 405 |
+
"step": 1175
|
| 406 |
+
},
|
| 407 |
+
{
|
| 408 |
+
"epoch": 6.154043645699615,
|
| 409 |
+
"grad_norm": 23.07004165649414,
|
| 410 |
+
"learning_rate": 1.539031339031339e-05,
|
| 411 |
+
"loss": 0.6124,
|
| 412 |
+
"step": 1200
|
| 413 |
+
},
|
| 414 |
+
{
|
| 415 |
+
"epoch": 6.282413350449294,
|
| 416 |
+
"grad_norm": 28.19152069091797,
|
| 417 |
+
"learning_rate": 1.5247863247863249e-05,
|
| 418 |
+
"loss": 0.6765,
|
| 419 |
+
"step": 1225
|
| 420 |
+
},
|
| 421 |
+
{
|
| 422 |
+
"epoch": 6.410783055198973,
|
| 423 |
+
"grad_norm": 9.752038955688477,
|
| 424 |
+
"learning_rate": 1.5105413105413106e-05,
|
| 425 |
+
"loss": 0.3463,
|
| 426 |
+
"step": 1250
|
| 427 |
+
},
|
| 428 |
+
{
|
| 429 |
+
"epoch": 6.539152759948652,
|
| 430 |
+
"grad_norm": 74.94021606445312,
|
| 431 |
+
"learning_rate": 1.4962962962962964e-05,
|
| 432 |
+
"loss": 0.4683,
|
| 433 |
+
"step": 1275
|
| 434 |
+
},
|
| 435 |
+
{
|
| 436 |
+
"epoch": 6.667522464698331,
|
| 437 |
+
"grad_norm": 76.32896423339844,
|
| 438 |
+
"learning_rate": 1.4820512820512823e-05,
|
| 439 |
+
"loss": 0.6158,
|
| 440 |
+
"step": 1300
|
| 441 |
+
},
|
| 442 |
+
{
|
| 443 |
+
"epoch": 6.7958921694480106,
|
| 444 |
+
"grad_norm": 11.797883987426758,
|
| 445 |
+
"learning_rate": 1.4678062678062678e-05,
|
| 446 |
+
"loss": 0.6217,
|
| 447 |
+
"step": 1325
|
| 448 |
+
},
|
| 449 |
+
{
|
| 450 |
+
"epoch": 6.92426187419769,
|
| 451 |
+
"grad_norm": 20.600454330444336,
|
| 452 |
+
"learning_rate": 1.4535612535612537e-05,
|
| 453 |
+
"loss": 0.5526,
|
| 454 |
+
"step": 1350
|
| 455 |
+
},
|
| 456 |
+
{
|
| 457 |
+
"epoch": 7.0,
|
| 458 |
+
"eval_accuracy": 0.842560553633218,
|
| 459 |
+
"eval_loss": 0.7795780897140503,
|
| 460 |
+
"eval_macro_f1": 0.8413026357896819,
|
| 461 |
+
"eval_runtime": 0.6988,
|
| 462 |
+
"eval_samples_per_second": 827.136,
|
| 463 |
+
"eval_steps_per_second": 52.948,
|
| 464 |
+
"eval_weighted_f1": 0.8413283564596132,
|
| 465 |
+
"step": 1365
|
| 466 |
+
},
|
| 467 |
+
{
|
| 468 |
+
"epoch": 7.051347881899872,
|
| 469 |
+
"grad_norm": 108.33708190917969,
|
| 470 |
+
"learning_rate": 1.4393162393162395e-05,
|
| 471 |
+
"loss": 0.3783,
|
| 472 |
+
"step": 1375
|
| 473 |
+
},
|
| 474 |
+
{
|
| 475 |
+
"epoch": 7.17971758664955,
|
| 476 |
+
"grad_norm": 101.03681945800781,
|
| 477 |
+
"learning_rate": 1.425071225071225e-05,
|
| 478 |
+
"loss": 0.6161,
|
| 479 |
+
"step": 1400
|
| 480 |
+
},
|
| 481 |
+
{
|
| 482 |
+
"epoch": 7.30808729139923,
|
| 483 |
+
"grad_norm": 17.134838104248047,
|
| 484 |
+
"learning_rate": 1.410826210826211e-05,
|
| 485 |
+
"loss": 0.5734,
|
| 486 |
+
"step": 1425
|
| 487 |
+
},
|
| 488 |
+
{
|
| 489 |
+
"epoch": 7.436456996148909,
|
| 490 |
+
"grad_norm": 28.270328521728516,
|
| 491 |
+
"learning_rate": 1.3965811965811967e-05,
|
| 492 |
+
"loss": 0.5108,
|
| 493 |
+
"step": 1450
|
| 494 |
+
},
|
| 495 |
+
{
|
| 496 |
+
"epoch": 7.564826700898588,
|
| 497 |
+
"grad_norm": 114.18989562988281,
|
| 498 |
+
"learning_rate": 1.3823361823361824e-05,
|
| 499 |
+
"loss": 0.4191,
|
| 500 |
+
"step": 1475
|
| 501 |
+
},
|
| 502 |
+
{
|
| 503 |
+
"epoch": 7.693196405648267,
|
| 504 |
+
"grad_norm": 58.8578987121582,
|
| 505 |
+
"learning_rate": 1.3680911680911682e-05,
|
| 506 |
+
"loss": 0.4254,
|
| 507 |
+
"step": 1500
|
| 508 |
+
},
|
| 509 |
+
{
|
| 510 |
+
"epoch": 7.821566110397946,
|
| 511 |
+
"grad_norm": 56.59989547729492,
|
| 512 |
+
"learning_rate": 1.353846153846154e-05,
|
| 513 |
+
"loss": 0.446,
|
| 514 |
+
"step": 1525
|
| 515 |
+
},
|
| 516 |
+
{
|
| 517 |
+
"epoch": 7.949935815147625,
|
| 518 |
+
"grad_norm": 11.307324409484863,
|
| 519 |
+
"learning_rate": 1.3396011396011396e-05,
|
| 520 |
+
"loss": 0.4073,
|
| 521 |
+
"step": 1550
|
| 522 |
+
},
|
| 523 |
+
{
|
| 524 |
+
"epoch": 8.0,
|
| 525 |
+
"eval_accuracy": 0.8304498269896193,
|
| 526 |
+
"eval_loss": 0.8245477676391602,
|
| 527 |
+
"eval_macro_f1": 0.8309519603637251,
|
| 528 |
+
"eval_runtime": 0.6907,
|
| 529 |
+
"eval_samples_per_second": 836.845,
|
| 530 |
+
"eval_steps_per_second": 53.57,
|
| 531 |
+
"eval_weighted_f1": 0.8307721754495618,
|
| 532 |
+
"step": 1560
|
| 533 |
+
},
|
| 534 |
+
{
|
| 535 |
+
"epoch": 8.077021822849808,
|
| 536 |
+
"grad_norm": 128.13253784179688,
|
| 537 |
+
"learning_rate": 1.3253561253561255e-05,
|
| 538 |
+
"loss": 0.3294,
|
| 539 |
+
"step": 1575
|
| 540 |
+
},
|
| 541 |
+
{
|
| 542 |
+
"epoch": 8.205391527599486,
|
| 543 |
+
"grad_norm": 8.089134216308594,
|
| 544 |
+
"learning_rate": 1.3111111111111113e-05,
|
| 545 |
+
"loss": 0.3693,
|
| 546 |
+
"step": 1600
|
| 547 |
+
},
|
| 548 |
+
{
|
| 549 |
+
"epoch": 8.333761232349165,
|
| 550 |
+
"grad_norm": 26.494630813598633,
|
| 551 |
+
"learning_rate": 1.2968660968660968e-05,
|
| 552 |
+
"loss": 0.2205,
|
| 553 |
+
"step": 1625
|
| 554 |
+
},
|
| 555 |
+
{
|
| 556 |
+
"epoch": 8.462130937098845,
|
| 557 |
+
"grad_norm": 86.39105987548828,
|
| 558 |
+
"learning_rate": 1.2826210826210828e-05,
|
| 559 |
+
"loss": 0.4239,
|
| 560 |
+
"step": 1650
|
| 561 |
+
},
|
| 562 |
+
{
|
| 563 |
+
"epoch": 8.590500641848523,
|
| 564 |
+
"grad_norm": 1.5420070886611938,
|
| 565 |
+
"learning_rate": 1.2683760683760685e-05,
|
| 566 |
+
"loss": 0.3752,
|
| 567 |
+
"step": 1675
|
| 568 |
+
},
|
| 569 |
+
{
|
| 570 |
+
"epoch": 8.718870346598203,
|
| 571 |
+
"grad_norm": 126.31005859375,
|
| 572 |
+
"learning_rate": 1.2541310541310542e-05,
|
| 573 |
+
"loss": 0.378,
|
| 574 |
+
"step": 1700
|
| 575 |
+
},
|
| 576 |
+
{
|
| 577 |
+
"epoch": 8.847240051347882,
|
| 578 |
+
"grad_norm": 75.65216064453125,
|
| 579 |
+
"learning_rate": 1.23988603988604e-05,
|
| 580 |
+
"loss": 0.4421,
|
| 581 |
+
"step": 1725
|
| 582 |
+
},
|
| 583 |
+
{
|
| 584 |
+
"epoch": 8.975609756097562,
|
| 585 |
+
"grad_norm": 4.920777320861816,
|
| 586 |
+
"learning_rate": 1.2256410256410259e-05,
|
| 587 |
+
"loss": 0.2636,
|
| 588 |
+
"step": 1750
|
| 589 |
+
},
|
| 590 |
+
{
|
| 591 |
+
"epoch": 9.0,
|
| 592 |
+
"eval_accuracy": 0.8269896193771626,
|
| 593 |
+
"eval_loss": 1.0120854377746582,
|
| 594 |
+
"eval_macro_f1": 0.8282596663406855,
|
| 595 |
+
"eval_runtime": 0.7899,
|
| 596 |
+
"eval_samples_per_second": 731.703,
|
| 597 |
+
"eval_steps_per_second": 46.839,
|
| 598 |
+
"eval_weighted_f1": 0.8282928488424939,
|
| 599 |
+
"step": 1755
|
| 600 |
+
},
|
| 601 |
+
{
|
| 602 |
+
"epoch": 9.102695763799744,
|
| 603 |
+
"grad_norm": 0.10547494143247604,
|
| 604 |
+
"learning_rate": 1.2113960113960114e-05,
|
| 605 |
+
"loss": 0.252,
|
| 606 |
+
"step": 1775
|
| 607 |
+
},
|
| 608 |
+
{
|
| 609 |
+
"epoch": 9.231065468549422,
|
| 610 |
+
"grad_norm": 56.235164642333984,
|
| 611 |
+
"learning_rate": 1.1971509971509972e-05,
|
| 612 |
+
"loss": 0.3121,
|
| 613 |
+
"step": 1800
|
| 614 |
+
},
|
| 615 |
+
{
|
| 616 |
+
"epoch": 9.3594351732991,
|
| 617 |
+
"grad_norm": 5.5893635749816895,
|
| 618 |
+
"learning_rate": 1.182905982905983e-05,
|
| 619 |
+
"loss": 0.3105,
|
| 620 |
+
"step": 1825
|
| 621 |
+
},
|
| 622 |
+
{
|
| 623 |
+
"epoch": 9.487804878048781,
|
| 624 |
+
"grad_norm": 0.08069439977407455,
|
| 625 |
+
"learning_rate": 1.1686609686609687e-05,
|
| 626 |
+
"loss": 0.166,
|
| 627 |
+
"step": 1850
|
| 628 |
+
},
|
| 629 |
+
{
|
| 630 |
+
"epoch": 9.61617458279846,
|
| 631 |
+
"grad_norm": 0.4873630702495575,
|
| 632 |
+
"learning_rate": 1.1544159544159546e-05,
|
| 633 |
+
"loss": 0.516,
|
| 634 |
+
"step": 1875
|
| 635 |
+
},
|
| 636 |
+
{
|
| 637 |
+
"epoch": 9.74454428754814,
|
| 638 |
+
"grad_norm": 3.2834293842315674,
|
| 639 |
+
"learning_rate": 1.1401709401709403e-05,
|
| 640 |
+
"loss": 0.2425,
|
| 641 |
+
"step": 1900
|
| 642 |
+
},
|
| 643 |
+
{
|
| 644 |
+
"epoch": 9.872913992297818,
|
| 645 |
+
"grad_norm": 8.29072093963623,
|
| 646 |
+
"learning_rate": 1.125925925925926e-05,
|
| 647 |
+
"loss": 0.4152,
|
| 648 |
+
"step": 1925
|
| 649 |
+
},
|
| 650 |
+
{
|
| 651 |
+
"epoch": 10.0,
|
| 652 |
+
"grad_norm": 0.07362756133079529,
|
| 653 |
+
"learning_rate": 1.1116809116809118e-05,
|
| 654 |
+
"loss": 0.238,
|
| 655 |
+
"step": 1950
|
| 656 |
+
},
|
| 657 |
+
{
|
| 658 |
+
"epoch": 10.0,
|
| 659 |
+
"eval_accuracy": 0.8321799307958477,
|
| 660 |
+
"eval_loss": 0.9812607169151306,
|
| 661 |
+
"eval_macro_f1": 0.8312622154499293,
|
| 662 |
+
"eval_runtime": 0.6823,
|
| 663 |
+
"eval_samples_per_second": 847.093,
|
| 664 |
+
"eval_steps_per_second": 54.226,
|
| 665 |
+
"eval_weighted_f1": 0.8315308489450359,
|
| 666 |
+
"step": 1950
|
| 667 |
+
}
|
| 668 |
+
],
|
| 669 |
+
"logging_steps": 25,
|
| 670 |
+
"max_steps": 3900,
|
| 671 |
+
"num_input_tokens_seen": 0,
|
| 672 |
+
"num_train_epochs": 20,
|
| 673 |
+
"save_steps": 500,
|
| 674 |
+
"stateful_callbacks": {
|
| 675 |
+
"EarlyStoppingCallback": {
|
| 676 |
+
"args": {
|
| 677 |
+
"early_stopping_patience": 3,
|
| 678 |
+
"early_stopping_threshold": 0.0
|
| 679 |
+
},
|
| 680 |
+
"attributes": {
|
| 681 |
+
"early_stopping_patience_counter": 3
|
| 682 |
+
}
|
| 683 |
+
},
|
| 684 |
+
"TrainerControl": {
|
| 685 |
+
"args": {
|
| 686 |
+
"should_epoch_stop": false,
|
| 687 |
+
"should_evaluate": false,
|
| 688 |
+
"should_log": false,
|
| 689 |
+
"should_save": true,
|
| 690 |
+
"should_training_stop": true
|
| 691 |
+
},
|
| 692 |
+
"attributes": {}
|
| 693 |
+
}
|
| 694 |
+
},
|
| 695 |
+
"total_flos": 0.0,
|
| 696 |
+
"train_batch_size": 8,
|
| 697 |
+
"trial_name": null,
|
| 698 |
+
"trial_params": null
|
| 699 |
+
}
|
ood-vifactcheck-bamibert-topic-domain-04-s43-20260827T194549Z-3858bb98/trainer/checkpoint-1692/special_tokens_map.json
ADDED
|
@@ -0,0 +1,51 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": {
|
| 3 |
+
"content": "<s>",
|
| 4 |
+
"lstrip": false,
|
| 5 |
+
"normalized": false,
|
| 6 |
+
"rstrip": false,
|
| 7 |
+
"single_word": false
|
| 8 |
+
},
|
| 9 |
+
"cls_token": {
|
| 10 |
+
"content": "<s>",
|
| 11 |
+
"lstrip": false,
|
| 12 |
+
"normalized": false,
|
| 13 |
+
"rstrip": false,
|
| 14 |
+
"single_word": false
|
| 15 |
+
},
|
| 16 |
+
"eos_token": {
|
| 17 |
+
"content": "</s>",
|
| 18 |
+
"lstrip": false,
|
| 19 |
+
"normalized": false,
|
| 20 |
+
"rstrip": false,
|
| 21 |
+
"single_word": false
|
| 22 |
+
},
|
| 23 |
+
"mask_token": {
|
| 24 |
+
"content": "<mask>",
|
| 25 |
+
"lstrip": false,
|
| 26 |
+
"normalized": false,
|
| 27 |
+
"rstrip": false,
|
| 28 |
+
"single_word": false
|
| 29 |
+
},
|
| 30 |
+
"pad_token": {
|
| 31 |
+
"content": "<pad>",
|
| 32 |
+
"lstrip": false,
|
| 33 |
+
"normalized": false,
|
| 34 |
+
"rstrip": false,
|
| 35 |
+
"single_word": false
|
| 36 |
+
},
|
| 37 |
+
"sep_token": {
|
| 38 |
+
"content": "</s>",
|
| 39 |
+
"lstrip": false,
|
| 40 |
+
"normalized": false,
|
| 41 |
+
"rstrip": false,
|
| 42 |
+
"single_word": false
|
| 43 |
+
},
|
| 44 |
+
"unk_token": {
|
| 45 |
+
"content": "<unk>",
|
| 46 |
+
"lstrip": false,
|
| 47 |
+
"normalized": false,
|
| 48 |
+
"rstrip": false,
|
| 49 |
+
"single_word": false
|
| 50 |
+
}
|
| 51 |
+
}
|
ood-vifactcheck-bamibert-topic-domain-04-s43-20260827T194549Z-3858bb98/trainer/checkpoint-1692/tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-bamibert-topic-domain-04-s43-20260827T194549Z-3858bb98/trainer/checkpoint-1692/tokenizer_config.json
ADDED
|
@@ -0,0 +1,55 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"added_tokens_decoder": {
|
| 3 |
+
"0": {
|
| 4 |
+
"content": "<s>",
|
| 5 |
+
"lstrip": false,
|
| 6 |
+
"normalized": false,
|
| 7 |
+
"rstrip": false,
|
| 8 |
+
"single_word": false,
|
| 9 |
+
"special": true
|
| 10 |
+
},
|
| 11 |
+
"1": {
|
| 12 |
+
"content": "<pad>",
|
| 13 |
+
"lstrip": false,
|
| 14 |
+
"normalized": false,
|
| 15 |
+
"rstrip": false,
|
| 16 |
+
"single_word": false,
|
| 17 |
+
"special": true
|
| 18 |
+
},
|
| 19 |
+
"2": {
|
| 20 |
+
"content": "</s>",
|
| 21 |
+
"lstrip": false,
|
| 22 |
+
"normalized": false,
|
| 23 |
+
"rstrip": false,
|
| 24 |
+
"single_word": false,
|
| 25 |
+
"special": true
|
| 26 |
+
},
|
| 27 |
+
"3": {
|
| 28 |
+
"content": "<unk>",
|
| 29 |
+
"lstrip": false,
|
| 30 |
+
"normalized": false,
|
| 31 |
+
"rstrip": false,
|
| 32 |
+
"single_word": false,
|
| 33 |
+
"special": true
|
| 34 |
+
},
|
| 35 |
+
"20480": {
|
| 36 |
+
"content": "<mask>",
|
| 37 |
+
"lstrip": false,
|
| 38 |
+
"normalized": false,
|
| 39 |
+
"rstrip": false,
|
| 40 |
+
"single_word": false,
|
| 41 |
+
"special": true
|
| 42 |
+
}
|
| 43 |
+
},
|
| 44 |
+
"bos_token": "<s>",
|
| 45 |
+
"clean_up_tokenization_spaces": false,
|
| 46 |
+
"cls_token": "<s>",
|
| 47 |
+
"eos_token": "</s>",
|
| 48 |
+
"extra_special_tokens": {},
|
| 49 |
+
"mask_token": "<mask>",
|
| 50 |
+
"model_max_length": 2048,
|
| 51 |
+
"pad_token": "<pad>",
|
| 52 |
+
"sep_token": "</s>",
|
| 53 |
+
"tokenizer_class": "XLMRobertaTokenizerFast",
|
| 54 |
+
"unk_token": "<unk>"
|
| 55 |
+
}
|
ood-vifactcheck-bamibert-topic-domain-04-s43-20260827T194549Z-3858bb98/trainer/checkpoint-1692/trainer_state.json
ADDED
|
@@ -0,0 +1,611 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 1692,
|
| 3 |
+
"best_metric": 0.8403151851486105,
|
| 4 |
+
"best_model_checkpoint": "/root/VNEncoder/results/checkpoints/ood-vifactcheck-bamibert-topic-domain-04-s43-20260827T194549Z-3858bb98/trainer/checkpoint-1692",
|
| 5 |
+
"epoch": 9.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 1692,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.13351134846461948,
|
| 14 |
+
"grad_norm": 12.041625022888184,
|
| 15 |
+
"learning_rate": 1.276595744680851e-06,
|
| 16 |
+
"loss": 4.4189,
|
| 17 |
+
"step": 25
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.26702269692923897,
|
| 21 |
+
"grad_norm": 23.579696655273438,
|
| 22 |
+
"learning_rate": 2.6063829787234047e-06,
|
| 23 |
+
"loss": 4.4199,
|
| 24 |
+
"step": 50
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.40053404539385845,
|
| 28 |
+
"grad_norm": 13.03511905670166,
|
| 29 |
+
"learning_rate": 3.936170212765958e-06,
|
| 30 |
+
"loss": 4.4164,
|
| 31 |
+
"step": 75
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.5340453938584779,
|
| 35 |
+
"grad_norm": 22.11226463317871,
|
| 36 |
+
"learning_rate": 5.265957446808511e-06,
|
| 37 |
+
"loss": 4.4016,
|
| 38 |
+
"step": 100
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.6675567423230975,
|
| 42 |
+
"grad_norm": 10.333353996276855,
|
| 43 |
+
"learning_rate": 6.595744680851064e-06,
|
| 44 |
+
"loss": 4.3491,
|
| 45 |
+
"step": 125
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.8010680907877169,
|
| 49 |
+
"grad_norm": 17.384540557861328,
|
| 50 |
+
"learning_rate": 7.925531914893617e-06,
|
| 51 |
+
"loss": 4.3027,
|
| 52 |
+
"step": 150
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"epoch": 0.9345794392523364,
|
| 56 |
+
"grad_norm": 17.556516647338867,
|
| 57 |
+
"learning_rate": 9.255319148936171e-06,
|
| 58 |
+
"loss": 4.2513,
|
| 59 |
+
"step": 175
|
| 60 |
+
},
|
| 61 |
+
{
|
| 62 |
+
"epoch": 1.0,
|
| 63 |
+
"eval_accuracy": 0.44168734491315137,
|
| 64 |
+
"eval_loss": 1.0277096033096313,
|
| 65 |
+
"eval_macro_f1": 0.38045091086483546,
|
| 66 |
+
"eval_runtime": 0.8805,
|
| 67 |
+
"eval_samples_per_second": 915.371,
|
| 68 |
+
"eval_steps_per_second": 57.92,
|
| 69 |
+
"eval_weighted_f1": 0.3833695067395448,
|
| 70 |
+
"step": 188
|
| 71 |
+
},
|
| 72 |
+
{
|
| 73 |
+
"epoch": 1.0640854472630175,
|
| 74 |
+
"grad_norm": 22.30445671081543,
|
| 75 |
+
"learning_rate": 1.0585106382978725e-05,
|
| 76 |
+
"loss": 3.9028,
|
| 77 |
+
"step": 200
|
| 78 |
+
},
|
| 79 |
+
{
|
| 80 |
+
"epoch": 1.1975967957276368,
|
| 81 |
+
"grad_norm": 37.01235580444336,
|
| 82 |
+
"learning_rate": 1.1914893617021277e-05,
|
| 83 |
+
"loss": 3.5939,
|
| 84 |
+
"step": 225
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 1.3311081441922563,
|
| 88 |
+
"grad_norm": 76.64301300048828,
|
| 89 |
+
"learning_rate": 1.3244680851063832e-05,
|
| 90 |
+
"loss": 3.4127,
|
| 91 |
+
"step": 250
|
| 92 |
+
},
|
| 93 |
+
{
|
| 94 |
+
"epoch": 1.4646194926568759,
|
| 95 |
+
"grad_norm": 137.1642303466797,
|
| 96 |
+
"learning_rate": 1.4574468085106384e-05,
|
| 97 |
+
"loss": 3.0726,
|
| 98 |
+
"step": 275
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 1.5981308411214954,
|
| 102 |
+
"grad_norm": 38.95008087158203,
|
| 103 |
+
"learning_rate": 1.590425531914894e-05,
|
| 104 |
+
"loss": 2.8287,
|
| 105 |
+
"step": 300
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 1.7316421895861147,
|
| 109 |
+
"grad_norm": 43.48196029663086,
|
| 110 |
+
"learning_rate": 1.723404255319149e-05,
|
| 111 |
+
"loss": 2.5254,
|
| 112 |
+
"step": 325
|
| 113 |
+
},
|
| 114 |
+
{
|
| 115 |
+
"epoch": 1.8651535380507345,
|
| 116 |
+
"grad_norm": 41.767425537109375,
|
| 117 |
+
"learning_rate": 1.8563829787234043e-05,
|
| 118 |
+
"loss": 2.5537,
|
| 119 |
+
"step": 350
|
| 120 |
+
},
|
| 121 |
+
{
|
| 122 |
+
"epoch": 1.9986648865153538,
|
| 123 |
+
"grad_norm": 51.959171295166016,
|
| 124 |
+
"learning_rate": 1.9893617021276595e-05,
|
| 125 |
+
"loss": 2.4611,
|
| 126 |
+
"step": 375
|
| 127 |
+
},
|
| 128 |
+
{
|
| 129 |
+
"epoch": 2.0,
|
| 130 |
+
"eval_accuracy": 0.7779156327543424,
|
| 131 |
+
"eval_loss": 0.5392054915428162,
|
| 132 |
+
"eval_macro_f1": 0.7731186959567543,
|
| 133 |
+
"eval_runtime": 0.8804,
|
| 134 |
+
"eval_samples_per_second": 915.476,
|
| 135 |
+
"eval_steps_per_second": 57.927,
|
| 136 |
+
"eval_weighted_f1": 0.7731682961222055,
|
| 137 |
+
"step": 376
|
| 138 |
+
},
|
| 139 |
+
{
|
| 140 |
+
"epoch": 2.128170894526035,
|
| 141 |
+
"grad_norm": 33.376922607421875,
|
| 142 |
+
"learning_rate": 1.986406619385343e-05,
|
| 143 |
+
"loss": 1.8897,
|
| 144 |
+
"step": 400
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 2.2616822429906542,
|
| 148 |
+
"grad_norm": 31.25908088684082,
|
| 149 |
+
"learning_rate": 1.971631205673759e-05,
|
| 150 |
+
"loss": 2.0112,
|
| 151 |
+
"step": 425
|
| 152 |
+
},
|
| 153 |
+
{
|
| 154 |
+
"epoch": 2.3951935914552736,
|
| 155 |
+
"grad_norm": 64.47969055175781,
|
| 156 |
+
"learning_rate": 1.956855791962175e-05,
|
| 157 |
+
"loss": 1.9916,
|
| 158 |
+
"step": 450
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 2.5287049399198933,
|
| 162 |
+
"grad_norm": 30.14402198791504,
|
| 163 |
+
"learning_rate": 1.9420803782505913e-05,
|
| 164 |
+
"loss": 1.801,
|
| 165 |
+
"step": 475
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 2.6622162883845126,
|
| 169 |
+
"grad_norm": 38.66526794433594,
|
| 170 |
+
"learning_rate": 1.927304964539007e-05,
|
| 171 |
+
"loss": 1.8946,
|
| 172 |
+
"step": 500
|
| 173 |
+
},
|
| 174 |
+
{
|
| 175 |
+
"epoch": 2.795727636849132,
|
| 176 |
+
"grad_norm": 99.96722412109375,
|
| 177 |
+
"learning_rate": 1.9125295508274234e-05,
|
| 178 |
+
"loss": 1.8492,
|
| 179 |
+
"step": 525
|
| 180 |
+
},
|
| 181 |
+
{
|
| 182 |
+
"epoch": 2.9292389853137517,
|
| 183 |
+
"grad_norm": 51.215087890625,
|
| 184 |
+
"learning_rate": 1.8977541371158394e-05,
|
| 185 |
+
"loss": 1.7098,
|
| 186 |
+
"step": 550
|
| 187 |
+
},
|
| 188 |
+
{
|
| 189 |
+
"epoch": 3.0,
|
| 190 |
+
"eval_accuracy": 0.8039702233250621,
|
| 191 |
+
"eval_loss": 0.5110635757446289,
|
| 192 |
+
"eval_macro_f1": 0.8061595909758944,
|
| 193 |
+
"eval_runtime": 0.8886,
|
| 194 |
+
"eval_samples_per_second": 907.074,
|
| 195 |
+
"eval_steps_per_second": 57.396,
|
| 196 |
+
"eval_weighted_f1": 0.8061343319914546,
|
| 197 |
+
"step": 564
|
| 198 |
+
},
|
| 199 |
+
{
|
| 200 |
+
"epoch": 3.0587449933244324,
|
| 201 |
+
"grad_norm": 48.343284606933594,
|
| 202 |
+
"learning_rate": 1.8829787234042557e-05,
|
| 203 |
+
"loss": 1.5609,
|
| 204 |
+
"step": 575
|
| 205 |
+
},
|
| 206 |
+
{
|
| 207 |
+
"epoch": 3.192256341789052,
|
| 208 |
+
"grad_norm": 52.88315200805664,
|
| 209 |
+
"learning_rate": 1.8682033096926714e-05,
|
| 210 |
+
"loss": 1.0574,
|
| 211 |
+
"step": 600
|
| 212 |
+
},
|
| 213 |
+
{
|
| 214 |
+
"epoch": 3.3257676902536715,
|
| 215 |
+
"grad_norm": 68.4620132446289,
|
| 216 |
+
"learning_rate": 1.8534278959810878e-05,
|
| 217 |
+
"loss": 0.9967,
|
| 218 |
+
"step": 625
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 3.4592790387182912,
|
| 222 |
+
"grad_norm": 46.33494186401367,
|
| 223 |
+
"learning_rate": 1.8386524822695038e-05,
|
| 224 |
+
"loss": 1.3332,
|
| 225 |
+
"step": 650
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"epoch": 3.5927903871829105,
|
| 229 |
+
"grad_norm": 46.911983489990234,
|
| 230 |
+
"learning_rate": 1.8238770685579198e-05,
|
| 231 |
+
"loss": 1.2784,
|
| 232 |
+
"step": 675
|
| 233 |
+
},
|
| 234 |
+
{
|
| 235 |
+
"epoch": 3.72630173564753,
|
| 236 |
+
"grad_norm": 50.444419860839844,
|
| 237 |
+
"learning_rate": 1.8091016548463358e-05,
|
| 238 |
+
"loss": 1.3525,
|
| 239 |
+
"step": 700
|
| 240 |
+
},
|
| 241 |
+
{
|
| 242 |
+
"epoch": 3.8598130841121496,
|
| 243 |
+
"grad_norm": 32.238704681396484,
|
| 244 |
+
"learning_rate": 1.7943262411347518e-05,
|
| 245 |
+
"loss": 1.3185,
|
| 246 |
+
"step": 725
|
| 247 |
+
},
|
| 248 |
+
{
|
| 249 |
+
"epoch": 3.993324432576769,
|
| 250 |
+
"grad_norm": 24.558015823364258,
|
| 251 |
+
"learning_rate": 1.779550827423168e-05,
|
| 252 |
+
"loss": 1.4304,
|
| 253 |
+
"step": 750
|
| 254 |
+
},
|
| 255 |
+
{
|
| 256 |
+
"epoch": 4.0,
|
| 257 |
+
"eval_accuracy": 0.826302729528536,
|
| 258 |
+
"eval_loss": 0.515184223651886,
|
| 259 |
+
"eval_macro_f1": 0.8249221583751791,
|
| 260 |
+
"eval_runtime": 0.8763,
|
| 261 |
+
"eval_samples_per_second": 919.813,
|
| 262 |
+
"eval_steps_per_second": 58.202,
|
| 263 |
+
"eval_weighted_f1": 0.8247302337608486,
|
| 264 |
+
"step": 752
|
| 265 |
+
},
|
| 266 |
+
{
|
| 267 |
+
"epoch": 4.12283044058745,
|
| 268 |
+
"grad_norm": 10.826188087463379,
|
| 269 |
+
"learning_rate": 1.7647754137115842e-05,
|
| 270 |
+
"loss": 0.6535,
|
| 271 |
+
"step": 775
|
| 272 |
+
},
|
| 273 |
+
{
|
| 274 |
+
"epoch": 4.25634178905207,
|
| 275 |
+
"grad_norm": 51.57345962524414,
|
| 276 |
+
"learning_rate": 1.7500000000000002e-05,
|
| 277 |
+
"loss": 0.8793,
|
| 278 |
+
"step": 800
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 4.389853137516689,
|
| 282 |
+
"grad_norm": 66.71621704101562,
|
| 283 |
+
"learning_rate": 1.7352245862884162e-05,
|
| 284 |
+
"loss": 0.9412,
|
| 285 |
+
"step": 825
|
| 286 |
+
},
|
| 287 |
+
{
|
| 288 |
+
"epoch": 4.5233644859813085,
|
| 289 |
+
"grad_norm": 52.68075942993164,
|
| 290 |
+
"learning_rate": 1.7204491725768322e-05,
|
| 291 |
+
"loss": 0.9856,
|
| 292 |
+
"step": 850
|
| 293 |
+
},
|
| 294 |
+
{
|
| 295 |
+
"epoch": 4.656875834445928,
|
| 296 |
+
"grad_norm": 26.891000747680664,
|
| 297 |
+
"learning_rate": 1.7056737588652482e-05,
|
| 298 |
+
"loss": 0.7824,
|
| 299 |
+
"step": 875
|
| 300 |
+
},
|
| 301 |
+
{
|
| 302 |
+
"epoch": 4.790387182910547,
|
| 303 |
+
"grad_norm": 49.95881271362305,
|
| 304 |
+
"learning_rate": 1.6908983451536646e-05,
|
| 305 |
+
"loss": 0.7956,
|
| 306 |
+
"step": 900
|
| 307 |
+
},
|
| 308 |
+
{
|
| 309 |
+
"epoch": 4.923898531375167,
|
| 310 |
+
"grad_norm": 60.759315490722656,
|
| 311 |
+
"learning_rate": 1.6761229314420806e-05,
|
| 312 |
+
"loss": 0.8574,
|
| 313 |
+
"step": 925
|
| 314 |
+
},
|
| 315 |
+
{
|
| 316 |
+
"epoch": 5.0,
|
| 317 |
+
"eval_accuracy": 0.8312655086848635,
|
| 318 |
+
"eval_loss": 0.5548403859138489,
|
| 319 |
+
"eval_macro_f1": 0.8300582328098267,
|
| 320 |
+
"eval_runtime": 0.8932,
|
| 321 |
+
"eval_samples_per_second": 902.398,
|
| 322 |
+
"eval_steps_per_second": 57.1,
|
| 323 |
+
"eval_weighted_f1": 0.8299638559696316,
|
| 324 |
+
"step": 940
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 5.0534045393858475,
|
| 328 |
+
"grad_norm": 34.98723220825195,
|
| 329 |
+
"learning_rate": 1.6613475177304966e-05,
|
| 330 |
+
"loss": 0.5591,
|
| 331 |
+
"step": 950
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 5.186915887850467,
|
| 335 |
+
"grad_norm": 47.1026496887207,
|
| 336 |
+
"learning_rate": 1.6465721040189126e-05,
|
| 337 |
+
"loss": 0.74,
|
| 338 |
+
"step": 975
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 5.320427236315087,
|
| 342 |
+
"grad_norm": 65.67156219482422,
|
| 343 |
+
"learning_rate": 1.6317966903073286e-05,
|
| 344 |
+
"loss": 0.5691,
|
| 345 |
+
"step": 1000
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 5.453938584779706,
|
| 349 |
+
"grad_norm": 114.16455078125,
|
| 350 |
+
"learning_rate": 1.6170212765957446e-05,
|
| 351 |
+
"loss": 0.559,
|
| 352 |
+
"step": 1025
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 5.587449933244326,
|
| 356 |
+
"grad_norm": 56.513545989990234,
|
| 357 |
+
"learning_rate": 1.602245862884161e-05,
|
| 358 |
+
"loss": 0.4883,
|
| 359 |
+
"step": 1050
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 5.720961281708945,
|
| 363 |
+
"grad_norm": 26.13739776611328,
|
| 364 |
+
"learning_rate": 1.587470449172577e-05,
|
| 365 |
+
"loss": 0.5287,
|
| 366 |
+
"step": 1075
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 5.854472630173564,
|
| 370 |
+
"grad_norm": 52.19254684448242,
|
| 371 |
+
"learning_rate": 1.572695035460993e-05,
|
| 372 |
+
"loss": 0.6611,
|
| 373 |
+
"step": 1100
|
| 374 |
+
},
|
| 375 |
+
{
|
| 376 |
+
"epoch": 5.9879839786381845,
|
| 377 |
+
"grad_norm": 149.34495544433594,
|
| 378 |
+
"learning_rate": 1.557919621749409e-05,
|
| 379 |
+
"loss": 0.5178,
|
| 380 |
+
"step": 1125
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 6.0,
|
| 384 |
+
"eval_accuracy": 0.8163771712158809,
|
| 385 |
+
"eval_loss": 0.6910102963447571,
|
| 386 |
+
"eval_macro_f1": 0.8147348816728024,
|
| 387 |
+
"eval_runtime": 0.933,
|
| 388 |
+
"eval_samples_per_second": 863.85,
|
| 389 |
+
"eval_steps_per_second": 54.66,
|
| 390 |
+
"eval_weighted_f1": 0.8147177066425263,
|
| 391 |
+
"step": 1128
|
| 392 |
+
},
|
| 393 |
+
{
|
| 394 |
+
"epoch": 6.117489986648865,
|
| 395 |
+
"grad_norm": 11.702919960021973,
|
| 396 |
+
"learning_rate": 1.543144208037825e-05,
|
| 397 |
+
"loss": 0.4257,
|
| 398 |
+
"step": 1150
|
| 399 |
+
},
|
| 400 |
+
{
|
| 401 |
+
"epoch": 6.251001335113485,
|
| 402 |
+
"grad_norm": 50.10649490356445,
|
| 403 |
+
"learning_rate": 1.5283687943262414e-05,
|
| 404 |
+
"loss": 0.5115,
|
| 405 |
+
"step": 1175
|
| 406 |
+
},
|
| 407 |
+
{
|
| 408 |
+
"epoch": 6.384512683578104,
|
| 409 |
+
"grad_norm": 1.3259004354476929,
|
| 410 |
+
"learning_rate": 1.5135933806146574e-05,
|
| 411 |
+
"loss": 0.3388,
|
| 412 |
+
"step": 1200
|
| 413 |
+
},
|
| 414 |
+
{
|
| 415 |
+
"epoch": 6.518024032042724,
|
| 416 |
+
"grad_norm": 25.595294952392578,
|
| 417 |
+
"learning_rate": 1.4988179669030734e-05,
|
| 418 |
+
"loss": 0.5207,
|
| 419 |
+
"step": 1225
|
| 420 |
+
},
|
| 421 |
+
{
|
| 422 |
+
"epoch": 6.651535380507343,
|
| 423 |
+
"grad_norm": 35.91559982299805,
|
| 424 |
+
"learning_rate": 1.4840425531914894e-05,
|
| 425 |
+
"loss": 0.4556,
|
| 426 |
+
"step": 1250
|
| 427 |
+
},
|
| 428 |
+
{
|
| 429 |
+
"epoch": 6.785046728971962,
|
| 430 |
+
"grad_norm": 120.19802856445312,
|
| 431 |
+
"learning_rate": 1.4692671394799056e-05,
|
| 432 |
+
"loss": 0.3891,
|
| 433 |
+
"step": 1275
|
| 434 |
+
},
|
| 435 |
+
{
|
| 436 |
+
"epoch": 6.9185580774365825,
|
| 437 |
+
"grad_norm": 42.937896728515625,
|
| 438 |
+
"learning_rate": 1.4544917257683215e-05,
|
| 439 |
+
"loss": 0.4652,
|
| 440 |
+
"step": 1300
|
| 441 |
+
},
|
| 442 |
+
{
|
| 443 |
+
"epoch": 7.0,
|
| 444 |
+
"eval_accuracy": 0.8325062034739454,
|
| 445 |
+
"eval_loss": 0.7512131333351135,
|
| 446 |
+
"eval_macro_f1": 0.8317569687847032,
|
| 447 |
+
"eval_runtime": 0.8853,
|
| 448 |
+
"eval_samples_per_second": 910.411,
|
| 449 |
+
"eval_steps_per_second": 57.607,
|
| 450 |
+
"eval_weighted_f1": 0.8317282250216943,
|
| 451 |
+
"step": 1316
|
| 452 |
+
},
|
| 453 |
+
{
|
| 454 |
+
"epoch": 7.048064085447263,
|
| 455 |
+
"grad_norm": 0.8096206188201904,
|
| 456 |
+
"learning_rate": 1.4397163120567377e-05,
|
| 457 |
+
"loss": 0.5165,
|
| 458 |
+
"step": 1325
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 7.181575433911883,
|
| 462 |
+
"grad_norm": 9.723548889160156,
|
| 463 |
+
"learning_rate": 1.4249408983451538e-05,
|
| 464 |
+
"loss": 0.3382,
|
| 465 |
+
"step": 1350
|
| 466 |
+
},
|
| 467 |
+
{
|
| 468 |
+
"epoch": 7.315086782376502,
|
| 469 |
+
"grad_norm": 97.81317138671875,
|
| 470 |
+
"learning_rate": 1.41016548463357e-05,
|
| 471 |
+
"loss": 0.2261,
|
| 472 |
+
"step": 1375
|
| 473 |
+
},
|
| 474 |
+
{
|
| 475 |
+
"epoch": 7.4485981308411215,
|
| 476 |
+
"grad_norm": 0.24998800456523895,
|
| 477 |
+
"learning_rate": 1.3953900709219859e-05,
|
| 478 |
+
"loss": 0.2369,
|
| 479 |
+
"step": 1400
|
| 480 |
+
},
|
| 481 |
+
{
|
| 482 |
+
"epoch": 7.582109479305741,
|
| 483 |
+
"grad_norm": 6.461670398712158,
|
| 484 |
+
"learning_rate": 1.380614657210402e-05,
|
| 485 |
+
"loss": 0.2873,
|
| 486 |
+
"step": 1425
|
| 487 |
+
},
|
| 488 |
+
{
|
| 489 |
+
"epoch": 7.71562082777036,
|
| 490 |
+
"grad_norm": 0.23604150116443634,
|
| 491 |
+
"learning_rate": 1.365839243498818e-05,
|
| 492 |
+
"loss": 0.398,
|
| 493 |
+
"step": 1450
|
| 494 |
+
},
|
| 495 |
+
{
|
| 496 |
+
"epoch": 7.8491321762349795,
|
| 497 |
+
"grad_norm": 52.03115463256836,
|
| 498 |
+
"learning_rate": 1.351063829787234e-05,
|
| 499 |
+
"loss": 0.5035,
|
| 500 |
+
"step": 1475
|
| 501 |
+
},
|
| 502 |
+
{
|
| 503 |
+
"epoch": 7.9826435246996,
|
| 504 |
+
"grad_norm": 69.65509796142578,
|
| 505 |
+
"learning_rate": 1.3362884160756503e-05,
|
| 506 |
+
"loss": 0.3796,
|
| 507 |
+
"step": 1500
|
| 508 |
+
},
|
| 509 |
+
{
|
| 510 |
+
"epoch": 8.0,
|
| 511 |
+
"eval_accuracy": 0.825062034739454,
|
| 512 |
+
"eval_loss": 0.9201409220695496,
|
| 513 |
+
"eval_macro_f1": 0.8237998988550653,
|
| 514 |
+
"eval_runtime": 0.8769,
|
| 515 |
+
"eval_samples_per_second": 919.161,
|
| 516 |
+
"eval_steps_per_second": 58.16,
|
| 517 |
+
"eval_weighted_f1": 0.8236155440779873,
|
| 518 |
+
"step": 1504
|
| 519 |
+
},
|
| 520 |
+
{
|
| 521 |
+
"epoch": 8.11214953271028,
|
| 522 |
+
"grad_norm": 66.33985137939453,
|
| 523 |
+
"learning_rate": 1.3215130023640663e-05,
|
| 524 |
+
"loss": 0.2118,
|
| 525 |
+
"step": 1525
|
| 526 |
+
},
|
| 527 |
+
{
|
| 528 |
+
"epoch": 8.2456608811749,
|
| 529 |
+
"grad_norm": 1.323639988899231,
|
| 530 |
+
"learning_rate": 1.3067375886524825e-05,
|
| 531 |
+
"loss": 0.3528,
|
| 532 |
+
"step": 1550
|
| 533 |
+
},
|
| 534 |
+
{
|
| 535 |
+
"epoch": 8.379172229639519,
|
| 536 |
+
"grad_norm": 0.2407652735710144,
|
| 537 |
+
"learning_rate": 1.2919621749408983e-05,
|
| 538 |
+
"loss": 0.1385,
|
| 539 |
+
"step": 1575
|
| 540 |
+
},
|
| 541 |
+
{
|
| 542 |
+
"epoch": 8.51268357810414,
|
| 543 |
+
"grad_norm": 246.52987670898438,
|
| 544 |
+
"learning_rate": 1.2771867612293145e-05,
|
| 545 |
+
"loss": 0.1627,
|
| 546 |
+
"step": 1600
|
| 547 |
+
},
|
| 548 |
+
{
|
| 549 |
+
"epoch": 8.646194926568759,
|
| 550 |
+
"grad_norm": 31.24919319152832,
|
| 551 |
+
"learning_rate": 1.2624113475177307e-05,
|
| 552 |
+
"loss": 0.1876,
|
| 553 |
+
"step": 1625
|
| 554 |
+
},
|
| 555 |
+
{
|
| 556 |
+
"epoch": 8.779706275033378,
|
| 557 |
+
"grad_norm": 60.09649658203125,
|
| 558 |
+
"learning_rate": 1.2476359338061469e-05,
|
| 559 |
+
"loss": 0.3086,
|
| 560 |
+
"step": 1650
|
| 561 |
+
},
|
| 562 |
+
{
|
| 563 |
+
"epoch": 8.913217623497998,
|
| 564 |
+
"grad_norm": 0.33130812644958496,
|
| 565 |
+
"learning_rate": 1.2328605200945627e-05,
|
| 566 |
+
"loss": 0.2882,
|
| 567 |
+
"step": 1675
|
| 568 |
+
},
|
| 569 |
+
{
|
| 570 |
+
"epoch": 9.0,
|
| 571 |
+
"eval_accuracy": 0.8399503722084367,
|
| 572 |
+
"eval_loss": 0.980648398399353,
|
| 573 |
+
"eval_macro_f1": 0.8403151851486105,
|
| 574 |
+
"eval_runtime": 0.9351,
|
| 575 |
+
"eval_samples_per_second": 861.939,
|
| 576 |
+
"eval_steps_per_second": 54.54,
|
| 577 |
+
"eval_weighted_f1": 0.8403824279884705,
|
| 578 |
+
"step": 1692
|
| 579 |
+
}
|
| 580 |
+
],
|
| 581 |
+
"logging_steps": 25,
|
| 582 |
+
"max_steps": 3760,
|
| 583 |
+
"num_input_tokens_seen": 0,
|
| 584 |
+
"num_train_epochs": 20,
|
| 585 |
+
"save_steps": 500,
|
| 586 |
+
"stateful_callbacks": {
|
| 587 |
+
"EarlyStoppingCallback": {
|
| 588 |
+
"args": {
|
| 589 |
+
"early_stopping_patience": 3,
|
| 590 |
+
"early_stopping_threshold": 0.0
|
| 591 |
+
},
|
| 592 |
+
"attributes": {
|
| 593 |
+
"early_stopping_patience_counter": 0
|
| 594 |
+
}
|
| 595 |
+
},
|
| 596 |
+
"TrainerControl": {
|
| 597 |
+
"args": {
|
| 598 |
+
"should_epoch_stop": false,
|
| 599 |
+
"should_evaluate": false,
|
| 600 |
+
"should_log": false,
|
| 601 |
+
"should_save": true,
|
| 602 |
+
"should_training_stop": false
|
| 603 |
+
},
|
| 604 |
+
"attributes": {}
|
| 605 |
+
}
|
| 606 |
+
},
|
| 607 |
+
"total_flos": 0.0,
|
| 608 |
+
"train_batch_size": 8,
|
| 609 |
+
"trial_name": null,
|
| 610 |
+
"trial_params": null
|
| 611 |
+
}
|
ood-vifactcheck-bamibert-topic-domain-04-s43-20260827T194549Z-3858bb98/trainer/checkpoint-2256/special_tokens_map.json
ADDED
|
@@ -0,0 +1,51 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": {
|
| 3 |
+
"content": "<s>",
|
| 4 |
+
"lstrip": false,
|
| 5 |
+
"normalized": false,
|
| 6 |
+
"rstrip": false,
|
| 7 |
+
"single_word": false
|
| 8 |
+
},
|
| 9 |
+
"cls_token": {
|
| 10 |
+
"content": "<s>",
|
| 11 |
+
"lstrip": false,
|
| 12 |
+
"normalized": false,
|
| 13 |
+
"rstrip": false,
|
| 14 |
+
"single_word": false
|
| 15 |
+
},
|
| 16 |
+
"eos_token": {
|
| 17 |
+
"content": "</s>",
|
| 18 |
+
"lstrip": false,
|
| 19 |
+
"normalized": false,
|
| 20 |
+
"rstrip": false,
|
| 21 |
+
"single_word": false
|
| 22 |
+
},
|
| 23 |
+
"mask_token": {
|
| 24 |
+
"content": "<mask>",
|
| 25 |
+
"lstrip": false,
|
| 26 |
+
"normalized": false,
|
| 27 |
+
"rstrip": false,
|
| 28 |
+
"single_word": false
|
| 29 |
+
},
|
| 30 |
+
"pad_token": {
|
| 31 |
+
"content": "<pad>",
|
| 32 |
+
"lstrip": false,
|
| 33 |
+
"normalized": false,
|
| 34 |
+
"rstrip": false,
|
| 35 |
+
"single_word": false
|
| 36 |
+
},
|
| 37 |
+
"sep_token": {
|
| 38 |
+
"content": "</s>",
|
| 39 |
+
"lstrip": false,
|
| 40 |
+
"normalized": false,
|
| 41 |
+
"rstrip": false,
|
| 42 |
+
"single_word": false
|
| 43 |
+
},
|
| 44 |
+
"unk_token": {
|
| 45 |
+
"content": "<unk>",
|
| 46 |
+
"lstrip": false,
|
| 47 |
+
"normalized": false,
|
| 48 |
+
"rstrip": false,
|
| 49 |
+
"single_word": false
|
| 50 |
+
}
|
| 51 |
+
}
|
ood-vifactcheck-bamibert-topic-domain-04-s43-20260827T194549Z-3858bb98/trainer/checkpoint-2256/tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-bamibert-topic-domain-04-s43-20260827T194549Z-3858bb98/trainer/checkpoint-2256/tokenizer_config.json
ADDED
|
@@ -0,0 +1,55 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"added_tokens_decoder": {
|
| 3 |
+
"0": {
|
| 4 |
+
"content": "<s>",
|
| 5 |
+
"lstrip": false,
|
| 6 |
+
"normalized": false,
|
| 7 |
+
"rstrip": false,
|
| 8 |
+
"single_word": false,
|
| 9 |
+
"special": true
|
| 10 |
+
},
|
| 11 |
+
"1": {
|
| 12 |
+
"content": "<pad>",
|
| 13 |
+
"lstrip": false,
|
| 14 |
+
"normalized": false,
|
| 15 |
+
"rstrip": false,
|
| 16 |
+
"single_word": false,
|
| 17 |
+
"special": true
|
| 18 |
+
},
|
| 19 |
+
"2": {
|
| 20 |
+
"content": "</s>",
|
| 21 |
+
"lstrip": false,
|
| 22 |
+
"normalized": false,
|
| 23 |
+
"rstrip": false,
|
| 24 |
+
"single_word": false,
|
| 25 |
+
"special": true
|
| 26 |
+
},
|
| 27 |
+
"3": {
|
| 28 |
+
"content": "<unk>",
|
| 29 |
+
"lstrip": false,
|
| 30 |
+
"normalized": false,
|
| 31 |
+
"rstrip": false,
|
| 32 |
+
"single_word": false,
|
| 33 |
+
"special": true
|
| 34 |
+
},
|
| 35 |
+
"20480": {
|
| 36 |
+
"content": "<mask>",
|
| 37 |
+
"lstrip": false,
|
| 38 |
+
"normalized": false,
|
| 39 |
+
"rstrip": false,
|
| 40 |
+
"single_word": false,
|
| 41 |
+
"special": true
|
| 42 |
+
}
|
| 43 |
+
},
|
| 44 |
+
"bos_token": "<s>",
|
| 45 |
+
"clean_up_tokenization_spaces": false,
|
| 46 |
+
"cls_token": "<s>",
|
| 47 |
+
"eos_token": "</s>",
|
| 48 |
+
"extra_special_tokens": {},
|
| 49 |
+
"mask_token": "<mask>",
|
| 50 |
+
"model_max_length": 2048,
|
| 51 |
+
"pad_token": "<pad>",
|
| 52 |
+
"sep_token": "</s>",
|
| 53 |
+
"tokenizer_class": "XLMRobertaTokenizerFast",
|
| 54 |
+
"unk_token": "<unk>"
|
| 55 |
+
}
|
ood-vifactcheck-bamibert-topic-domain-04-s43-20260827T194549Z-3858bb98/trainer/checkpoint-2256/trainer_state.json
ADDED
|
@@ -0,0 +1,805 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 1692,
|
| 3 |
+
"best_metric": 0.8403151851486105,
|
| 4 |
+
"best_model_checkpoint": "/root/VNEncoder/results/checkpoints/ood-vifactcheck-bamibert-topic-domain-04-s43-20260827T194549Z-3858bb98/trainer/checkpoint-1692",
|
| 5 |
+
"epoch": 12.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 2256,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.13351134846461948,
|
| 14 |
+
"grad_norm": 12.041625022888184,
|
| 15 |
+
"learning_rate": 1.276595744680851e-06,
|
| 16 |
+
"loss": 4.4189,
|
| 17 |
+
"step": 25
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.26702269692923897,
|
| 21 |
+
"grad_norm": 23.579696655273438,
|
| 22 |
+
"learning_rate": 2.6063829787234047e-06,
|
| 23 |
+
"loss": 4.4199,
|
| 24 |
+
"step": 50
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.40053404539385845,
|
| 28 |
+
"grad_norm": 13.03511905670166,
|
| 29 |
+
"learning_rate": 3.936170212765958e-06,
|
| 30 |
+
"loss": 4.4164,
|
| 31 |
+
"step": 75
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.5340453938584779,
|
| 35 |
+
"grad_norm": 22.11226463317871,
|
| 36 |
+
"learning_rate": 5.265957446808511e-06,
|
| 37 |
+
"loss": 4.4016,
|
| 38 |
+
"step": 100
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.6675567423230975,
|
| 42 |
+
"grad_norm": 10.333353996276855,
|
| 43 |
+
"learning_rate": 6.595744680851064e-06,
|
| 44 |
+
"loss": 4.3491,
|
| 45 |
+
"step": 125
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.8010680907877169,
|
| 49 |
+
"grad_norm": 17.384540557861328,
|
| 50 |
+
"learning_rate": 7.925531914893617e-06,
|
| 51 |
+
"loss": 4.3027,
|
| 52 |
+
"step": 150
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"epoch": 0.9345794392523364,
|
| 56 |
+
"grad_norm": 17.556516647338867,
|
| 57 |
+
"learning_rate": 9.255319148936171e-06,
|
| 58 |
+
"loss": 4.2513,
|
| 59 |
+
"step": 175
|
| 60 |
+
},
|
| 61 |
+
{
|
| 62 |
+
"epoch": 1.0,
|
| 63 |
+
"eval_accuracy": 0.44168734491315137,
|
| 64 |
+
"eval_loss": 1.0277096033096313,
|
| 65 |
+
"eval_macro_f1": 0.38045091086483546,
|
| 66 |
+
"eval_runtime": 0.8805,
|
| 67 |
+
"eval_samples_per_second": 915.371,
|
| 68 |
+
"eval_steps_per_second": 57.92,
|
| 69 |
+
"eval_weighted_f1": 0.3833695067395448,
|
| 70 |
+
"step": 188
|
| 71 |
+
},
|
| 72 |
+
{
|
| 73 |
+
"epoch": 1.0640854472630175,
|
| 74 |
+
"grad_norm": 22.30445671081543,
|
| 75 |
+
"learning_rate": 1.0585106382978725e-05,
|
| 76 |
+
"loss": 3.9028,
|
| 77 |
+
"step": 200
|
| 78 |
+
},
|
| 79 |
+
{
|
| 80 |
+
"epoch": 1.1975967957276368,
|
| 81 |
+
"grad_norm": 37.01235580444336,
|
| 82 |
+
"learning_rate": 1.1914893617021277e-05,
|
| 83 |
+
"loss": 3.5939,
|
| 84 |
+
"step": 225
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 1.3311081441922563,
|
| 88 |
+
"grad_norm": 76.64301300048828,
|
| 89 |
+
"learning_rate": 1.3244680851063832e-05,
|
| 90 |
+
"loss": 3.4127,
|
| 91 |
+
"step": 250
|
| 92 |
+
},
|
| 93 |
+
{
|
| 94 |
+
"epoch": 1.4646194926568759,
|
| 95 |
+
"grad_norm": 137.1642303466797,
|
| 96 |
+
"learning_rate": 1.4574468085106384e-05,
|
| 97 |
+
"loss": 3.0726,
|
| 98 |
+
"step": 275
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 1.5981308411214954,
|
| 102 |
+
"grad_norm": 38.95008087158203,
|
| 103 |
+
"learning_rate": 1.590425531914894e-05,
|
| 104 |
+
"loss": 2.8287,
|
| 105 |
+
"step": 300
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 1.7316421895861147,
|
| 109 |
+
"grad_norm": 43.48196029663086,
|
| 110 |
+
"learning_rate": 1.723404255319149e-05,
|
| 111 |
+
"loss": 2.5254,
|
| 112 |
+
"step": 325
|
| 113 |
+
},
|
| 114 |
+
{
|
| 115 |
+
"epoch": 1.8651535380507345,
|
| 116 |
+
"grad_norm": 41.767425537109375,
|
| 117 |
+
"learning_rate": 1.8563829787234043e-05,
|
| 118 |
+
"loss": 2.5537,
|
| 119 |
+
"step": 350
|
| 120 |
+
},
|
| 121 |
+
{
|
| 122 |
+
"epoch": 1.9986648865153538,
|
| 123 |
+
"grad_norm": 51.959171295166016,
|
| 124 |
+
"learning_rate": 1.9893617021276595e-05,
|
| 125 |
+
"loss": 2.4611,
|
| 126 |
+
"step": 375
|
| 127 |
+
},
|
| 128 |
+
{
|
| 129 |
+
"epoch": 2.0,
|
| 130 |
+
"eval_accuracy": 0.7779156327543424,
|
| 131 |
+
"eval_loss": 0.5392054915428162,
|
| 132 |
+
"eval_macro_f1": 0.7731186959567543,
|
| 133 |
+
"eval_runtime": 0.8804,
|
| 134 |
+
"eval_samples_per_second": 915.476,
|
| 135 |
+
"eval_steps_per_second": 57.927,
|
| 136 |
+
"eval_weighted_f1": 0.7731682961222055,
|
| 137 |
+
"step": 376
|
| 138 |
+
},
|
| 139 |
+
{
|
| 140 |
+
"epoch": 2.128170894526035,
|
| 141 |
+
"grad_norm": 33.376922607421875,
|
| 142 |
+
"learning_rate": 1.986406619385343e-05,
|
| 143 |
+
"loss": 1.8897,
|
| 144 |
+
"step": 400
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 2.2616822429906542,
|
| 148 |
+
"grad_norm": 31.25908088684082,
|
| 149 |
+
"learning_rate": 1.971631205673759e-05,
|
| 150 |
+
"loss": 2.0112,
|
| 151 |
+
"step": 425
|
| 152 |
+
},
|
| 153 |
+
{
|
| 154 |
+
"epoch": 2.3951935914552736,
|
| 155 |
+
"grad_norm": 64.47969055175781,
|
| 156 |
+
"learning_rate": 1.956855791962175e-05,
|
| 157 |
+
"loss": 1.9916,
|
| 158 |
+
"step": 450
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 2.5287049399198933,
|
| 162 |
+
"grad_norm": 30.14402198791504,
|
| 163 |
+
"learning_rate": 1.9420803782505913e-05,
|
| 164 |
+
"loss": 1.801,
|
| 165 |
+
"step": 475
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 2.6622162883845126,
|
| 169 |
+
"grad_norm": 38.66526794433594,
|
| 170 |
+
"learning_rate": 1.927304964539007e-05,
|
| 171 |
+
"loss": 1.8946,
|
| 172 |
+
"step": 500
|
| 173 |
+
},
|
| 174 |
+
{
|
| 175 |
+
"epoch": 2.795727636849132,
|
| 176 |
+
"grad_norm": 99.96722412109375,
|
| 177 |
+
"learning_rate": 1.9125295508274234e-05,
|
| 178 |
+
"loss": 1.8492,
|
| 179 |
+
"step": 525
|
| 180 |
+
},
|
| 181 |
+
{
|
| 182 |
+
"epoch": 2.9292389853137517,
|
| 183 |
+
"grad_norm": 51.215087890625,
|
| 184 |
+
"learning_rate": 1.8977541371158394e-05,
|
| 185 |
+
"loss": 1.7098,
|
| 186 |
+
"step": 550
|
| 187 |
+
},
|
| 188 |
+
{
|
| 189 |
+
"epoch": 3.0,
|
| 190 |
+
"eval_accuracy": 0.8039702233250621,
|
| 191 |
+
"eval_loss": 0.5110635757446289,
|
| 192 |
+
"eval_macro_f1": 0.8061595909758944,
|
| 193 |
+
"eval_runtime": 0.8886,
|
| 194 |
+
"eval_samples_per_second": 907.074,
|
| 195 |
+
"eval_steps_per_second": 57.396,
|
| 196 |
+
"eval_weighted_f1": 0.8061343319914546,
|
| 197 |
+
"step": 564
|
| 198 |
+
},
|
| 199 |
+
{
|
| 200 |
+
"epoch": 3.0587449933244324,
|
| 201 |
+
"grad_norm": 48.343284606933594,
|
| 202 |
+
"learning_rate": 1.8829787234042557e-05,
|
| 203 |
+
"loss": 1.5609,
|
| 204 |
+
"step": 575
|
| 205 |
+
},
|
| 206 |
+
{
|
| 207 |
+
"epoch": 3.192256341789052,
|
| 208 |
+
"grad_norm": 52.88315200805664,
|
| 209 |
+
"learning_rate": 1.8682033096926714e-05,
|
| 210 |
+
"loss": 1.0574,
|
| 211 |
+
"step": 600
|
| 212 |
+
},
|
| 213 |
+
{
|
| 214 |
+
"epoch": 3.3257676902536715,
|
| 215 |
+
"grad_norm": 68.4620132446289,
|
| 216 |
+
"learning_rate": 1.8534278959810878e-05,
|
| 217 |
+
"loss": 0.9967,
|
| 218 |
+
"step": 625
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 3.4592790387182912,
|
| 222 |
+
"grad_norm": 46.33494186401367,
|
| 223 |
+
"learning_rate": 1.8386524822695038e-05,
|
| 224 |
+
"loss": 1.3332,
|
| 225 |
+
"step": 650
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"epoch": 3.5927903871829105,
|
| 229 |
+
"grad_norm": 46.911983489990234,
|
| 230 |
+
"learning_rate": 1.8238770685579198e-05,
|
| 231 |
+
"loss": 1.2784,
|
| 232 |
+
"step": 675
|
| 233 |
+
},
|
| 234 |
+
{
|
| 235 |
+
"epoch": 3.72630173564753,
|
| 236 |
+
"grad_norm": 50.444419860839844,
|
| 237 |
+
"learning_rate": 1.8091016548463358e-05,
|
| 238 |
+
"loss": 1.3525,
|
| 239 |
+
"step": 700
|
| 240 |
+
},
|
| 241 |
+
{
|
| 242 |
+
"epoch": 3.8598130841121496,
|
| 243 |
+
"grad_norm": 32.238704681396484,
|
| 244 |
+
"learning_rate": 1.7943262411347518e-05,
|
| 245 |
+
"loss": 1.3185,
|
| 246 |
+
"step": 725
|
| 247 |
+
},
|
| 248 |
+
{
|
| 249 |
+
"epoch": 3.993324432576769,
|
| 250 |
+
"grad_norm": 24.558015823364258,
|
| 251 |
+
"learning_rate": 1.779550827423168e-05,
|
| 252 |
+
"loss": 1.4304,
|
| 253 |
+
"step": 750
|
| 254 |
+
},
|
| 255 |
+
{
|
| 256 |
+
"epoch": 4.0,
|
| 257 |
+
"eval_accuracy": 0.826302729528536,
|
| 258 |
+
"eval_loss": 0.515184223651886,
|
| 259 |
+
"eval_macro_f1": 0.8249221583751791,
|
| 260 |
+
"eval_runtime": 0.8763,
|
| 261 |
+
"eval_samples_per_second": 919.813,
|
| 262 |
+
"eval_steps_per_second": 58.202,
|
| 263 |
+
"eval_weighted_f1": 0.8247302337608486,
|
| 264 |
+
"step": 752
|
| 265 |
+
},
|
| 266 |
+
{
|
| 267 |
+
"epoch": 4.12283044058745,
|
| 268 |
+
"grad_norm": 10.826188087463379,
|
| 269 |
+
"learning_rate": 1.7647754137115842e-05,
|
| 270 |
+
"loss": 0.6535,
|
| 271 |
+
"step": 775
|
| 272 |
+
},
|
| 273 |
+
{
|
| 274 |
+
"epoch": 4.25634178905207,
|
| 275 |
+
"grad_norm": 51.57345962524414,
|
| 276 |
+
"learning_rate": 1.7500000000000002e-05,
|
| 277 |
+
"loss": 0.8793,
|
| 278 |
+
"step": 800
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 4.389853137516689,
|
| 282 |
+
"grad_norm": 66.71621704101562,
|
| 283 |
+
"learning_rate": 1.7352245862884162e-05,
|
| 284 |
+
"loss": 0.9412,
|
| 285 |
+
"step": 825
|
| 286 |
+
},
|
| 287 |
+
{
|
| 288 |
+
"epoch": 4.5233644859813085,
|
| 289 |
+
"grad_norm": 52.68075942993164,
|
| 290 |
+
"learning_rate": 1.7204491725768322e-05,
|
| 291 |
+
"loss": 0.9856,
|
| 292 |
+
"step": 850
|
| 293 |
+
},
|
| 294 |
+
{
|
| 295 |
+
"epoch": 4.656875834445928,
|
| 296 |
+
"grad_norm": 26.891000747680664,
|
| 297 |
+
"learning_rate": 1.7056737588652482e-05,
|
| 298 |
+
"loss": 0.7824,
|
| 299 |
+
"step": 875
|
| 300 |
+
},
|
| 301 |
+
{
|
| 302 |
+
"epoch": 4.790387182910547,
|
| 303 |
+
"grad_norm": 49.95881271362305,
|
| 304 |
+
"learning_rate": 1.6908983451536646e-05,
|
| 305 |
+
"loss": 0.7956,
|
| 306 |
+
"step": 900
|
| 307 |
+
},
|
| 308 |
+
{
|
| 309 |
+
"epoch": 4.923898531375167,
|
| 310 |
+
"grad_norm": 60.759315490722656,
|
| 311 |
+
"learning_rate": 1.6761229314420806e-05,
|
| 312 |
+
"loss": 0.8574,
|
| 313 |
+
"step": 925
|
| 314 |
+
},
|
| 315 |
+
{
|
| 316 |
+
"epoch": 5.0,
|
| 317 |
+
"eval_accuracy": 0.8312655086848635,
|
| 318 |
+
"eval_loss": 0.5548403859138489,
|
| 319 |
+
"eval_macro_f1": 0.8300582328098267,
|
| 320 |
+
"eval_runtime": 0.8932,
|
| 321 |
+
"eval_samples_per_second": 902.398,
|
| 322 |
+
"eval_steps_per_second": 57.1,
|
| 323 |
+
"eval_weighted_f1": 0.8299638559696316,
|
| 324 |
+
"step": 940
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 5.0534045393858475,
|
| 328 |
+
"grad_norm": 34.98723220825195,
|
| 329 |
+
"learning_rate": 1.6613475177304966e-05,
|
| 330 |
+
"loss": 0.5591,
|
| 331 |
+
"step": 950
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 5.186915887850467,
|
| 335 |
+
"grad_norm": 47.1026496887207,
|
| 336 |
+
"learning_rate": 1.6465721040189126e-05,
|
| 337 |
+
"loss": 0.74,
|
| 338 |
+
"step": 975
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 5.320427236315087,
|
| 342 |
+
"grad_norm": 65.67156219482422,
|
| 343 |
+
"learning_rate": 1.6317966903073286e-05,
|
| 344 |
+
"loss": 0.5691,
|
| 345 |
+
"step": 1000
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 5.453938584779706,
|
| 349 |
+
"grad_norm": 114.16455078125,
|
| 350 |
+
"learning_rate": 1.6170212765957446e-05,
|
| 351 |
+
"loss": 0.559,
|
| 352 |
+
"step": 1025
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 5.587449933244326,
|
| 356 |
+
"grad_norm": 56.513545989990234,
|
| 357 |
+
"learning_rate": 1.602245862884161e-05,
|
| 358 |
+
"loss": 0.4883,
|
| 359 |
+
"step": 1050
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 5.720961281708945,
|
| 363 |
+
"grad_norm": 26.13739776611328,
|
| 364 |
+
"learning_rate": 1.587470449172577e-05,
|
| 365 |
+
"loss": 0.5287,
|
| 366 |
+
"step": 1075
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 5.854472630173564,
|
| 370 |
+
"grad_norm": 52.19254684448242,
|
| 371 |
+
"learning_rate": 1.572695035460993e-05,
|
| 372 |
+
"loss": 0.6611,
|
| 373 |
+
"step": 1100
|
| 374 |
+
},
|
| 375 |
+
{
|
| 376 |
+
"epoch": 5.9879839786381845,
|
| 377 |
+
"grad_norm": 149.34495544433594,
|
| 378 |
+
"learning_rate": 1.557919621749409e-05,
|
| 379 |
+
"loss": 0.5178,
|
| 380 |
+
"step": 1125
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 6.0,
|
| 384 |
+
"eval_accuracy": 0.8163771712158809,
|
| 385 |
+
"eval_loss": 0.6910102963447571,
|
| 386 |
+
"eval_macro_f1": 0.8147348816728024,
|
| 387 |
+
"eval_runtime": 0.933,
|
| 388 |
+
"eval_samples_per_second": 863.85,
|
| 389 |
+
"eval_steps_per_second": 54.66,
|
| 390 |
+
"eval_weighted_f1": 0.8147177066425263,
|
| 391 |
+
"step": 1128
|
| 392 |
+
},
|
| 393 |
+
{
|
| 394 |
+
"epoch": 6.117489986648865,
|
| 395 |
+
"grad_norm": 11.702919960021973,
|
| 396 |
+
"learning_rate": 1.543144208037825e-05,
|
| 397 |
+
"loss": 0.4257,
|
| 398 |
+
"step": 1150
|
| 399 |
+
},
|
| 400 |
+
{
|
| 401 |
+
"epoch": 6.251001335113485,
|
| 402 |
+
"grad_norm": 50.10649490356445,
|
| 403 |
+
"learning_rate": 1.5283687943262414e-05,
|
| 404 |
+
"loss": 0.5115,
|
| 405 |
+
"step": 1175
|
| 406 |
+
},
|
| 407 |
+
{
|
| 408 |
+
"epoch": 6.384512683578104,
|
| 409 |
+
"grad_norm": 1.3259004354476929,
|
| 410 |
+
"learning_rate": 1.5135933806146574e-05,
|
| 411 |
+
"loss": 0.3388,
|
| 412 |
+
"step": 1200
|
| 413 |
+
},
|
| 414 |
+
{
|
| 415 |
+
"epoch": 6.518024032042724,
|
| 416 |
+
"grad_norm": 25.595294952392578,
|
| 417 |
+
"learning_rate": 1.4988179669030734e-05,
|
| 418 |
+
"loss": 0.5207,
|
| 419 |
+
"step": 1225
|
| 420 |
+
},
|
| 421 |
+
{
|
| 422 |
+
"epoch": 6.651535380507343,
|
| 423 |
+
"grad_norm": 35.91559982299805,
|
| 424 |
+
"learning_rate": 1.4840425531914894e-05,
|
| 425 |
+
"loss": 0.4556,
|
| 426 |
+
"step": 1250
|
| 427 |
+
},
|
| 428 |
+
{
|
| 429 |
+
"epoch": 6.785046728971962,
|
| 430 |
+
"grad_norm": 120.19802856445312,
|
| 431 |
+
"learning_rate": 1.4692671394799056e-05,
|
| 432 |
+
"loss": 0.3891,
|
| 433 |
+
"step": 1275
|
| 434 |
+
},
|
| 435 |
+
{
|
| 436 |
+
"epoch": 6.9185580774365825,
|
| 437 |
+
"grad_norm": 42.937896728515625,
|
| 438 |
+
"learning_rate": 1.4544917257683215e-05,
|
| 439 |
+
"loss": 0.4652,
|
| 440 |
+
"step": 1300
|
| 441 |
+
},
|
| 442 |
+
{
|
| 443 |
+
"epoch": 7.0,
|
| 444 |
+
"eval_accuracy": 0.8325062034739454,
|
| 445 |
+
"eval_loss": 0.7512131333351135,
|
| 446 |
+
"eval_macro_f1": 0.8317569687847032,
|
| 447 |
+
"eval_runtime": 0.8853,
|
| 448 |
+
"eval_samples_per_second": 910.411,
|
| 449 |
+
"eval_steps_per_second": 57.607,
|
| 450 |
+
"eval_weighted_f1": 0.8317282250216943,
|
| 451 |
+
"step": 1316
|
| 452 |
+
},
|
| 453 |
+
{
|
| 454 |
+
"epoch": 7.048064085447263,
|
| 455 |
+
"grad_norm": 0.8096206188201904,
|
| 456 |
+
"learning_rate": 1.4397163120567377e-05,
|
| 457 |
+
"loss": 0.5165,
|
| 458 |
+
"step": 1325
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 7.181575433911883,
|
| 462 |
+
"grad_norm": 9.723548889160156,
|
| 463 |
+
"learning_rate": 1.4249408983451538e-05,
|
| 464 |
+
"loss": 0.3382,
|
| 465 |
+
"step": 1350
|
| 466 |
+
},
|
| 467 |
+
{
|
| 468 |
+
"epoch": 7.315086782376502,
|
| 469 |
+
"grad_norm": 97.81317138671875,
|
| 470 |
+
"learning_rate": 1.41016548463357e-05,
|
| 471 |
+
"loss": 0.2261,
|
| 472 |
+
"step": 1375
|
| 473 |
+
},
|
| 474 |
+
{
|
| 475 |
+
"epoch": 7.4485981308411215,
|
| 476 |
+
"grad_norm": 0.24998800456523895,
|
| 477 |
+
"learning_rate": 1.3953900709219859e-05,
|
| 478 |
+
"loss": 0.2369,
|
| 479 |
+
"step": 1400
|
| 480 |
+
},
|
| 481 |
+
{
|
| 482 |
+
"epoch": 7.582109479305741,
|
| 483 |
+
"grad_norm": 6.461670398712158,
|
| 484 |
+
"learning_rate": 1.380614657210402e-05,
|
| 485 |
+
"loss": 0.2873,
|
| 486 |
+
"step": 1425
|
| 487 |
+
},
|
| 488 |
+
{
|
| 489 |
+
"epoch": 7.71562082777036,
|
| 490 |
+
"grad_norm": 0.23604150116443634,
|
| 491 |
+
"learning_rate": 1.365839243498818e-05,
|
| 492 |
+
"loss": 0.398,
|
| 493 |
+
"step": 1450
|
| 494 |
+
},
|
| 495 |
+
{
|
| 496 |
+
"epoch": 7.8491321762349795,
|
| 497 |
+
"grad_norm": 52.03115463256836,
|
| 498 |
+
"learning_rate": 1.351063829787234e-05,
|
| 499 |
+
"loss": 0.5035,
|
| 500 |
+
"step": 1475
|
| 501 |
+
},
|
| 502 |
+
{
|
| 503 |
+
"epoch": 7.9826435246996,
|
| 504 |
+
"grad_norm": 69.65509796142578,
|
| 505 |
+
"learning_rate": 1.3362884160756503e-05,
|
| 506 |
+
"loss": 0.3796,
|
| 507 |
+
"step": 1500
|
| 508 |
+
},
|
| 509 |
+
{
|
| 510 |
+
"epoch": 8.0,
|
| 511 |
+
"eval_accuracy": 0.825062034739454,
|
| 512 |
+
"eval_loss": 0.9201409220695496,
|
| 513 |
+
"eval_macro_f1": 0.8237998988550653,
|
| 514 |
+
"eval_runtime": 0.8769,
|
| 515 |
+
"eval_samples_per_second": 919.161,
|
| 516 |
+
"eval_steps_per_second": 58.16,
|
| 517 |
+
"eval_weighted_f1": 0.8236155440779873,
|
| 518 |
+
"step": 1504
|
| 519 |
+
},
|
| 520 |
+
{
|
| 521 |
+
"epoch": 8.11214953271028,
|
| 522 |
+
"grad_norm": 66.33985137939453,
|
| 523 |
+
"learning_rate": 1.3215130023640663e-05,
|
| 524 |
+
"loss": 0.2118,
|
| 525 |
+
"step": 1525
|
| 526 |
+
},
|
| 527 |
+
{
|
| 528 |
+
"epoch": 8.2456608811749,
|
| 529 |
+
"grad_norm": 1.323639988899231,
|
| 530 |
+
"learning_rate": 1.3067375886524825e-05,
|
| 531 |
+
"loss": 0.3528,
|
| 532 |
+
"step": 1550
|
| 533 |
+
},
|
| 534 |
+
{
|
| 535 |
+
"epoch": 8.379172229639519,
|
| 536 |
+
"grad_norm": 0.2407652735710144,
|
| 537 |
+
"learning_rate": 1.2919621749408983e-05,
|
| 538 |
+
"loss": 0.1385,
|
| 539 |
+
"step": 1575
|
| 540 |
+
},
|
| 541 |
+
{
|
| 542 |
+
"epoch": 8.51268357810414,
|
| 543 |
+
"grad_norm": 246.52987670898438,
|
| 544 |
+
"learning_rate": 1.2771867612293145e-05,
|
| 545 |
+
"loss": 0.1627,
|
| 546 |
+
"step": 1600
|
| 547 |
+
},
|
| 548 |
+
{
|
| 549 |
+
"epoch": 8.646194926568759,
|
| 550 |
+
"grad_norm": 31.24919319152832,
|
| 551 |
+
"learning_rate": 1.2624113475177307e-05,
|
| 552 |
+
"loss": 0.1876,
|
| 553 |
+
"step": 1625
|
| 554 |
+
},
|
| 555 |
+
{
|
| 556 |
+
"epoch": 8.779706275033378,
|
| 557 |
+
"grad_norm": 60.09649658203125,
|
| 558 |
+
"learning_rate": 1.2476359338061469e-05,
|
| 559 |
+
"loss": 0.3086,
|
| 560 |
+
"step": 1650
|
| 561 |
+
},
|
| 562 |
+
{
|
| 563 |
+
"epoch": 8.913217623497998,
|
| 564 |
+
"grad_norm": 0.33130812644958496,
|
| 565 |
+
"learning_rate": 1.2328605200945627e-05,
|
| 566 |
+
"loss": 0.2882,
|
| 567 |
+
"step": 1675
|
| 568 |
+
},
|
| 569 |
+
{
|
| 570 |
+
"epoch": 9.0,
|
| 571 |
+
"eval_accuracy": 0.8399503722084367,
|
| 572 |
+
"eval_loss": 0.980648398399353,
|
| 573 |
+
"eval_macro_f1": 0.8403151851486105,
|
| 574 |
+
"eval_runtime": 0.9351,
|
| 575 |
+
"eval_samples_per_second": 861.939,
|
| 576 |
+
"eval_steps_per_second": 54.54,
|
| 577 |
+
"eval_weighted_f1": 0.8403824279884705,
|
| 578 |
+
"step": 1692
|
| 579 |
+
},
|
| 580 |
+
{
|
| 581 |
+
"epoch": 9.042723631508679,
|
| 582 |
+
"grad_norm": 41.748104095458984,
|
| 583 |
+
"learning_rate": 1.2180851063829789e-05,
|
| 584 |
+
"loss": 0.3227,
|
| 585 |
+
"step": 1700
|
| 586 |
+
},
|
| 587 |
+
{
|
| 588 |
+
"epoch": 9.176234979973298,
|
| 589 |
+
"grad_norm": 0.1747618168592453,
|
| 590 |
+
"learning_rate": 1.2033096926713949e-05,
|
| 591 |
+
"loss": 0.3277,
|
| 592 |
+
"step": 1725
|
| 593 |
+
},
|
| 594 |
+
{
|
| 595 |
+
"epoch": 9.309746328437917,
|
| 596 |
+
"grad_norm": 33.10725402832031,
|
| 597 |
+
"learning_rate": 1.1885342789598109e-05,
|
| 598 |
+
"loss": 0.207,
|
| 599 |
+
"step": 1750
|
| 600 |
+
},
|
| 601 |
+
{
|
| 602 |
+
"epoch": 9.443257676902537,
|
| 603 |
+
"grad_norm": 0.09367255121469498,
|
| 604 |
+
"learning_rate": 1.1737588652482271e-05,
|
| 605 |
+
"loss": 0.209,
|
| 606 |
+
"step": 1775
|
| 607 |
+
},
|
| 608 |
+
{
|
| 609 |
+
"epoch": 9.576769025367156,
|
| 610 |
+
"grad_norm": 0.6507932543754578,
|
| 611 |
+
"learning_rate": 1.1589834515366431e-05,
|
| 612 |
+
"loss": 0.0832,
|
| 613 |
+
"step": 1800
|
| 614 |
+
},
|
| 615 |
+
{
|
| 616 |
+
"epoch": 9.710280373831775,
|
| 617 |
+
"grad_norm": 0.1129131093621254,
|
| 618 |
+
"learning_rate": 1.1442080378250593e-05,
|
| 619 |
+
"loss": 0.2333,
|
| 620 |
+
"step": 1825
|
| 621 |
+
},
|
| 622 |
+
{
|
| 623 |
+
"epoch": 9.843791722296395,
|
| 624 |
+
"grad_norm": 0.07176733762025833,
|
| 625 |
+
"learning_rate": 1.1294326241134751e-05,
|
| 626 |
+
"loss": 0.1906,
|
| 627 |
+
"step": 1850
|
| 628 |
+
},
|
| 629 |
+
{
|
| 630 |
+
"epoch": 9.977303070761014,
|
| 631 |
+
"grad_norm": 0.2096477448940277,
|
| 632 |
+
"learning_rate": 1.1146572104018913e-05,
|
| 633 |
+
"loss": 0.3196,
|
| 634 |
+
"step": 1875
|
| 635 |
+
},
|
| 636 |
+
{
|
| 637 |
+
"epoch": 10.0,
|
| 638 |
+
"eval_accuracy": 0.8362282878411911,
|
| 639 |
+
"eval_loss": 1.0051621198654175,
|
| 640 |
+
"eval_macro_f1": 0.8348796160995325,
|
| 641 |
+
"eval_runtime": 0.8914,
|
| 642 |
+
"eval_samples_per_second": 904.228,
|
| 643 |
+
"eval_steps_per_second": 57.215,
|
| 644 |
+
"eval_weighted_f1": 0.8347690228773457,
|
| 645 |
+
"step": 1880
|
| 646 |
+
},
|
| 647 |
+
{
|
| 648 |
+
"epoch": 10.106809078771695,
|
| 649 |
+
"grad_norm": 14.35685920715332,
|
| 650 |
+
"learning_rate": 1.0998817966903075e-05,
|
| 651 |
+
"loss": 0.1879,
|
| 652 |
+
"step": 1900
|
| 653 |
+
},
|
| 654 |
+
{
|
| 655 |
+
"epoch": 10.240320427236314,
|
| 656 |
+
"grad_norm": 0.13472262024879456,
|
| 657 |
+
"learning_rate": 1.0851063829787233e-05,
|
| 658 |
+
"loss": 0.1757,
|
| 659 |
+
"step": 1925
|
| 660 |
+
},
|
| 661 |
+
{
|
| 662 |
+
"epoch": 10.373831775700934,
|
| 663 |
+
"grad_norm": 214.050048828125,
|
| 664 |
+
"learning_rate": 1.0703309692671395e-05,
|
| 665 |
+
"loss": 0.2152,
|
| 666 |
+
"step": 1950
|
| 667 |
+
},
|
| 668 |
+
{
|
| 669 |
+
"epoch": 10.507343124165555,
|
| 670 |
+
"grad_norm": 0.04622187465429306,
|
| 671 |
+
"learning_rate": 1.0555555555555557e-05,
|
| 672 |
+
"loss": 0.1124,
|
| 673 |
+
"step": 1975
|
| 674 |
+
},
|
| 675 |
+
{
|
| 676 |
+
"epoch": 10.640854472630174,
|
| 677 |
+
"grad_norm": 96.18000793457031,
|
| 678 |
+
"learning_rate": 1.0407801418439717e-05,
|
| 679 |
+
"loss": 0.2627,
|
| 680 |
+
"step": 2000
|
| 681 |
+
},
|
| 682 |
+
{
|
| 683 |
+
"epoch": 10.774365821094793,
|
| 684 |
+
"grad_norm": 140.4534454345703,
|
| 685 |
+
"learning_rate": 1.0260047281323877e-05,
|
| 686 |
+
"loss": 0.1471,
|
| 687 |
+
"step": 2025
|
| 688 |
+
},
|
| 689 |
+
{
|
| 690 |
+
"epoch": 10.907877169559413,
|
| 691 |
+
"grad_norm": 1.6079241037368774,
|
| 692 |
+
"learning_rate": 1.011229314420804e-05,
|
| 693 |
+
"loss": 0.0841,
|
| 694 |
+
"step": 2050
|
| 695 |
+
},
|
| 696 |
+
{
|
| 697 |
+
"epoch": 11.0,
|
| 698 |
+
"eval_accuracy": 0.8362282878411911,
|
| 699 |
+
"eval_loss": 1.0941190719604492,
|
| 700 |
+
"eval_macro_f1": 0.836169357748632,
|
| 701 |
+
"eval_runtime": 0.8835,
|
| 702 |
+
"eval_samples_per_second": 912.244,
|
| 703 |
+
"eval_steps_per_second": 57.723,
|
| 704 |
+
"eval_weighted_f1": 0.8360102608673665,
|
| 705 |
+
"step": 2068
|
| 706 |
+
},
|
| 707 |
+
{
|
| 708 |
+
"epoch": 11.037383177570094,
|
| 709 |
+
"grad_norm": 0.06270817667245865,
|
| 710 |
+
"learning_rate": 9.9645390070922e-06,
|
| 711 |
+
"loss": 0.1672,
|
| 712 |
+
"step": 2075
|
| 713 |
+
},
|
| 714 |
+
{
|
| 715 |
+
"epoch": 11.170894526034713,
|
| 716 |
+
"grad_norm": 7.355442047119141,
|
| 717 |
+
"learning_rate": 9.81678486997636e-06,
|
| 718 |
+
"loss": 0.1417,
|
| 719 |
+
"step": 2100
|
| 720 |
+
},
|
| 721 |
+
{
|
| 722 |
+
"epoch": 11.304405874499333,
|
| 723 |
+
"grad_norm": 0.06039491295814514,
|
| 724 |
+
"learning_rate": 9.66903073286052e-06,
|
| 725 |
+
"loss": 0.1179,
|
| 726 |
+
"step": 2125
|
| 727 |
+
},
|
| 728 |
+
{
|
| 729 |
+
"epoch": 11.437917222963952,
|
| 730 |
+
"grad_norm": 0.11568111181259155,
|
| 731 |
+
"learning_rate": 9.521276595744681e-06,
|
| 732 |
+
"loss": 0.1319,
|
| 733 |
+
"step": 2150
|
| 734 |
+
},
|
| 735 |
+
{
|
| 736 |
+
"epoch": 11.571428571428571,
|
| 737 |
+
"grad_norm": 25.074573516845703,
|
| 738 |
+
"learning_rate": 9.373522458628842e-06,
|
| 739 |
+
"loss": 0.1443,
|
| 740 |
+
"step": 2175
|
| 741 |
+
},
|
| 742 |
+
{
|
| 743 |
+
"epoch": 11.70493991989319,
|
| 744 |
+
"grad_norm": 0.10229099541902542,
|
| 745 |
+
"learning_rate": 9.225768321513003e-06,
|
| 746 |
+
"loss": 0.1286,
|
| 747 |
+
"step": 2200
|
| 748 |
+
},
|
| 749 |
+
{
|
| 750 |
+
"epoch": 11.83845126835781,
|
| 751 |
+
"grad_norm": 0.03965507447719574,
|
| 752 |
+
"learning_rate": 9.078014184397164e-06,
|
| 753 |
+
"loss": 0.1009,
|
| 754 |
+
"step": 2225
|
| 755 |
+
},
|
| 756 |
+
{
|
| 757 |
+
"epoch": 11.97196261682243,
|
| 758 |
+
"grad_norm": 0.7916253209114075,
|
| 759 |
+
"learning_rate": 8.930260047281325e-06,
|
| 760 |
+
"loss": 0.1324,
|
| 761 |
+
"step": 2250
|
| 762 |
+
},
|
| 763 |
+
{
|
| 764 |
+
"epoch": 12.0,
|
| 765 |
+
"eval_accuracy": 0.8374689826302729,
|
| 766 |
+
"eval_loss": 1.128941535949707,
|
| 767 |
+
"eval_macro_f1": 0.837426678830166,
|
| 768 |
+
"eval_runtime": 0.8782,
|
| 769 |
+
"eval_samples_per_second": 917.799,
|
| 770 |
+
"eval_steps_per_second": 58.074,
|
| 771 |
+
"eval_weighted_f1": 0.8371058384380181,
|
| 772 |
+
"step": 2256
|
| 773 |
+
}
|
| 774 |
+
],
|
| 775 |
+
"logging_steps": 25,
|
| 776 |
+
"max_steps": 3760,
|
| 777 |
+
"num_input_tokens_seen": 0,
|
| 778 |
+
"num_train_epochs": 20,
|
| 779 |
+
"save_steps": 500,
|
| 780 |
+
"stateful_callbacks": {
|
| 781 |
+
"EarlyStoppingCallback": {
|
| 782 |
+
"args": {
|
| 783 |
+
"early_stopping_patience": 3,
|
| 784 |
+
"early_stopping_threshold": 0.0
|
| 785 |
+
},
|
| 786 |
+
"attributes": {
|
| 787 |
+
"early_stopping_patience_counter": 3
|
| 788 |
+
}
|
| 789 |
+
},
|
| 790 |
+
"TrainerControl": {
|
| 791 |
+
"args": {
|
| 792 |
+
"should_epoch_stop": false,
|
| 793 |
+
"should_evaluate": false,
|
| 794 |
+
"should_log": false,
|
| 795 |
+
"should_save": true,
|
| 796 |
+
"should_training_stop": true
|
| 797 |
+
},
|
| 798 |
+
"attributes": {}
|
| 799 |
+
}
|
| 800 |
+
},
|
| 801 |
+
"total_flos": 0.0,
|
| 802 |
+
"train_batch_size": 8,
|
| 803 |
+
"trial_name": null,
|
| 804 |
+
"trial_params": null
|
| 805 |
+
}
|
ood-vifactcheck-bamibert-topic-domain-05-s43-20260827T195058Z-3858bb98/trainer/checkpoint-1575/special_tokens_map.json
ADDED
|
@@ -0,0 +1,51 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": {
|
| 3 |
+
"content": "<s>",
|
| 4 |
+
"lstrip": false,
|
| 5 |
+
"normalized": false,
|
| 6 |
+
"rstrip": false,
|
| 7 |
+
"single_word": false
|
| 8 |
+
},
|
| 9 |
+
"cls_token": {
|
| 10 |
+
"content": "<s>",
|
| 11 |
+
"lstrip": false,
|
| 12 |
+
"normalized": false,
|
| 13 |
+
"rstrip": false,
|
| 14 |
+
"single_word": false
|
| 15 |
+
},
|
| 16 |
+
"eos_token": {
|
| 17 |
+
"content": "</s>",
|
| 18 |
+
"lstrip": false,
|
| 19 |
+
"normalized": false,
|
| 20 |
+
"rstrip": false,
|
| 21 |
+
"single_word": false
|
| 22 |
+
},
|
| 23 |
+
"mask_token": {
|
| 24 |
+
"content": "<mask>",
|
| 25 |
+
"lstrip": false,
|
| 26 |
+
"normalized": false,
|
| 27 |
+
"rstrip": false,
|
| 28 |
+
"single_word": false
|
| 29 |
+
},
|
| 30 |
+
"pad_token": {
|
| 31 |
+
"content": "<pad>",
|
| 32 |
+
"lstrip": false,
|
| 33 |
+
"normalized": false,
|
| 34 |
+
"rstrip": false,
|
| 35 |
+
"single_word": false
|
| 36 |
+
},
|
| 37 |
+
"sep_token": {
|
| 38 |
+
"content": "</s>",
|
| 39 |
+
"lstrip": false,
|
| 40 |
+
"normalized": false,
|
| 41 |
+
"rstrip": false,
|
| 42 |
+
"single_word": false
|
| 43 |
+
},
|
| 44 |
+
"unk_token": {
|
| 45 |
+
"content": "<unk>",
|
| 46 |
+
"lstrip": false,
|
| 47 |
+
"normalized": false,
|
| 48 |
+
"rstrip": false,
|
| 49 |
+
"single_word": false
|
| 50 |
+
}
|
| 51 |
+
}
|
ood-vifactcheck-bamibert-topic-domain-05-s43-20260827T195058Z-3858bb98/trainer/checkpoint-1575/tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-bamibert-topic-domain-05-s43-20260827T195058Z-3858bb98/trainer/checkpoint-1575/tokenizer_config.json
ADDED
|
@@ -0,0 +1,55 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"added_tokens_decoder": {
|
| 3 |
+
"0": {
|
| 4 |
+
"content": "<s>",
|
| 5 |
+
"lstrip": false,
|
| 6 |
+
"normalized": false,
|
| 7 |
+
"rstrip": false,
|
| 8 |
+
"single_word": false,
|
| 9 |
+
"special": true
|
| 10 |
+
},
|
| 11 |
+
"1": {
|
| 12 |
+
"content": "<pad>",
|
| 13 |
+
"lstrip": false,
|
| 14 |
+
"normalized": false,
|
| 15 |
+
"rstrip": false,
|
| 16 |
+
"single_word": false,
|
| 17 |
+
"special": true
|
| 18 |
+
},
|
| 19 |
+
"2": {
|
| 20 |
+
"content": "</s>",
|
| 21 |
+
"lstrip": false,
|
| 22 |
+
"normalized": false,
|
| 23 |
+
"rstrip": false,
|
| 24 |
+
"single_word": false,
|
| 25 |
+
"special": true
|
| 26 |
+
},
|
| 27 |
+
"3": {
|
| 28 |
+
"content": "<unk>",
|
| 29 |
+
"lstrip": false,
|
| 30 |
+
"normalized": false,
|
| 31 |
+
"rstrip": false,
|
| 32 |
+
"single_word": false,
|
| 33 |
+
"special": true
|
| 34 |
+
},
|
| 35 |
+
"20480": {
|
| 36 |
+
"content": "<mask>",
|
| 37 |
+
"lstrip": false,
|
| 38 |
+
"normalized": false,
|
| 39 |
+
"rstrip": false,
|
| 40 |
+
"single_word": false,
|
| 41 |
+
"special": true
|
| 42 |
+
}
|
| 43 |
+
},
|
| 44 |
+
"bos_token": "<s>",
|
| 45 |
+
"clean_up_tokenization_spaces": false,
|
| 46 |
+
"cls_token": "<s>",
|
| 47 |
+
"eos_token": "</s>",
|
| 48 |
+
"extra_special_tokens": {},
|
| 49 |
+
"mask_token": "<mask>",
|
| 50 |
+
"model_max_length": 2048,
|
| 51 |
+
"pad_token": "<pad>",
|
| 52 |
+
"sep_token": "</s>",
|
| 53 |
+
"tokenizer_class": "XLMRobertaTokenizerFast",
|
| 54 |
+
"unk_token": "<unk>"
|
| 55 |
+
}
|
ood-vifactcheck-bamibert-topic-domain-05-s43-20260827T195058Z-3858bb98/trainer/checkpoint-1575/trainer_state.json
ADDED
|
@@ -0,0 +1,583 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 1575,
|
| 3 |
+
"best_metric": 0.8385877780941896,
|
| 4 |
+
"best_model_checkpoint": "/root/VNEncoder/results/checkpoints/ood-vifactcheck-bamibert-topic-domain-05-s43-20260827T195058Z-3858bb98/trainer/checkpoint-1575",
|
| 5 |
+
"epoch": 9.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 1575,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.1430615164520744,
|
| 14 |
+
"grad_norm": 14.410805702209473,
|
| 15 |
+
"learning_rate": 1.3714285714285717e-06,
|
| 16 |
+
"loss": 4.3997,
|
| 17 |
+
"step": 25
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.2861230329041488,
|
| 21 |
+
"grad_norm": 9.47131633758545,
|
| 22 |
+
"learning_rate": 2.8000000000000003e-06,
|
| 23 |
+
"loss": 4.4248,
|
| 24 |
+
"step": 50
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.4291845493562232,
|
| 28 |
+
"grad_norm": 6.229504585266113,
|
| 29 |
+
"learning_rate": 4.228571428571429e-06,
|
| 30 |
+
"loss": 4.4037,
|
| 31 |
+
"step": 75
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.5722460658082976,
|
| 35 |
+
"grad_norm": 9.012520790100098,
|
| 36 |
+
"learning_rate": 5.6571428571428576e-06,
|
| 37 |
+
"loss": 4.3548,
|
| 38 |
+
"step": 100
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.7153075822603719,
|
| 42 |
+
"grad_norm": 14.192707061767578,
|
| 43 |
+
"learning_rate": 7.085714285714286e-06,
|
| 44 |
+
"loss": 4.3811,
|
| 45 |
+
"step": 125
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.8583690987124464,
|
| 49 |
+
"grad_norm": 9.538655281066895,
|
| 50 |
+
"learning_rate": 8.514285714285715e-06,
|
| 51 |
+
"loss": 4.3266,
|
| 52 |
+
"step": 150
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"epoch": 1.0,
|
| 56 |
+
"grad_norm": 9.140969276428223,
|
| 57 |
+
"learning_rate": 9.942857142857145e-06,
|
| 58 |
+
"loss": 4.245,
|
| 59 |
+
"step": 175
|
| 60 |
+
},
|
| 61 |
+
{
|
| 62 |
+
"epoch": 1.0,
|
| 63 |
+
"eval_accuracy": 0.45956873315363883,
|
| 64 |
+
"eval_loss": 1.0388729572296143,
|
| 65 |
+
"eval_macro_f1": 0.39965935091985516,
|
| 66 |
+
"eval_runtime": 0.8432,
|
| 67 |
+
"eval_samples_per_second": 879.999,
|
| 68 |
+
"eval_steps_per_second": 55.741,
|
| 69 |
+
"eval_weighted_f1": 0.40267089211799173,
|
| 70 |
+
"step": 175
|
| 71 |
+
},
|
| 72 |
+
{
|
| 73 |
+
"epoch": 1.1430615164520743,
|
| 74 |
+
"grad_norm": 22.30449676513672,
|
| 75 |
+
"learning_rate": 1.1371428571428573e-05,
|
| 76 |
+
"loss": 4.169,
|
| 77 |
+
"step": 200
|
| 78 |
+
},
|
| 79 |
+
{
|
| 80 |
+
"epoch": 1.2861230329041489,
|
| 81 |
+
"grad_norm": 52.536251068115234,
|
| 82 |
+
"learning_rate": 1.2800000000000001e-05,
|
| 83 |
+
"loss": 3.7333,
|
| 84 |
+
"step": 225
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 1.4291845493562232,
|
| 88 |
+
"grad_norm": 34.337257385253906,
|
| 89 |
+
"learning_rate": 1.422857142857143e-05,
|
| 90 |
+
"loss": 3.4397,
|
| 91 |
+
"step": 250
|
| 92 |
+
},
|
| 93 |
+
{
|
| 94 |
+
"epoch": 1.5722460658082977,
|
| 95 |
+
"grad_norm": 43.14560317993164,
|
| 96 |
+
"learning_rate": 1.5657142857142856e-05,
|
| 97 |
+
"loss": 2.9114,
|
| 98 |
+
"step": 275
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 1.715307582260372,
|
| 102 |
+
"grad_norm": 64.7920913696289,
|
| 103 |
+
"learning_rate": 1.708571428571429e-05,
|
| 104 |
+
"loss": 2.5674,
|
| 105 |
+
"step": 300
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 1.8583690987124464,
|
| 109 |
+
"grad_norm": 44.81905746459961,
|
| 110 |
+
"learning_rate": 1.8514285714285716e-05,
|
| 111 |
+
"loss": 2.5042,
|
| 112 |
+
"step": 325
|
| 113 |
+
},
|
| 114 |
+
{
|
| 115 |
+
"epoch": 2.0,
|
| 116 |
+
"grad_norm": 31.639690399169922,
|
| 117 |
+
"learning_rate": 1.9942857142857142e-05,
|
| 118 |
+
"loss": 2.3798,
|
| 119 |
+
"step": 350
|
| 120 |
+
},
|
| 121 |
+
{
|
| 122 |
+
"epoch": 2.0,
|
| 123 |
+
"eval_accuracy": 0.7547169811320755,
|
| 124 |
+
"eval_loss": 0.5748988389968872,
|
| 125 |
+
"eval_macro_f1": 0.7549644564941046,
|
| 126 |
+
"eval_runtime": 0.8586,
|
| 127 |
+
"eval_samples_per_second": 864.159,
|
| 128 |
+
"eval_steps_per_second": 54.738,
|
| 129 |
+
"eval_weighted_f1": 0.7554270901473462,
|
| 130 |
+
"step": 350
|
| 131 |
+
},
|
| 132 |
+
{
|
| 133 |
+
"epoch": 2.1430615164520743,
|
| 134 |
+
"grad_norm": 58.28070831298828,
|
| 135 |
+
"learning_rate": 1.984761904761905e-05,
|
| 136 |
+
"loss": 1.9731,
|
| 137 |
+
"step": 375
|
| 138 |
+
},
|
| 139 |
+
{
|
| 140 |
+
"epoch": 2.2861230329041486,
|
| 141 |
+
"grad_norm": 36.115516662597656,
|
| 142 |
+
"learning_rate": 1.968888888888889e-05,
|
| 143 |
+
"loss": 1.8605,
|
| 144 |
+
"step": 400
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 2.429184549356223,
|
| 148 |
+
"grad_norm": 46.78152084350586,
|
| 149 |
+
"learning_rate": 1.953015873015873e-05,
|
| 150 |
+
"loss": 2.007,
|
| 151 |
+
"step": 425
|
| 152 |
+
},
|
| 153 |
+
{
|
| 154 |
+
"epoch": 2.5722460658082977,
|
| 155 |
+
"grad_norm": 55.40108871459961,
|
| 156 |
+
"learning_rate": 1.937142857142857e-05,
|
| 157 |
+
"loss": 1.9006,
|
| 158 |
+
"step": 450
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 2.715307582260372,
|
| 162 |
+
"grad_norm": 45.07542037963867,
|
| 163 |
+
"learning_rate": 1.9212698412698414e-05,
|
| 164 |
+
"loss": 1.6802,
|
| 165 |
+
"step": 475
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 2.8583690987124464,
|
| 169 |
+
"grad_norm": 27.396472930908203,
|
| 170 |
+
"learning_rate": 1.9053968253968257e-05,
|
| 171 |
+
"loss": 1.7654,
|
| 172 |
+
"step": 500
|
| 173 |
+
},
|
| 174 |
+
{
|
| 175 |
+
"epoch": 3.0,
|
| 176 |
+
"grad_norm": 31.59823989868164,
|
| 177 |
+
"learning_rate": 1.8895238095238096e-05,
|
| 178 |
+
"loss": 1.7634,
|
| 179 |
+
"step": 525
|
| 180 |
+
},
|
| 181 |
+
{
|
| 182 |
+
"epoch": 3.0,
|
| 183 |
+
"eval_accuracy": 0.807277628032345,
|
| 184 |
+
"eval_loss": 0.4905075430870056,
|
| 185 |
+
"eval_macro_f1": 0.8061139250249086,
|
| 186 |
+
"eval_runtime": 0.8535,
|
| 187 |
+
"eval_samples_per_second": 869.407,
|
| 188 |
+
"eval_steps_per_second": 55.07,
|
| 189 |
+
"eval_weighted_f1": 0.8064427025314329,
|
| 190 |
+
"step": 525
|
| 191 |
+
},
|
| 192 |
+
{
|
| 193 |
+
"epoch": 3.1430615164520743,
|
| 194 |
+
"grad_norm": 67.2063217163086,
|
| 195 |
+
"learning_rate": 1.8736507936507936e-05,
|
| 196 |
+
"loss": 0.996,
|
| 197 |
+
"step": 550
|
| 198 |
+
},
|
| 199 |
+
{
|
| 200 |
+
"epoch": 3.2861230329041486,
|
| 201 |
+
"grad_norm": 83.634521484375,
|
| 202 |
+
"learning_rate": 1.857777777777778e-05,
|
| 203 |
+
"loss": 1.2307,
|
| 204 |
+
"step": 575
|
| 205 |
+
},
|
| 206 |
+
{
|
| 207 |
+
"epoch": 3.429184549356223,
|
| 208 |
+
"grad_norm": 32.015079498291016,
|
| 209 |
+
"learning_rate": 1.8419047619047622e-05,
|
| 210 |
+
"loss": 1.1427,
|
| 211 |
+
"step": 600
|
| 212 |
+
},
|
| 213 |
+
{
|
| 214 |
+
"epoch": 3.5722460658082977,
|
| 215 |
+
"grad_norm": 41.71841812133789,
|
| 216 |
+
"learning_rate": 1.8260317460317462e-05,
|
| 217 |
+
"loss": 1.3008,
|
| 218 |
+
"step": 625
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 3.715307582260372,
|
| 222 |
+
"grad_norm": 59.132320404052734,
|
| 223 |
+
"learning_rate": 1.81015873015873e-05,
|
| 224 |
+
"loss": 1.2368,
|
| 225 |
+
"step": 650
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"epoch": 3.8583690987124464,
|
| 229 |
+
"grad_norm": 38.083065032958984,
|
| 230 |
+
"learning_rate": 1.7942857142857144e-05,
|
| 231 |
+
"loss": 1.2518,
|
| 232 |
+
"step": 675
|
| 233 |
+
},
|
| 234 |
+
{
|
| 235 |
+
"epoch": 4.0,
|
| 236 |
+
"grad_norm": 42.52290725708008,
|
| 237 |
+
"learning_rate": 1.7784126984126987e-05,
|
| 238 |
+
"loss": 1.2294,
|
| 239 |
+
"step": 700
|
| 240 |
+
},
|
| 241 |
+
{
|
| 242 |
+
"epoch": 4.0,
|
| 243 |
+
"eval_accuracy": 0.8032345013477089,
|
| 244 |
+
"eval_loss": 0.5453498363494873,
|
| 245 |
+
"eval_macro_f1": 0.8010587518679261,
|
| 246 |
+
"eval_runtime": 0.8577,
|
| 247 |
+
"eval_samples_per_second": 865.116,
|
| 248 |
+
"eval_steps_per_second": 54.798,
|
| 249 |
+
"eval_weighted_f1": 0.8013903122767734,
|
| 250 |
+
"step": 700
|
| 251 |
+
},
|
| 252 |
+
{
|
| 253 |
+
"epoch": 4.143061516452074,
|
| 254 |
+
"grad_norm": 62.421905517578125,
|
| 255 |
+
"learning_rate": 1.7625396825396827e-05,
|
| 256 |
+
"loss": 0.9209,
|
| 257 |
+
"step": 725
|
| 258 |
+
},
|
| 259 |
+
{
|
| 260 |
+
"epoch": 4.286123032904149,
|
| 261 |
+
"grad_norm": 30.84873390197754,
|
| 262 |
+
"learning_rate": 1.7466666666666667e-05,
|
| 263 |
+
"loss": 0.6917,
|
| 264 |
+
"step": 750
|
| 265 |
+
},
|
| 266 |
+
{
|
| 267 |
+
"epoch": 4.429184549356223,
|
| 268 |
+
"grad_norm": 37.597694396972656,
|
| 269 |
+
"learning_rate": 1.730793650793651e-05,
|
| 270 |
+
"loss": 0.753,
|
| 271 |
+
"step": 775
|
| 272 |
+
},
|
| 273 |
+
{
|
| 274 |
+
"epoch": 4.572246065808297,
|
| 275 |
+
"grad_norm": 75.85980987548828,
|
| 276 |
+
"learning_rate": 1.7149206349206353e-05,
|
| 277 |
+
"loss": 0.7342,
|
| 278 |
+
"step": 800
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 4.715307582260372,
|
| 282 |
+
"grad_norm": 53.75027084350586,
|
| 283 |
+
"learning_rate": 1.6990476190476192e-05,
|
| 284 |
+
"loss": 0.6969,
|
| 285 |
+
"step": 825
|
| 286 |
+
},
|
| 287 |
+
{
|
| 288 |
+
"epoch": 4.858369098712446,
|
| 289 |
+
"grad_norm": 7.4808173179626465,
|
| 290 |
+
"learning_rate": 1.6831746031746032e-05,
|
| 291 |
+
"loss": 0.7179,
|
| 292 |
+
"step": 850
|
| 293 |
+
},
|
| 294 |
+
{
|
| 295 |
+
"epoch": 5.0,
|
| 296 |
+
"grad_norm": 34.891693115234375,
|
| 297 |
+
"learning_rate": 1.6673015873015875e-05,
|
| 298 |
+
"loss": 0.9722,
|
| 299 |
+
"step": 875
|
| 300 |
+
},
|
| 301 |
+
{
|
| 302 |
+
"epoch": 5.0,
|
| 303 |
+
"eval_accuracy": 0.8207547169811321,
|
| 304 |
+
"eval_loss": 0.5585728287696838,
|
| 305 |
+
"eval_macro_f1": 0.8175550071748489,
|
| 306 |
+
"eval_runtime": 0.8441,
|
| 307 |
+
"eval_samples_per_second": 879.007,
|
| 308 |
+
"eval_steps_per_second": 55.678,
|
| 309 |
+
"eval_weighted_f1": 0.8176335983763029,
|
| 310 |
+
"step": 875
|
| 311 |
+
},
|
| 312 |
+
{
|
| 313 |
+
"epoch": 5.143061516452074,
|
| 314 |
+
"grad_norm": 37.72514724731445,
|
| 315 |
+
"learning_rate": 1.6514285714285714e-05,
|
| 316 |
+
"loss": 0.4463,
|
| 317 |
+
"step": 900
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 5.286123032904149,
|
| 321 |
+
"grad_norm": 67.63714599609375,
|
| 322 |
+
"learning_rate": 1.6355555555555557e-05,
|
| 323 |
+
"loss": 0.5709,
|
| 324 |
+
"step": 925
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 5.429184549356223,
|
| 328 |
+
"grad_norm": 20.562917709350586,
|
| 329 |
+
"learning_rate": 1.6196825396825397e-05,
|
| 330 |
+
"loss": 0.6067,
|
| 331 |
+
"step": 950
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 5.572246065808297,
|
| 335 |
+
"grad_norm": 64.2474365234375,
|
| 336 |
+
"learning_rate": 1.603809523809524e-05,
|
| 337 |
+
"loss": 0.4798,
|
| 338 |
+
"step": 975
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 5.715307582260372,
|
| 342 |
+
"grad_norm": 73.35626983642578,
|
| 343 |
+
"learning_rate": 1.587936507936508e-05,
|
| 344 |
+
"loss": 0.6928,
|
| 345 |
+
"step": 1000
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 5.858369098712446,
|
| 349 |
+
"grad_norm": 20.787761688232422,
|
| 350 |
+
"learning_rate": 1.5720634920634923e-05,
|
| 351 |
+
"loss": 0.5191,
|
| 352 |
+
"step": 1025
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 6.0,
|
| 356 |
+
"grad_norm": 39.936492919921875,
|
| 357 |
+
"learning_rate": 1.5561904761904762e-05,
|
| 358 |
+
"loss": 0.6271,
|
| 359 |
+
"step": 1050
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 6.0,
|
| 363 |
+
"eval_accuracy": 0.8261455525606469,
|
| 364 |
+
"eval_loss": 0.6241559386253357,
|
| 365 |
+
"eval_macro_f1": 0.825354877680459,
|
| 366 |
+
"eval_runtime": 0.8496,
|
| 367 |
+
"eval_samples_per_second": 873.399,
|
| 368 |
+
"eval_steps_per_second": 55.323,
|
| 369 |
+
"eval_weighted_f1": 0.8256656525947569,
|
| 370 |
+
"step": 1050
|
| 371 |
+
},
|
| 372 |
+
{
|
| 373 |
+
"epoch": 6.143061516452074,
|
| 374 |
+
"grad_norm": 114.56380462646484,
|
| 375 |
+
"learning_rate": 1.5403174603174605e-05,
|
| 376 |
+
"loss": 0.3623,
|
| 377 |
+
"step": 1075
|
| 378 |
+
},
|
| 379 |
+
{
|
| 380 |
+
"epoch": 6.286123032904149,
|
| 381 |
+
"grad_norm": 77.52657318115234,
|
| 382 |
+
"learning_rate": 1.5244444444444447e-05,
|
| 383 |
+
"loss": 0.391,
|
| 384 |
+
"step": 1100
|
| 385 |
+
},
|
| 386 |
+
{
|
| 387 |
+
"epoch": 6.429184549356223,
|
| 388 |
+
"grad_norm": 39.765647888183594,
|
| 389 |
+
"learning_rate": 1.5085714285714288e-05,
|
| 390 |
+
"loss": 0.5178,
|
| 391 |
+
"step": 1125
|
| 392 |
+
},
|
| 393 |
+
{
|
| 394 |
+
"epoch": 6.572246065808297,
|
| 395 |
+
"grad_norm": 6.1532087326049805,
|
| 396 |
+
"learning_rate": 1.4926984126984128e-05,
|
| 397 |
+
"loss": 0.5117,
|
| 398 |
+
"step": 1150
|
| 399 |
+
},
|
| 400 |
+
{
|
| 401 |
+
"epoch": 6.715307582260372,
|
| 402 |
+
"grad_norm": 73.53240966796875,
|
| 403 |
+
"learning_rate": 1.4768253968253969e-05,
|
| 404 |
+
"loss": 0.4157,
|
| 405 |
+
"step": 1175
|
| 406 |
+
},
|
| 407 |
+
{
|
| 408 |
+
"epoch": 6.858369098712446,
|
| 409 |
+
"grad_norm": 195.1527099609375,
|
| 410 |
+
"learning_rate": 1.4609523809523812e-05,
|
| 411 |
+
"loss": 0.5763,
|
| 412 |
+
"step": 1200
|
| 413 |
+
},
|
| 414 |
+
{
|
| 415 |
+
"epoch": 7.0,
|
| 416 |
+
"grad_norm": 81.26548767089844,
|
| 417 |
+
"learning_rate": 1.4450793650793653e-05,
|
| 418 |
+
"loss": 0.4112,
|
| 419 |
+
"step": 1225
|
| 420 |
+
},
|
| 421 |
+
{
|
| 422 |
+
"epoch": 7.0,
|
| 423 |
+
"eval_accuracy": 0.8018867924528302,
|
| 424 |
+
"eval_loss": 0.9285014867782593,
|
| 425 |
+
"eval_macro_f1": 0.8004517524849147,
|
| 426 |
+
"eval_runtime": 0.8406,
|
| 427 |
+
"eval_samples_per_second": 882.746,
|
| 428 |
+
"eval_steps_per_second": 55.915,
|
| 429 |
+
"eval_weighted_f1": 0.8001643468723298,
|
| 430 |
+
"step": 1225
|
| 431 |
+
},
|
| 432 |
+
{
|
| 433 |
+
"epoch": 7.143061516452074,
|
| 434 |
+
"grad_norm": 0.7759934663772583,
|
| 435 |
+
"learning_rate": 1.4292063492063493e-05,
|
| 436 |
+
"loss": 0.2987,
|
| 437 |
+
"step": 1250
|
| 438 |
+
},
|
| 439 |
+
{
|
| 440 |
+
"epoch": 7.286123032904149,
|
| 441 |
+
"grad_norm": 17.462034225463867,
|
| 442 |
+
"learning_rate": 1.4133333333333334e-05,
|
| 443 |
+
"loss": 0.3494,
|
| 444 |
+
"step": 1275
|
| 445 |
+
},
|
| 446 |
+
{
|
| 447 |
+
"epoch": 7.429184549356223,
|
| 448 |
+
"grad_norm": 0.44466474652290344,
|
| 449 |
+
"learning_rate": 1.3974603174603177e-05,
|
| 450 |
+
"loss": 0.3154,
|
| 451 |
+
"step": 1300
|
| 452 |
+
},
|
| 453 |
+
{
|
| 454 |
+
"epoch": 7.572246065808297,
|
| 455 |
+
"grad_norm": 26.82601547241211,
|
| 456 |
+
"learning_rate": 1.3815873015873018e-05,
|
| 457 |
+
"loss": 0.2817,
|
| 458 |
+
"step": 1325
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 7.715307582260372,
|
| 462 |
+
"grad_norm": 23.987606048583984,
|
| 463 |
+
"learning_rate": 1.3657142857142858e-05,
|
| 464 |
+
"loss": 0.3081,
|
| 465 |
+
"step": 1350
|
| 466 |
+
},
|
| 467 |
+
{
|
| 468 |
+
"epoch": 7.858369098712446,
|
| 469 |
+
"grad_norm": 30.050395965576172,
|
| 470 |
+
"learning_rate": 1.34984126984127e-05,
|
| 471 |
+
"loss": 0.6405,
|
| 472 |
+
"step": 1375
|
| 473 |
+
},
|
| 474 |
+
{
|
| 475 |
+
"epoch": 8.0,
|
| 476 |
+
"grad_norm": 72.27727508544922,
|
| 477 |
+
"learning_rate": 1.333968253968254e-05,
|
| 478 |
+
"loss": 0.2966,
|
| 479 |
+
"step": 1400
|
| 480 |
+
},
|
| 481 |
+
{
|
| 482 |
+
"epoch": 8.0,
|
| 483 |
+
"eval_accuracy": 0.8369272237196765,
|
| 484 |
+
"eval_loss": 0.8155926465988159,
|
| 485 |
+
"eval_macro_f1": 0.8373888166781356,
|
| 486 |
+
"eval_runtime": 0.8439,
|
| 487 |
+
"eval_samples_per_second": 879.23,
|
| 488 |
+
"eval_steps_per_second": 55.692,
|
| 489 |
+
"eval_weighted_f1": 0.8373416527059593,
|
| 490 |
+
"step": 1400
|
| 491 |
+
},
|
| 492 |
+
{
|
| 493 |
+
"epoch": 8.143061516452075,
|
| 494 |
+
"grad_norm": 55.8216667175293,
|
| 495 |
+
"learning_rate": 1.318095238095238e-05,
|
| 496 |
+
"loss": 0.3157,
|
| 497 |
+
"step": 1425
|
| 498 |
+
},
|
| 499 |
+
{
|
| 500 |
+
"epoch": 8.286123032904149,
|
| 501 |
+
"grad_norm": 197.83203125,
|
| 502 |
+
"learning_rate": 1.3022222222222223e-05,
|
| 503 |
+
"loss": 0.4096,
|
| 504 |
+
"step": 1450
|
| 505 |
+
},
|
| 506 |
+
{
|
| 507 |
+
"epoch": 8.429184549356224,
|
| 508 |
+
"grad_norm": 22.691734313964844,
|
| 509 |
+
"learning_rate": 1.2863492063492065e-05,
|
| 510 |
+
"loss": 0.267,
|
| 511 |
+
"step": 1475
|
| 512 |
+
},
|
| 513 |
+
{
|
| 514 |
+
"epoch": 8.572246065808297,
|
| 515 |
+
"grad_norm": 139.1099090576172,
|
| 516 |
+
"learning_rate": 1.2704761904761906e-05,
|
| 517 |
+
"loss": 0.4216,
|
| 518 |
+
"step": 1500
|
| 519 |
+
},
|
| 520 |
+
{
|
| 521 |
+
"epoch": 8.715307582260372,
|
| 522 |
+
"grad_norm": 15.543232917785645,
|
| 523 |
+
"learning_rate": 1.2546031746031746e-05,
|
| 524 |
+
"loss": 0.2027,
|
| 525 |
+
"step": 1525
|
| 526 |
+
},
|
| 527 |
+
{
|
| 528 |
+
"epoch": 8.858369098712446,
|
| 529 |
+
"grad_norm": 44.792259216308594,
|
| 530 |
+
"learning_rate": 1.2387301587301589e-05,
|
| 531 |
+
"loss": 0.2466,
|
| 532 |
+
"step": 1550
|
| 533 |
+
},
|
| 534 |
+
{
|
| 535 |
+
"epoch": 9.0,
|
| 536 |
+
"grad_norm": 70.58670806884766,
|
| 537 |
+
"learning_rate": 1.222857142857143e-05,
|
| 538 |
+
"loss": 0.2962,
|
| 539 |
+
"step": 1575
|
| 540 |
+
},
|
| 541 |
+
{
|
| 542 |
+
"epoch": 9.0,
|
| 543 |
+
"eval_accuracy": 0.839622641509434,
|
| 544 |
+
"eval_loss": 0.9224327802658081,
|
| 545 |
+
"eval_macro_f1": 0.8385877780941896,
|
| 546 |
+
"eval_runtime": 0.8504,
|
| 547 |
+
"eval_samples_per_second": 872.525,
|
| 548 |
+
"eval_steps_per_second": 55.268,
|
| 549 |
+
"eval_weighted_f1": 0.8389274395689376,
|
| 550 |
+
"step": 1575
|
| 551 |
+
}
|
| 552 |
+
],
|
| 553 |
+
"logging_steps": 25,
|
| 554 |
+
"max_steps": 3500,
|
| 555 |
+
"num_input_tokens_seen": 0,
|
| 556 |
+
"num_train_epochs": 20,
|
| 557 |
+
"save_steps": 500,
|
| 558 |
+
"stateful_callbacks": {
|
| 559 |
+
"EarlyStoppingCallback": {
|
| 560 |
+
"args": {
|
| 561 |
+
"early_stopping_patience": 3,
|
| 562 |
+
"early_stopping_threshold": 0.0
|
| 563 |
+
},
|
| 564 |
+
"attributes": {
|
| 565 |
+
"early_stopping_patience_counter": 0
|
| 566 |
+
}
|
| 567 |
+
},
|
| 568 |
+
"TrainerControl": {
|
| 569 |
+
"args": {
|
| 570 |
+
"should_epoch_stop": false,
|
| 571 |
+
"should_evaluate": false,
|
| 572 |
+
"should_log": false,
|
| 573 |
+
"should_save": true,
|
| 574 |
+
"should_training_stop": false
|
| 575 |
+
},
|
| 576 |
+
"attributes": {}
|
| 577 |
+
}
|
| 578 |
+
},
|
| 579 |
+
"total_flos": 0.0,
|
| 580 |
+
"train_batch_size": 8,
|
| 581 |
+
"trial_name": null,
|
| 582 |
+
"trial_params": null
|
| 583 |
+
}
|
ood-vifactcheck-bamibert-topic-domain-05-s43-20260827T195058Z-3858bb98/trainer/checkpoint-2100/special_tokens_map.json
ADDED
|
@@ -0,0 +1,51 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": {
|
| 3 |
+
"content": "<s>",
|
| 4 |
+
"lstrip": false,
|
| 5 |
+
"normalized": false,
|
| 6 |
+
"rstrip": false,
|
| 7 |
+
"single_word": false
|
| 8 |
+
},
|
| 9 |
+
"cls_token": {
|
| 10 |
+
"content": "<s>",
|
| 11 |
+
"lstrip": false,
|
| 12 |
+
"normalized": false,
|
| 13 |
+
"rstrip": false,
|
| 14 |
+
"single_word": false
|
| 15 |
+
},
|
| 16 |
+
"eos_token": {
|
| 17 |
+
"content": "</s>",
|
| 18 |
+
"lstrip": false,
|
| 19 |
+
"normalized": false,
|
| 20 |
+
"rstrip": false,
|
| 21 |
+
"single_word": false
|
| 22 |
+
},
|
| 23 |
+
"mask_token": {
|
| 24 |
+
"content": "<mask>",
|
| 25 |
+
"lstrip": false,
|
| 26 |
+
"normalized": false,
|
| 27 |
+
"rstrip": false,
|
| 28 |
+
"single_word": false
|
| 29 |
+
},
|
| 30 |
+
"pad_token": {
|
| 31 |
+
"content": "<pad>",
|
| 32 |
+
"lstrip": false,
|
| 33 |
+
"normalized": false,
|
| 34 |
+
"rstrip": false,
|
| 35 |
+
"single_word": false
|
| 36 |
+
},
|
| 37 |
+
"sep_token": {
|
| 38 |
+
"content": "</s>",
|
| 39 |
+
"lstrip": false,
|
| 40 |
+
"normalized": false,
|
| 41 |
+
"rstrip": false,
|
| 42 |
+
"single_word": false
|
| 43 |
+
},
|
| 44 |
+
"unk_token": {
|
| 45 |
+
"content": "<unk>",
|
| 46 |
+
"lstrip": false,
|
| 47 |
+
"normalized": false,
|
| 48 |
+
"rstrip": false,
|
| 49 |
+
"single_word": false
|
| 50 |
+
}
|
| 51 |
+
}
|
ood-vifactcheck-bamibert-topic-domain-05-s43-20260827T195058Z-3858bb98/trainer/checkpoint-2100/tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-bamibert-topic-domain-05-s43-20260827T195058Z-3858bb98/trainer/checkpoint-2100/tokenizer_config.json
ADDED
|
@@ -0,0 +1,55 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"added_tokens_decoder": {
|
| 3 |
+
"0": {
|
| 4 |
+
"content": "<s>",
|
| 5 |
+
"lstrip": false,
|
| 6 |
+
"normalized": false,
|
| 7 |
+
"rstrip": false,
|
| 8 |
+
"single_word": false,
|
| 9 |
+
"special": true
|
| 10 |
+
},
|
| 11 |
+
"1": {
|
| 12 |
+
"content": "<pad>",
|
| 13 |
+
"lstrip": false,
|
| 14 |
+
"normalized": false,
|
| 15 |
+
"rstrip": false,
|
| 16 |
+
"single_word": false,
|
| 17 |
+
"special": true
|
| 18 |
+
},
|
| 19 |
+
"2": {
|
| 20 |
+
"content": "</s>",
|
| 21 |
+
"lstrip": false,
|
| 22 |
+
"normalized": false,
|
| 23 |
+
"rstrip": false,
|
| 24 |
+
"single_word": false,
|
| 25 |
+
"special": true
|
| 26 |
+
},
|
| 27 |
+
"3": {
|
| 28 |
+
"content": "<unk>",
|
| 29 |
+
"lstrip": false,
|
| 30 |
+
"normalized": false,
|
| 31 |
+
"rstrip": false,
|
| 32 |
+
"single_word": false,
|
| 33 |
+
"special": true
|
| 34 |
+
},
|
| 35 |
+
"20480": {
|
| 36 |
+
"content": "<mask>",
|
| 37 |
+
"lstrip": false,
|
| 38 |
+
"normalized": false,
|
| 39 |
+
"rstrip": false,
|
| 40 |
+
"single_word": false,
|
| 41 |
+
"special": true
|
| 42 |
+
}
|
| 43 |
+
},
|
| 44 |
+
"bos_token": "<s>",
|
| 45 |
+
"clean_up_tokenization_spaces": false,
|
| 46 |
+
"cls_token": "<s>",
|
| 47 |
+
"eos_token": "</s>",
|
| 48 |
+
"extra_special_tokens": {},
|
| 49 |
+
"mask_token": "<mask>",
|
| 50 |
+
"model_max_length": 2048,
|
| 51 |
+
"pad_token": "<pad>",
|
| 52 |
+
"sep_token": "</s>",
|
| 53 |
+
"tokenizer_class": "XLMRobertaTokenizerFast",
|
| 54 |
+
"unk_token": "<unk>"
|
| 55 |
+
}
|
ood-vifactcheck-bamibert-topic-domain-05-s43-20260827T195058Z-3858bb98/trainer/checkpoint-2100/trainer_state.json
ADDED
|
@@ -0,0 +1,763 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 1575,
|
| 3 |
+
"best_metric": 0.8385877780941896,
|
| 4 |
+
"best_model_checkpoint": "/root/VNEncoder/results/checkpoints/ood-vifactcheck-bamibert-topic-domain-05-s43-20260827T195058Z-3858bb98/trainer/checkpoint-1575",
|
| 5 |
+
"epoch": 12.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 2100,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.1430615164520744,
|
| 14 |
+
"grad_norm": 14.410805702209473,
|
| 15 |
+
"learning_rate": 1.3714285714285717e-06,
|
| 16 |
+
"loss": 4.3997,
|
| 17 |
+
"step": 25
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.2861230329041488,
|
| 21 |
+
"grad_norm": 9.47131633758545,
|
| 22 |
+
"learning_rate": 2.8000000000000003e-06,
|
| 23 |
+
"loss": 4.4248,
|
| 24 |
+
"step": 50
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.4291845493562232,
|
| 28 |
+
"grad_norm": 6.229504585266113,
|
| 29 |
+
"learning_rate": 4.228571428571429e-06,
|
| 30 |
+
"loss": 4.4037,
|
| 31 |
+
"step": 75
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.5722460658082976,
|
| 35 |
+
"grad_norm": 9.012520790100098,
|
| 36 |
+
"learning_rate": 5.6571428571428576e-06,
|
| 37 |
+
"loss": 4.3548,
|
| 38 |
+
"step": 100
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.7153075822603719,
|
| 42 |
+
"grad_norm": 14.192707061767578,
|
| 43 |
+
"learning_rate": 7.085714285714286e-06,
|
| 44 |
+
"loss": 4.3811,
|
| 45 |
+
"step": 125
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.8583690987124464,
|
| 49 |
+
"grad_norm": 9.538655281066895,
|
| 50 |
+
"learning_rate": 8.514285714285715e-06,
|
| 51 |
+
"loss": 4.3266,
|
| 52 |
+
"step": 150
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"epoch": 1.0,
|
| 56 |
+
"grad_norm": 9.140969276428223,
|
| 57 |
+
"learning_rate": 9.942857142857145e-06,
|
| 58 |
+
"loss": 4.245,
|
| 59 |
+
"step": 175
|
| 60 |
+
},
|
| 61 |
+
{
|
| 62 |
+
"epoch": 1.0,
|
| 63 |
+
"eval_accuracy": 0.45956873315363883,
|
| 64 |
+
"eval_loss": 1.0388729572296143,
|
| 65 |
+
"eval_macro_f1": 0.39965935091985516,
|
| 66 |
+
"eval_runtime": 0.8432,
|
| 67 |
+
"eval_samples_per_second": 879.999,
|
| 68 |
+
"eval_steps_per_second": 55.741,
|
| 69 |
+
"eval_weighted_f1": 0.40267089211799173,
|
| 70 |
+
"step": 175
|
| 71 |
+
},
|
| 72 |
+
{
|
| 73 |
+
"epoch": 1.1430615164520743,
|
| 74 |
+
"grad_norm": 22.30449676513672,
|
| 75 |
+
"learning_rate": 1.1371428571428573e-05,
|
| 76 |
+
"loss": 4.169,
|
| 77 |
+
"step": 200
|
| 78 |
+
},
|
| 79 |
+
{
|
| 80 |
+
"epoch": 1.2861230329041489,
|
| 81 |
+
"grad_norm": 52.536251068115234,
|
| 82 |
+
"learning_rate": 1.2800000000000001e-05,
|
| 83 |
+
"loss": 3.7333,
|
| 84 |
+
"step": 225
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 1.4291845493562232,
|
| 88 |
+
"grad_norm": 34.337257385253906,
|
| 89 |
+
"learning_rate": 1.422857142857143e-05,
|
| 90 |
+
"loss": 3.4397,
|
| 91 |
+
"step": 250
|
| 92 |
+
},
|
| 93 |
+
{
|
| 94 |
+
"epoch": 1.5722460658082977,
|
| 95 |
+
"grad_norm": 43.14560317993164,
|
| 96 |
+
"learning_rate": 1.5657142857142856e-05,
|
| 97 |
+
"loss": 2.9114,
|
| 98 |
+
"step": 275
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 1.715307582260372,
|
| 102 |
+
"grad_norm": 64.7920913696289,
|
| 103 |
+
"learning_rate": 1.708571428571429e-05,
|
| 104 |
+
"loss": 2.5674,
|
| 105 |
+
"step": 300
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 1.8583690987124464,
|
| 109 |
+
"grad_norm": 44.81905746459961,
|
| 110 |
+
"learning_rate": 1.8514285714285716e-05,
|
| 111 |
+
"loss": 2.5042,
|
| 112 |
+
"step": 325
|
| 113 |
+
},
|
| 114 |
+
{
|
| 115 |
+
"epoch": 2.0,
|
| 116 |
+
"grad_norm": 31.639690399169922,
|
| 117 |
+
"learning_rate": 1.9942857142857142e-05,
|
| 118 |
+
"loss": 2.3798,
|
| 119 |
+
"step": 350
|
| 120 |
+
},
|
| 121 |
+
{
|
| 122 |
+
"epoch": 2.0,
|
| 123 |
+
"eval_accuracy": 0.7547169811320755,
|
| 124 |
+
"eval_loss": 0.5748988389968872,
|
| 125 |
+
"eval_macro_f1": 0.7549644564941046,
|
| 126 |
+
"eval_runtime": 0.8586,
|
| 127 |
+
"eval_samples_per_second": 864.159,
|
| 128 |
+
"eval_steps_per_second": 54.738,
|
| 129 |
+
"eval_weighted_f1": 0.7554270901473462,
|
| 130 |
+
"step": 350
|
| 131 |
+
},
|
| 132 |
+
{
|
| 133 |
+
"epoch": 2.1430615164520743,
|
| 134 |
+
"grad_norm": 58.28070831298828,
|
| 135 |
+
"learning_rate": 1.984761904761905e-05,
|
| 136 |
+
"loss": 1.9731,
|
| 137 |
+
"step": 375
|
| 138 |
+
},
|
| 139 |
+
{
|
| 140 |
+
"epoch": 2.2861230329041486,
|
| 141 |
+
"grad_norm": 36.115516662597656,
|
| 142 |
+
"learning_rate": 1.968888888888889e-05,
|
| 143 |
+
"loss": 1.8605,
|
| 144 |
+
"step": 400
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 2.429184549356223,
|
| 148 |
+
"grad_norm": 46.78152084350586,
|
| 149 |
+
"learning_rate": 1.953015873015873e-05,
|
| 150 |
+
"loss": 2.007,
|
| 151 |
+
"step": 425
|
| 152 |
+
},
|
| 153 |
+
{
|
| 154 |
+
"epoch": 2.5722460658082977,
|
| 155 |
+
"grad_norm": 55.40108871459961,
|
| 156 |
+
"learning_rate": 1.937142857142857e-05,
|
| 157 |
+
"loss": 1.9006,
|
| 158 |
+
"step": 450
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 2.715307582260372,
|
| 162 |
+
"grad_norm": 45.07542037963867,
|
| 163 |
+
"learning_rate": 1.9212698412698414e-05,
|
| 164 |
+
"loss": 1.6802,
|
| 165 |
+
"step": 475
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 2.8583690987124464,
|
| 169 |
+
"grad_norm": 27.396472930908203,
|
| 170 |
+
"learning_rate": 1.9053968253968257e-05,
|
| 171 |
+
"loss": 1.7654,
|
| 172 |
+
"step": 500
|
| 173 |
+
},
|
| 174 |
+
{
|
| 175 |
+
"epoch": 3.0,
|
| 176 |
+
"grad_norm": 31.59823989868164,
|
| 177 |
+
"learning_rate": 1.8895238095238096e-05,
|
| 178 |
+
"loss": 1.7634,
|
| 179 |
+
"step": 525
|
| 180 |
+
},
|
| 181 |
+
{
|
| 182 |
+
"epoch": 3.0,
|
| 183 |
+
"eval_accuracy": 0.807277628032345,
|
| 184 |
+
"eval_loss": 0.4905075430870056,
|
| 185 |
+
"eval_macro_f1": 0.8061139250249086,
|
| 186 |
+
"eval_runtime": 0.8535,
|
| 187 |
+
"eval_samples_per_second": 869.407,
|
| 188 |
+
"eval_steps_per_second": 55.07,
|
| 189 |
+
"eval_weighted_f1": 0.8064427025314329,
|
| 190 |
+
"step": 525
|
| 191 |
+
},
|
| 192 |
+
{
|
| 193 |
+
"epoch": 3.1430615164520743,
|
| 194 |
+
"grad_norm": 67.2063217163086,
|
| 195 |
+
"learning_rate": 1.8736507936507936e-05,
|
| 196 |
+
"loss": 0.996,
|
| 197 |
+
"step": 550
|
| 198 |
+
},
|
| 199 |
+
{
|
| 200 |
+
"epoch": 3.2861230329041486,
|
| 201 |
+
"grad_norm": 83.634521484375,
|
| 202 |
+
"learning_rate": 1.857777777777778e-05,
|
| 203 |
+
"loss": 1.2307,
|
| 204 |
+
"step": 575
|
| 205 |
+
},
|
| 206 |
+
{
|
| 207 |
+
"epoch": 3.429184549356223,
|
| 208 |
+
"grad_norm": 32.015079498291016,
|
| 209 |
+
"learning_rate": 1.8419047619047622e-05,
|
| 210 |
+
"loss": 1.1427,
|
| 211 |
+
"step": 600
|
| 212 |
+
},
|
| 213 |
+
{
|
| 214 |
+
"epoch": 3.5722460658082977,
|
| 215 |
+
"grad_norm": 41.71841812133789,
|
| 216 |
+
"learning_rate": 1.8260317460317462e-05,
|
| 217 |
+
"loss": 1.3008,
|
| 218 |
+
"step": 625
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 3.715307582260372,
|
| 222 |
+
"grad_norm": 59.132320404052734,
|
| 223 |
+
"learning_rate": 1.81015873015873e-05,
|
| 224 |
+
"loss": 1.2368,
|
| 225 |
+
"step": 650
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"epoch": 3.8583690987124464,
|
| 229 |
+
"grad_norm": 38.083065032958984,
|
| 230 |
+
"learning_rate": 1.7942857142857144e-05,
|
| 231 |
+
"loss": 1.2518,
|
| 232 |
+
"step": 675
|
| 233 |
+
},
|
| 234 |
+
{
|
| 235 |
+
"epoch": 4.0,
|
| 236 |
+
"grad_norm": 42.52290725708008,
|
| 237 |
+
"learning_rate": 1.7784126984126987e-05,
|
| 238 |
+
"loss": 1.2294,
|
| 239 |
+
"step": 700
|
| 240 |
+
},
|
| 241 |
+
{
|
| 242 |
+
"epoch": 4.0,
|
| 243 |
+
"eval_accuracy": 0.8032345013477089,
|
| 244 |
+
"eval_loss": 0.5453498363494873,
|
| 245 |
+
"eval_macro_f1": 0.8010587518679261,
|
| 246 |
+
"eval_runtime": 0.8577,
|
| 247 |
+
"eval_samples_per_second": 865.116,
|
| 248 |
+
"eval_steps_per_second": 54.798,
|
| 249 |
+
"eval_weighted_f1": 0.8013903122767734,
|
| 250 |
+
"step": 700
|
| 251 |
+
},
|
| 252 |
+
{
|
| 253 |
+
"epoch": 4.143061516452074,
|
| 254 |
+
"grad_norm": 62.421905517578125,
|
| 255 |
+
"learning_rate": 1.7625396825396827e-05,
|
| 256 |
+
"loss": 0.9209,
|
| 257 |
+
"step": 725
|
| 258 |
+
},
|
| 259 |
+
{
|
| 260 |
+
"epoch": 4.286123032904149,
|
| 261 |
+
"grad_norm": 30.84873390197754,
|
| 262 |
+
"learning_rate": 1.7466666666666667e-05,
|
| 263 |
+
"loss": 0.6917,
|
| 264 |
+
"step": 750
|
| 265 |
+
},
|
| 266 |
+
{
|
| 267 |
+
"epoch": 4.429184549356223,
|
| 268 |
+
"grad_norm": 37.597694396972656,
|
| 269 |
+
"learning_rate": 1.730793650793651e-05,
|
| 270 |
+
"loss": 0.753,
|
| 271 |
+
"step": 775
|
| 272 |
+
},
|
| 273 |
+
{
|
| 274 |
+
"epoch": 4.572246065808297,
|
| 275 |
+
"grad_norm": 75.85980987548828,
|
| 276 |
+
"learning_rate": 1.7149206349206353e-05,
|
| 277 |
+
"loss": 0.7342,
|
| 278 |
+
"step": 800
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 4.715307582260372,
|
| 282 |
+
"grad_norm": 53.75027084350586,
|
| 283 |
+
"learning_rate": 1.6990476190476192e-05,
|
| 284 |
+
"loss": 0.6969,
|
| 285 |
+
"step": 825
|
| 286 |
+
},
|
| 287 |
+
{
|
| 288 |
+
"epoch": 4.858369098712446,
|
| 289 |
+
"grad_norm": 7.4808173179626465,
|
| 290 |
+
"learning_rate": 1.6831746031746032e-05,
|
| 291 |
+
"loss": 0.7179,
|
| 292 |
+
"step": 850
|
| 293 |
+
},
|
| 294 |
+
{
|
| 295 |
+
"epoch": 5.0,
|
| 296 |
+
"grad_norm": 34.891693115234375,
|
| 297 |
+
"learning_rate": 1.6673015873015875e-05,
|
| 298 |
+
"loss": 0.9722,
|
| 299 |
+
"step": 875
|
| 300 |
+
},
|
| 301 |
+
{
|
| 302 |
+
"epoch": 5.0,
|
| 303 |
+
"eval_accuracy": 0.8207547169811321,
|
| 304 |
+
"eval_loss": 0.5585728287696838,
|
| 305 |
+
"eval_macro_f1": 0.8175550071748489,
|
| 306 |
+
"eval_runtime": 0.8441,
|
| 307 |
+
"eval_samples_per_second": 879.007,
|
| 308 |
+
"eval_steps_per_second": 55.678,
|
| 309 |
+
"eval_weighted_f1": 0.8176335983763029,
|
| 310 |
+
"step": 875
|
| 311 |
+
},
|
| 312 |
+
{
|
| 313 |
+
"epoch": 5.143061516452074,
|
| 314 |
+
"grad_norm": 37.72514724731445,
|
| 315 |
+
"learning_rate": 1.6514285714285714e-05,
|
| 316 |
+
"loss": 0.4463,
|
| 317 |
+
"step": 900
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 5.286123032904149,
|
| 321 |
+
"grad_norm": 67.63714599609375,
|
| 322 |
+
"learning_rate": 1.6355555555555557e-05,
|
| 323 |
+
"loss": 0.5709,
|
| 324 |
+
"step": 925
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 5.429184549356223,
|
| 328 |
+
"grad_norm": 20.562917709350586,
|
| 329 |
+
"learning_rate": 1.6196825396825397e-05,
|
| 330 |
+
"loss": 0.6067,
|
| 331 |
+
"step": 950
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 5.572246065808297,
|
| 335 |
+
"grad_norm": 64.2474365234375,
|
| 336 |
+
"learning_rate": 1.603809523809524e-05,
|
| 337 |
+
"loss": 0.4798,
|
| 338 |
+
"step": 975
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 5.715307582260372,
|
| 342 |
+
"grad_norm": 73.35626983642578,
|
| 343 |
+
"learning_rate": 1.587936507936508e-05,
|
| 344 |
+
"loss": 0.6928,
|
| 345 |
+
"step": 1000
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 5.858369098712446,
|
| 349 |
+
"grad_norm": 20.787761688232422,
|
| 350 |
+
"learning_rate": 1.5720634920634923e-05,
|
| 351 |
+
"loss": 0.5191,
|
| 352 |
+
"step": 1025
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 6.0,
|
| 356 |
+
"grad_norm": 39.936492919921875,
|
| 357 |
+
"learning_rate": 1.5561904761904762e-05,
|
| 358 |
+
"loss": 0.6271,
|
| 359 |
+
"step": 1050
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 6.0,
|
| 363 |
+
"eval_accuracy": 0.8261455525606469,
|
| 364 |
+
"eval_loss": 0.6241559386253357,
|
| 365 |
+
"eval_macro_f1": 0.825354877680459,
|
| 366 |
+
"eval_runtime": 0.8496,
|
| 367 |
+
"eval_samples_per_second": 873.399,
|
| 368 |
+
"eval_steps_per_second": 55.323,
|
| 369 |
+
"eval_weighted_f1": 0.8256656525947569,
|
| 370 |
+
"step": 1050
|
| 371 |
+
},
|
| 372 |
+
{
|
| 373 |
+
"epoch": 6.143061516452074,
|
| 374 |
+
"grad_norm": 114.56380462646484,
|
| 375 |
+
"learning_rate": 1.5403174603174605e-05,
|
| 376 |
+
"loss": 0.3623,
|
| 377 |
+
"step": 1075
|
| 378 |
+
},
|
| 379 |
+
{
|
| 380 |
+
"epoch": 6.286123032904149,
|
| 381 |
+
"grad_norm": 77.52657318115234,
|
| 382 |
+
"learning_rate": 1.5244444444444447e-05,
|
| 383 |
+
"loss": 0.391,
|
| 384 |
+
"step": 1100
|
| 385 |
+
},
|
| 386 |
+
{
|
| 387 |
+
"epoch": 6.429184549356223,
|
| 388 |
+
"grad_norm": 39.765647888183594,
|
| 389 |
+
"learning_rate": 1.5085714285714288e-05,
|
| 390 |
+
"loss": 0.5178,
|
| 391 |
+
"step": 1125
|
| 392 |
+
},
|
| 393 |
+
{
|
| 394 |
+
"epoch": 6.572246065808297,
|
| 395 |
+
"grad_norm": 6.1532087326049805,
|
| 396 |
+
"learning_rate": 1.4926984126984128e-05,
|
| 397 |
+
"loss": 0.5117,
|
| 398 |
+
"step": 1150
|
| 399 |
+
},
|
| 400 |
+
{
|
| 401 |
+
"epoch": 6.715307582260372,
|
| 402 |
+
"grad_norm": 73.53240966796875,
|
| 403 |
+
"learning_rate": 1.4768253968253969e-05,
|
| 404 |
+
"loss": 0.4157,
|
| 405 |
+
"step": 1175
|
| 406 |
+
},
|
| 407 |
+
{
|
| 408 |
+
"epoch": 6.858369098712446,
|
| 409 |
+
"grad_norm": 195.1527099609375,
|
| 410 |
+
"learning_rate": 1.4609523809523812e-05,
|
| 411 |
+
"loss": 0.5763,
|
| 412 |
+
"step": 1200
|
| 413 |
+
},
|
| 414 |
+
{
|
| 415 |
+
"epoch": 7.0,
|
| 416 |
+
"grad_norm": 81.26548767089844,
|
| 417 |
+
"learning_rate": 1.4450793650793653e-05,
|
| 418 |
+
"loss": 0.4112,
|
| 419 |
+
"step": 1225
|
| 420 |
+
},
|
| 421 |
+
{
|
| 422 |
+
"epoch": 7.0,
|
| 423 |
+
"eval_accuracy": 0.8018867924528302,
|
| 424 |
+
"eval_loss": 0.9285014867782593,
|
| 425 |
+
"eval_macro_f1": 0.8004517524849147,
|
| 426 |
+
"eval_runtime": 0.8406,
|
| 427 |
+
"eval_samples_per_second": 882.746,
|
| 428 |
+
"eval_steps_per_second": 55.915,
|
| 429 |
+
"eval_weighted_f1": 0.8001643468723298,
|
| 430 |
+
"step": 1225
|
| 431 |
+
},
|
| 432 |
+
{
|
| 433 |
+
"epoch": 7.143061516452074,
|
| 434 |
+
"grad_norm": 0.7759934663772583,
|
| 435 |
+
"learning_rate": 1.4292063492063493e-05,
|
| 436 |
+
"loss": 0.2987,
|
| 437 |
+
"step": 1250
|
| 438 |
+
},
|
| 439 |
+
{
|
| 440 |
+
"epoch": 7.286123032904149,
|
| 441 |
+
"grad_norm": 17.462034225463867,
|
| 442 |
+
"learning_rate": 1.4133333333333334e-05,
|
| 443 |
+
"loss": 0.3494,
|
| 444 |
+
"step": 1275
|
| 445 |
+
},
|
| 446 |
+
{
|
| 447 |
+
"epoch": 7.429184549356223,
|
| 448 |
+
"grad_norm": 0.44466474652290344,
|
| 449 |
+
"learning_rate": 1.3974603174603177e-05,
|
| 450 |
+
"loss": 0.3154,
|
| 451 |
+
"step": 1300
|
| 452 |
+
},
|
| 453 |
+
{
|
| 454 |
+
"epoch": 7.572246065808297,
|
| 455 |
+
"grad_norm": 26.82601547241211,
|
| 456 |
+
"learning_rate": 1.3815873015873018e-05,
|
| 457 |
+
"loss": 0.2817,
|
| 458 |
+
"step": 1325
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 7.715307582260372,
|
| 462 |
+
"grad_norm": 23.987606048583984,
|
| 463 |
+
"learning_rate": 1.3657142857142858e-05,
|
| 464 |
+
"loss": 0.3081,
|
| 465 |
+
"step": 1350
|
| 466 |
+
},
|
| 467 |
+
{
|
| 468 |
+
"epoch": 7.858369098712446,
|
| 469 |
+
"grad_norm": 30.050395965576172,
|
| 470 |
+
"learning_rate": 1.34984126984127e-05,
|
| 471 |
+
"loss": 0.6405,
|
| 472 |
+
"step": 1375
|
| 473 |
+
},
|
| 474 |
+
{
|
| 475 |
+
"epoch": 8.0,
|
| 476 |
+
"grad_norm": 72.27727508544922,
|
| 477 |
+
"learning_rate": 1.333968253968254e-05,
|
| 478 |
+
"loss": 0.2966,
|
| 479 |
+
"step": 1400
|
| 480 |
+
},
|
| 481 |
+
{
|
| 482 |
+
"epoch": 8.0,
|
| 483 |
+
"eval_accuracy": 0.8369272237196765,
|
| 484 |
+
"eval_loss": 0.8155926465988159,
|
| 485 |
+
"eval_macro_f1": 0.8373888166781356,
|
| 486 |
+
"eval_runtime": 0.8439,
|
| 487 |
+
"eval_samples_per_second": 879.23,
|
| 488 |
+
"eval_steps_per_second": 55.692,
|
| 489 |
+
"eval_weighted_f1": 0.8373416527059593,
|
| 490 |
+
"step": 1400
|
| 491 |
+
},
|
| 492 |
+
{
|
| 493 |
+
"epoch": 8.143061516452075,
|
| 494 |
+
"grad_norm": 55.8216667175293,
|
| 495 |
+
"learning_rate": 1.318095238095238e-05,
|
| 496 |
+
"loss": 0.3157,
|
| 497 |
+
"step": 1425
|
| 498 |
+
},
|
| 499 |
+
{
|
| 500 |
+
"epoch": 8.286123032904149,
|
| 501 |
+
"grad_norm": 197.83203125,
|
| 502 |
+
"learning_rate": 1.3022222222222223e-05,
|
| 503 |
+
"loss": 0.4096,
|
| 504 |
+
"step": 1450
|
| 505 |
+
},
|
| 506 |
+
{
|
| 507 |
+
"epoch": 8.429184549356224,
|
| 508 |
+
"grad_norm": 22.691734313964844,
|
| 509 |
+
"learning_rate": 1.2863492063492065e-05,
|
| 510 |
+
"loss": 0.267,
|
| 511 |
+
"step": 1475
|
| 512 |
+
},
|
| 513 |
+
{
|
| 514 |
+
"epoch": 8.572246065808297,
|
| 515 |
+
"grad_norm": 139.1099090576172,
|
| 516 |
+
"learning_rate": 1.2704761904761906e-05,
|
| 517 |
+
"loss": 0.4216,
|
| 518 |
+
"step": 1500
|
| 519 |
+
},
|
| 520 |
+
{
|
| 521 |
+
"epoch": 8.715307582260372,
|
| 522 |
+
"grad_norm": 15.543232917785645,
|
| 523 |
+
"learning_rate": 1.2546031746031746e-05,
|
| 524 |
+
"loss": 0.2027,
|
| 525 |
+
"step": 1525
|
| 526 |
+
},
|
| 527 |
+
{
|
| 528 |
+
"epoch": 8.858369098712446,
|
| 529 |
+
"grad_norm": 44.792259216308594,
|
| 530 |
+
"learning_rate": 1.2387301587301589e-05,
|
| 531 |
+
"loss": 0.2466,
|
| 532 |
+
"step": 1550
|
| 533 |
+
},
|
| 534 |
+
{
|
| 535 |
+
"epoch": 9.0,
|
| 536 |
+
"grad_norm": 70.58670806884766,
|
| 537 |
+
"learning_rate": 1.222857142857143e-05,
|
| 538 |
+
"loss": 0.2962,
|
| 539 |
+
"step": 1575
|
| 540 |
+
},
|
| 541 |
+
{
|
| 542 |
+
"epoch": 9.0,
|
| 543 |
+
"eval_accuracy": 0.839622641509434,
|
| 544 |
+
"eval_loss": 0.9224327802658081,
|
| 545 |
+
"eval_macro_f1": 0.8385877780941896,
|
| 546 |
+
"eval_runtime": 0.8504,
|
| 547 |
+
"eval_samples_per_second": 872.525,
|
| 548 |
+
"eval_steps_per_second": 55.268,
|
| 549 |
+
"eval_weighted_f1": 0.8389274395689376,
|
| 550 |
+
"step": 1575
|
| 551 |
+
},
|
| 552 |
+
{
|
| 553 |
+
"epoch": 9.143061516452075,
|
| 554 |
+
"grad_norm": 1.0777536630630493,
|
| 555 |
+
"learning_rate": 1.2069841269841271e-05,
|
| 556 |
+
"loss": 0.09,
|
| 557 |
+
"step": 1600
|
| 558 |
+
},
|
| 559 |
+
{
|
| 560 |
+
"epoch": 9.286123032904149,
|
| 561 |
+
"grad_norm": 0.11101091653108597,
|
| 562 |
+
"learning_rate": 1.191111111111111e-05,
|
| 563 |
+
"loss": 0.0521,
|
| 564 |
+
"step": 1625
|
| 565 |
+
},
|
| 566 |
+
{
|
| 567 |
+
"epoch": 9.429184549356224,
|
| 568 |
+
"grad_norm": 138.81204223632812,
|
| 569 |
+
"learning_rate": 1.1752380952380952e-05,
|
| 570 |
+
"loss": 0.1425,
|
| 571 |
+
"step": 1650
|
| 572 |
+
},
|
| 573 |
+
{
|
| 574 |
+
"epoch": 9.572246065808297,
|
| 575 |
+
"grad_norm": 0.24475054442882538,
|
| 576 |
+
"learning_rate": 1.1593650793650795e-05,
|
| 577 |
+
"loss": 0.2814,
|
| 578 |
+
"step": 1675
|
| 579 |
+
},
|
| 580 |
+
{
|
| 581 |
+
"epoch": 9.715307582260372,
|
| 582 |
+
"grad_norm": 0.20813189446926117,
|
| 583 |
+
"learning_rate": 1.1434920634920636e-05,
|
| 584 |
+
"loss": 0.2128,
|
| 585 |
+
"step": 1700
|
| 586 |
+
},
|
| 587 |
+
{
|
| 588 |
+
"epoch": 9.858369098712446,
|
| 589 |
+
"grad_norm": 216.3400115966797,
|
| 590 |
+
"learning_rate": 1.1276190476190476e-05,
|
| 591 |
+
"loss": 0.2524,
|
| 592 |
+
"step": 1725
|
| 593 |
+
},
|
| 594 |
+
{
|
| 595 |
+
"epoch": 10.0,
|
| 596 |
+
"grad_norm": 4.022839069366455,
|
| 597 |
+
"learning_rate": 1.1117460317460317e-05,
|
| 598 |
+
"loss": 0.2348,
|
| 599 |
+
"step": 1750
|
| 600 |
+
},
|
| 601 |
+
{
|
| 602 |
+
"epoch": 10.0,
|
| 603 |
+
"eval_accuracy": 0.8274932614555256,
|
| 604 |
+
"eval_loss": 1.0882335901260376,
|
| 605 |
+
"eval_macro_f1": 0.8267393286654454,
|
| 606 |
+
"eval_runtime": 0.8538,
|
| 607 |
+
"eval_samples_per_second": 869.069,
|
| 608 |
+
"eval_steps_per_second": 55.049,
|
| 609 |
+
"eval_weighted_f1": 0.8269063068076206,
|
| 610 |
+
"step": 1750
|
| 611 |
+
},
|
| 612 |
+
{
|
| 613 |
+
"epoch": 10.143061516452075,
|
| 614 |
+
"grad_norm": 286.20477294921875,
|
| 615 |
+
"learning_rate": 1.095873015873016e-05,
|
| 616 |
+
"loss": 0.2291,
|
| 617 |
+
"step": 1775
|
| 618 |
+
},
|
| 619 |
+
{
|
| 620 |
+
"epoch": 10.286123032904149,
|
| 621 |
+
"grad_norm": 1.6473817825317383,
|
| 622 |
+
"learning_rate": 1.0800000000000002e-05,
|
| 623 |
+
"loss": 0.1185,
|
| 624 |
+
"step": 1800
|
| 625 |
+
},
|
| 626 |
+
{
|
| 627 |
+
"epoch": 10.429184549356224,
|
| 628 |
+
"grad_norm": 497.3121337890625,
|
| 629 |
+
"learning_rate": 1.0641269841269841e-05,
|
| 630 |
+
"loss": 0.2163,
|
| 631 |
+
"step": 1825
|
| 632 |
+
},
|
| 633 |
+
{
|
| 634 |
+
"epoch": 10.572246065808297,
|
| 635 |
+
"grad_norm": 0.2167133092880249,
|
| 636 |
+
"learning_rate": 1.0482539682539683e-05,
|
| 637 |
+
"loss": 0.1819,
|
| 638 |
+
"step": 1850
|
| 639 |
+
},
|
| 640 |
+
{
|
| 641 |
+
"epoch": 10.715307582260372,
|
| 642 |
+
"grad_norm": 0.1071733683347702,
|
| 643 |
+
"learning_rate": 1.0323809523809526e-05,
|
| 644 |
+
"loss": 0.1972,
|
| 645 |
+
"step": 1875
|
| 646 |
+
},
|
| 647 |
+
{
|
| 648 |
+
"epoch": 10.858369098712446,
|
| 649 |
+
"grad_norm": 0.06342184543609619,
|
| 650 |
+
"learning_rate": 1.0165079365079367e-05,
|
| 651 |
+
"loss": 0.0698,
|
| 652 |
+
"step": 1900
|
| 653 |
+
},
|
| 654 |
+
{
|
| 655 |
+
"epoch": 11.0,
|
| 656 |
+
"grad_norm": 97.03350830078125,
|
| 657 |
+
"learning_rate": 1.0006349206349206e-05,
|
| 658 |
+
"loss": 0.1744,
|
| 659 |
+
"step": 1925
|
| 660 |
+
},
|
| 661 |
+
{
|
| 662 |
+
"epoch": 11.0,
|
| 663 |
+
"eval_accuracy": 0.8194070080862533,
|
| 664 |
+
"eval_loss": 1.1149522066116333,
|
| 665 |
+
"eval_macro_f1": 0.8181709600755375,
|
| 666 |
+
"eval_runtime": 0.8435,
|
| 667 |
+
"eval_samples_per_second": 879.676,
|
| 668 |
+
"eval_steps_per_second": 55.721,
|
| 669 |
+
"eval_weighted_f1": 0.818140192014889,
|
| 670 |
+
"step": 1925
|
| 671 |
+
},
|
| 672 |
+
{
|
| 673 |
+
"epoch": 11.143061516452075,
|
| 674 |
+
"grad_norm": 0.08893871307373047,
|
| 675 |
+
"learning_rate": 9.847619047619048e-06,
|
| 676 |
+
"loss": 0.0298,
|
| 677 |
+
"step": 1950
|
| 678 |
+
},
|
| 679 |
+
{
|
| 680 |
+
"epoch": 11.286123032904149,
|
| 681 |
+
"grad_norm": 247.6279296875,
|
| 682 |
+
"learning_rate": 9.688888888888889e-06,
|
| 683 |
+
"loss": 0.1249,
|
| 684 |
+
"step": 1975
|
| 685 |
+
},
|
| 686 |
+
{
|
| 687 |
+
"epoch": 11.429184549356224,
|
| 688 |
+
"grad_norm": 0.0632932037115097,
|
| 689 |
+
"learning_rate": 9.53015873015873e-06,
|
| 690 |
+
"loss": 0.0954,
|
| 691 |
+
"step": 2000
|
| 692 |
+
},
|
| 693 |
+
{
|
| 694 |
+
"epoch": 11.572246065808297,
|
| 695 |
+
"grad_norm": 83.19509887695312,
|
| 696 |
+
"learning_rate": 9.371428571428572e-06,
|
| 697 |
+
"loss": 0.1412,
|
| 698 |
+
"step": 2025
|
| 699 |
+
},
|
| 700 |
+
{
|
| 701 |
+
"epoch": 11.715307582260372,
|
| 702 |
+
"grad_norm": 0.059804923832416534,
|
| 703 |
+
"learning_rate": 9.212698412698413e-06,
|
| 704 |
+
"loss": 0.0819,
|
| 705 |
+
"step": 2050
|
| 706 |
+
},
|
| 707 |
+
{
|
| 708 |
+
"epoch": 11.858369098712446,
|
| 709 |
+
"grad_norm": 0.2570134103298187,
|
| 710 |
+
"learning_rate": 9.053968253968254e-06,
|
| 711 |
+
"loss": 0.0802,
|
| 712 |
+
"step": 2075
|
| 713 |
+
},
|
| 714 |
+
{
|
| 715 |
+
"epoch": 12.0,
|
| 716 |
+
"grad_norm": 0.20984600484371185,
|
| 717 |
+
"learning_rate": 8.895238095238096e-06,
|
| 718 |
+
"loss": 0.1918,
|
| 719 |
+
"step": 2100
|
| 720 |
+
},
|
| 721 |
+
{
|
| 722 |
+
"epoch": 12.0,
|
| 723 |
+
"eval_accuracy": 0.8342318059299192,
|
| 724 |
+
"eval_loss": 1.0476006269454956,
|
| 725 |
+
"eval_macro_f1": 0.8343047283909296,
|
| 726 |
+
"eval_runtime": 0.8674,
|
| 727 |
+
"eval_samples_per_second": 855.425,
|
| 728 |
+
"eval_steps_per_second": 54.185,
|
| 729 |
+
"eval_weighted_f1": 0.8344074536884716,
|
| 730 |
+
"step": 2100
|
| 731 |
+
}
|
| 732 |
+
],
|
| 733 |
+
"logging_steps": 25,
|
| 734 |
+
"max_steps": 3500,
|
| 735 |
+
"num_input_tokens_seen": 0,
|
| 736 |
+
"num_train_epochs": 20,
|
| 737 |
+
"save_steps": 500,
|
| 738 |
+
"stateful_callbacks": {
|
| 739 |
+
"EarlyStoppingCallback": {
|
| 740 |
+
"args": {
|
| 741 |
+
"early_stopping_patience": 3,
|
| 742 |
+
"early_stopping_threshold": 0.0
|
| 743 |
+
},
|
| 744 |
+
"attributes": {
|
| 745 |
+
"early_stopping_patience_counter": 3
|
| 746 |
+
}
|
| 747 |
+
},
|
| 748 |
+
"TrainerControl": {
|
| 749 |
+
"args": {
|
| 750 |
+
"should_epoch_stop": false,
|
| 751 |
+
"should_evaluate": false,
|
| 752 |
+
"should_log": false,
|
| 753 |
+
"should_save": true,
|
| 754 |
+
"should_training_stop": true
|
| 755 |
+
},
|
| 756 |
+
"attributes": {}
|
| 757 |
+
}
|
| 758 |
+
},
|
| 759 |
+
"total_flos": 0.0,
|
| 760 |
+
"train_batch_size": 8,
|
| 761 |
+
"trial_name": null,
|
| 762 |
+
"trial_params": null
|
| 763 |
+
}
|
ood-vifactcheck-bamibert-topic-domain-06-s43-20260827T195550Z-3858bb98/trainer/checkpoint-1104/special_tokens_map.json
ADDED
|
@@ -0,0 +1,51 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": {
|
| 3 |
+
"content": "<s>",
|
| 4 |
+
"lstrip": false,
|
| 5 |
+
"normalized": false,
|
| 6 |
+
"rstrip": false,
|
| 7 |
+
"single_word": false
|
| 8 |
+
},
|
| 9 |
+
"cls_token": {
|
| 10 |
+
"content": "<s>",
|
| 11 |
+
"lstrip": false,
|
| 12 |
+
"normalized": false,
|
| 13 |
+
"rstrip": false,
|
| 14 |
+
"single_word": false
|
| 15 |
+
},
|
| 16 |
+
"eos_token": {
|
| 17 |
+
"content": "</s>",
|
| 18 |
+
"lstrip": false,
|
| 19 |
+
"normalized": false,
|
| 20 |
+
"rstrip": false,
|
| 21 |
+
"single_word": false
|
| 22 |
+
},
|
| 23 |
+
"mask_token": {
|
| 24 |
+
"content": "<mask>",
|
| 25 |
+
"lstrip": false,
|
| 26 |
+
"normalized": false,
|
| 27 |
+
"rstrip": false,
|
| 28 |
+
"single_word": false
|
| 29 |
+
},
|
| 30 |
+
"pad_token": {
|
| 31 |
+
"content": "<pad>",
|
| 32 |
+
"lstrip": false,
|
| 33 |
+
"normalized": false,
|
| 34 |
+
"rstrip": false,
|
| 35 |
+
"single_word": false
|
| 36 |
+
},
|
| 37 |
+
"sep_token": {
|
| 38 |
+
"content": "</s>",
|
| 39 |
+
"lstrip": false,
|
| 40 |
+
"normalized": false,
|
| 41 |
+
"rstrip": false,
|
| 42 |
+
"single_word": false
|
| 43 |
+
},
|
| 44 |
+
"unk_token": {
|
| 45 |
+
"content": "<unk>",
|
| 46 |
+
"lstrip": false,
|
| 47 |
+
"normalized": false,
|
| 48 |
+
"rstrip": false,
|
| 49 |
+
"single_word": false
|
| 50 |
+
}
|
| 51 |
+
}
|
ood-vifactcheck-bamibert-topic-domain-06-s43-20260827T195550Z-3858bb98/trainer/checkpoint-1104/tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-bamibert-topic-domain-06-s43-20260827T195550Z-3858bb98/trainer/checkpoint-1104/tokenizer_config.json
ADDED
|
@@ -0,0 +1,55 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"added_tokens_decoder": {
|
| 3 |
+
"0": {
|
| 4 |
+
"content": "<s>",
|
| 5 |
+
"lstrip": false,
|
| 6 |
+
"normalized": false,
|
| 7 |
+
"rstrip": false,
|
| 8 |
+
"single_word": false,
|
| 9 |
+
"special": true
|
| 10 |
+
},
|
| 11 |
+
"1": {
|
| 12 |
+
"content": "<pad>",
|
| 13 |
+
"lstrip": false,
|
| 14 |
+
"normalized": false,
|
| 15 |
+
"rstrip": false,
|
| 16 |
+
"single_word": false,
|
| 17 |
+
"special": true
|
| 18 |
+
},
|
| 19 |
+
"2": {
|
| 20 |
+
"content": "</s>",
|
| 21 |
+
"lstrip": false,
|
| 22 |
+
"normalized": false,
|
| 23 |
+
"rstrip": false,
|
| 24 |
+
"single_word": false,
|
| 25 |
+
"special": true
|
| 26 |
+
},
|
| 27 |
+
"3": {
|
| 28 |
+
"content": "<unk>",
|
| 29 |
+
"lstrip": false,
|
| 30 |
+
"normalized": false,
|
| 31 |
+
"rstrip": false,
|
| 32 |
+
"single_word": false,
|
| 33 |
+
"special": true
|
| 34 |
+
},
|
| 35 |
+
"20480": {
|
| 36 |
+
"content": "<mask>",
|
| 37 |
+
"lstrip": false,
|
| 38 |
+
"normalized": false,
|
| 39 |
+
"rstrip": false,
|
| 40 |
+
"single_word": false,
|
| 41 |
+
"special": true
|
| 42 |
+
}
|
| 43 |
+
},
|
| 44 |
+
"bos_token": "<s>",
|
| 45 |
+
"clean_up_tokenization_spaces": false,
|
| 46 |
+
"cls_token": "<s>",
|
| 47 |
+
"eos_token": "</s>",
|
| 48 |
+
"extra_special_tokens": {},
|
| 49 |
+
"mask_token": "<mask>",
|
| 50 |
+
"model_max_length": 2048,
|
| 51 |
+
"pad_token": "<pad>",
|
| 52 |
+
"sep_token": "</s>",
|
| 53 |
+
"tokenizer_class": "XLMRobertaTokenizerFast",
|
| 54 |
+
"unk_token": "<unk>"
|
| 55 |
+
}
|
ood-vifactcheck-bamibert-topic-domain-06-s43-20260827T195550Z-3858bb98/trainer/checkpoint-1104/trainer_state.json
ADDED
|
@@ -0,0 +1,417 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 552,
|
| 3 |
+
"best_metric": 0.8170391895457513,
|
| 4 |
+
"best_model_checkpoint": "/root/VNEncoder/results/checkpoints/ood-vifactcheck-bamibert-topic-domain-06-s43-20260827T195550Z-3858bb98/trainer/checkpoint-552",
|
| 5 |
+
"epoch": 6.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 1104,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.1364256480218281,
|
| 14 |
+
"grad_norm": 13.91116714477539,
|
| 15 |
+
"learning_rate": 1.3043478260869566e-06,
|
| 16 |
+
"loss": 4.4188,
|
| 17 |
+
"step": 25
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.2728512960436562,
|
| 21 |
+
"grad_norm": 13.288341522216797,
|
| 22 |
+
"learning_rate": 2.6630434782608698e-06,
|
| 23 |
+
"loss": 4.3869,
|
| 24 |
+
"step": 50
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.4092769440654843,
|
| 28 |
+
"grad_norm": 10.405351638793945,
|
| 29 |
+
"learning_rate": 4.021739130434783e-06,
|
| 30 |
+
"loss": 4.4184,
|
| 31 |
+
"step": 75
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.5457025920873124,
|
| 35 |
+
"grad_norm": 13.788087844848633,
|
| 36 |
+
"learning_rate": 5.380434782608695e-06,
|
| 37 |
+
"loss": 4.3772,
|
| 38 |
+
"step": 100
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.6821282401091405,
|
| 42 |
+
"grad_norm": 18.209796905517578,
|
| 43 |
+
"learning_rate": 6.739130434782609e-06,
|
| 44 |
+
"loss": 4.3307,
|
| 45 |
+
"step": 125
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.8185538881309686,
|
| 49 |
+
"grad_norm": 21.05099105834961,
|
| 50 |
+
"learning_rate": 8.097826086956523e-06,
|
| 51 |
+
"loss": 4.3437,
|
| 52 |
+
"step": 150
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"epoch": 0.9549795361527967,
|
| 56 |
+
"grad_norm": 15.863072395324707,
|
| 57 |
+
"learning_rate": 9.456521739130436e-06,
|
| 58 |
+
"loss": 4.2182,
|
| 59 |
+
"step": 175
|
| 60 |
+
},
|
| 61 |
+
{
|
| 62 |
+
"epoch": 1.0,
|
| 63 |
+
"eval_accuracy": 0.49732620320855614,
|
| 64 |
+
"eval_loss": 1.0000888109207153,
|
| 65 |
+
"eval_macro_f1": 0.4271992952816907,
|
| 66 |
+
"eval_runtime": 0.8087,
|
| 67 |
+
"eval_samples_per_second": 924.95,
|
| 68 |
+
"eval_steps_per_second": 58.118,
|
| 69 |
+
"eval_weighted_f1": 0.4321530862066507,
|
| 70 |
+
"step": 184
|
| 71 |
+
},
|
| 72 |
+
{
|
| 73 |
+
"epoch": 1.08731241473397,
|
| 74 |
+
"grad_norm": 21.70718765258789,
|
| 75 |
+
"learning_rate": 1.0815217391304349e-05,
|
| 76 |
+
"loss": 3.9352,
|
| 77 |
+
"step": 200
|
| 78 |
+
},
|
| 79 |
+
{
|
| 80 |
+
"epoch": 1.223738062755798,
|
| 81 |
+
"grad_norm": 70.66429138183594,
|
| 82 |
+
"learning_rate": 1.2173913043478263e-05,
|
| 83 |
+
"loss": 3.7315,
|
| 84 |
+
"step": 225
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 1.360163710777626,
|
| 88 |
+
"grad_norm": 82.18575286865234,
|
| 89 |
+
"learning_rate": 1.3532608695652174e-05,
|
| 90 |
+
"loss": 3.3856,
|
| 91 |
+
"step": 250
|
| 92 |
+
},
|
| 93 |
+
{
|
| 94 |
+
"epoch": 1.4965893587994543,
|
| 95 |
+
"grad_norm": 72.28233337402344,
|
| 96 |
+
"learning_rate": 1.4891304347826087e-05,
|
| 97 |
+
"loss": 3.109,
|
| 98 |
+
"step": 275
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 1.6330150068212825,
|
| 102 |
+
"grad_norm": 38.895118713378906,
|
| 103 |
+
"learning_rate": 1.6250000000000002e-05,
|
| 104 |
+
"loss": 3.0423,
|
| 105 |
+
"step": 300
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 1.7694406548431105,
|
| 109 |
+
"grad_norm": 46.553977966308594,
|
| 110 |
+
"learning_rate": 1.7608695652173915e-05,
|
| 111 |
+
"loss": 2.6991,
|
| 112 |
+
"step": 325
|
| 113 |
+
},
|
| 114 |
+
{
|
| 115 |
+
"epoch": 1.9058663028649385,
|
| 116 |
+
"grad_norm": 73.34980773925781,
|
| 117 |
+
"learning_rate": 1.8967391304347828e-05,
|
| 118 |
+
"loss": 2.4286,
|
| 119 |
+
"step": 350
|
| 120 |
+
},
|
| 121 |
+
{
|
| 122 |
+
"epoch": 2.0,
|
| 123 |
+
"eval_accuracy": 0.6818181818181818,
|
| 124 |
+
"eval_loss": 0.7317593097686768,
|
| 125 |
+
"eval_macro_f1": 0.6724501069918717,
|
| 126 |
+
"eval_runtime": 0.8509,
|
| 127 |
+
"eval_samples_per_second": 879.105,
|
| 128 |
+
"eval_steps_per_second": 55.238,
|
| 129 |
+
"eval_weighted_f1": 0.6733933760795694,
|
| 130 |
+
"step": 368
|
| 131 |
+
},
|
| 132 |
+
{
|
| 133 |
+
"epoch": 2.038199181446112,
|
| 134 |
+
"grad_norm": 40.529052734375,
|
| 135 |
+
"learning_rate": 1.9963768115942032e-05,
|
| 136 |
+
"loss": 2.4305,
|
| 137 |
+
"step": 375
|
| 138 |
+
},
|
| 139 |
+
{
|
| 140 |
+
"epoch": 2.17462482946794,
|
| 141 |
+
"grad_norm": 41.45320510864258,
|
| 142 |
+
"learning_rate": 1.9812801932367153e-05,
|
| 143 |
+
"loss": 2.1477,
|
| 144 |
+
"step": 400
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 2.311050477489768,
|
| 148 |
+
"grad_norm": 29.474023818969727,
|
| 149 |
+
"learning_rate": 1.966183574879227e-05,
|
| 150 |
+
"loss": 2.0595,
|
| 151 |
+
"step": 425
|
| 152 |
+
},
|
| 153 |
+
{
|
| 154 |
+
"epoch": 2.447476125511596,
|
| 155 |
+
"grad_norm": 38.42793655395508,
|
| 156 |
+
"learning_rate": 1.9510869565217393e-05,
|
| 157 |
+
"loss": 1.9622,
|
| 158 |
+
"step": 450
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 2.5839017735334244,
|
| 162 |
+
"grad_norm": 32.58100128173828,
|
| 163 |
+
"learning_rate": 1.9359903381642515e-05,
|
| 164 |
+
"loss": 1.9382,
|
| 165 |
+
"step": 475
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 2.720327421555252,
|
| 169 |
+
"grad_norm": 34.730350494384766,
|
| 170 |
+
"learning_rate": 1.9208937198067633e-05,
|
| 171 |
+
"loss": 2.1929,
|
| 172 |
+
"step": 500
|
| 173 |
+
},
|
| 174 |
+
{
|
| 175 |
+
"epoch": 2.8567530695770804,
|
| 176 |
+
"grad_norm": 31.687461853027344,
|
| 177 |
+
"learning_rate": 1.9057971014492754e-05,
|
| 178 |
+
"loss": 1.9919,
|
| 179 |
+
"step": 525
|
| 180 |
+
},
|
| 181 |
+
{
|
| 182 |
+
"epoch": 2.9931787175989086,
|
| 183 |
+
"grad_norm": 33.949790954589844,
|
| 184 |
+
"learning_rate": 1.8907004830917876e-05,
|
| 185 |
+
"loss": 2.1664,
|
| 186 |
+
"step": 550
|
| 187 |
+
},
|
| 188 |
+
{
|
| 189 |
+
"epoch": 3.0,
|
| 190 |
+
"eval_accuracy": 0.81951871657754,
|
| 191 |
+
"eval_loss": 0.49579721689224243,
|
| 192 |
+
"eval_macro_f1": 0.8170391895457513,
|
| 193 |
+
"eval_runtime": 0.798,
|
| 194 |
+
"eval_samples_per_second": 937.352,
|
| 195 |
+
"eval_steps_per_second": 58.898,
|
| 196 |
+
"eval_weighted_f1": 0.8176412342953566,
|
| 197 |
+
"step": 552
|
| 198 |
+
},
|
| 199 |
+
{
|
| 200 |
+
"epoch": 3.1255115961800817,
|
| 201 |
+
"grad_norm": 55.468326568603516,
|
| 202 |
+
"learning_rate": 1.8756038647342997e-05,
|
| 203 |
+
"loss": 1.1882,
|
| 204 |
+
"step": 575
|
| 205 |
+
},
|
| 206 |
+
{
|
| 207 |
+
"epoch": 3.26193724420191,
|
| 208 |
+
"grad_norm": 48.688045501708984,
|
| 209 |
+
"learning_rate": 1.8605072463768115e-05,
|
| 210 |
+
"loss": 1.4517,
|
| 211 |
+
"step": 600
|
| 212 |
+
},
|
| 213 |
+
{
|
| 214 |
+
"epoch": 3.398362892223738,
|
| 215 |
+
"grad_norm": 47.39004898071289,
|
| 216 |
+
"learning_rate": 1.8454106280193237e-05,
|
| 217 |
+
"loss": 1.371,
|
| 218 |
+
"step": 625
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 3.5347885402455663,
|
| 222 |
+
"grad_norm": 109.52153015136719,
|
| 223 |
+
"learning_rate": 1.830314009661836e-05,
|
| 224 |
+
"loss": 1.3941,
|
| 225 |
+
"step": 650
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"epoch": 3.6712141882673945,
|
| 229 |
+
"grad_norm": 52.8348274230957,
|
| 230 |
+
"learning_rate": 1.815217391304348e-05,
|
| 231 |
+
"loss": 1.3808,
|
| 232 |
+
"step": 675
|
| 233 |
+
},
|
| 234 |
+
{
|
| 235 |
+
"epoch": 3.8076398362892223,
|
| 236 |
+
"grad_norm": 35.71442794799805,
|
| 237 |
+
"learning_rate": 1.8001207729468598e-05,
|
| 238 |
+
"loss": 1.3991,
|
| 239 |
+
"step": 700
|
| 240 |
+
},
|
| 241 |
+
{
|
| 242 |
+
"epoch": 3.9440654843110505,
|
| 243 |
+
"grad_norm": 35.96894073486328,
|
| 244 |
+
"learning_rate": 1.785024154589372e-05,
|
| 245 |
+
"loss": 1.4121,
|
| 246 |
+
"step": 725
|
| 247 |
+
},
|
| 248 |
+
{
|
| 249 |
+
"epoch": 4.0,
|
| 250 |
+
"eval_accuracy": 0.8101604278074866,
|
| 251 |
+
"eval_loss": 0.55982506275177,
|
| 252 |
+
"eval_macro_f1": 0.8094223219903971,
|
| 253 |
+
"eval_runtime": 0.7864,
|
| 254 |
+
"eval_samples_per_second": 951.134,
|
| 255 |
+
"eval_steps_per_second": 59.764,
|
| 256 |
+
"eval_weighted_f1": 0.8098770559042625,
|
| 257 |
+
"step": 736
|
| 258 |
+
},
|
| 259 |
+
{
|
| 260 |
+
"epoch": 4.076398362892224,
|
| 261 |
+
"grad_norm": 49.468929290771484,
|
| 262 |
+
"learning_rate": 1.769927536231884e-05,
|
| 263 |
+
"loss": 1.1229,
|
| 264 |
+
"step": 750
|
| 265 |
+
},
|
| 266 |
+
{
|
| 267 |
+
"epoch": 4.212824010914052,
|
| 268 |
+
"grad_norm": 74.64271545410156,
|
| 269 |
+
"learning_rate": 1.7548309178743963e-05,
|
| 270 |
+
"loss": 0.9943,
|
| 271 |
+
"step": 775
|
| 272 |
+
},
|
| 273 |
+
{
|
| 274 |
+
"epoch": 4.34924965893588,
|
| 275 |
+
"grad_norm": 41.42115020751953,
|
| 276 |
+
"learning_rate": 1.7397342995169084e-05,
|
| 277 |
+
"loss": 1.1671,
|
| 278 |
+
"step": 800
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 4.485675306957708,
|
| 282 |
+
"grad_norm": 31.184541702270508,
|
| 283 |
+
"learning_rate": 1.7246376811594206e-05,
|
| 284 |
+
"loss": 0.8163,
|
| 285 |
+
"step": 825
|
| 286 |
+
},
|
| 287 |
+
{
|
| 288 |
+
"epoch": 4.622100954979536,
|
| 289 |
+
"grad_norm": 46.80307388305664,
|
| 290 |
+
"learning_rate": 1.7095410628019324e-05,
|
| 291 |
+
"loss": 0.7534,
|
| 292 |
+
"step": 850
|
| 293 |
+
},
|
| 294 |
+
{
|
| 295 |
+
"epoch": 4.758526603001364,
|
| 296 |
+
"grad_norm": 69.3770751953125,
|
| 297 |
+
"learning_rate": 1.6944444444444446e-05,
|
| 298 |
+
"loss": 0.9733,
|
| 299 |
+
"step": 875
|
| 300 |
+
},
|
| 301 |
+
{
|
| 302 |
+
"epoch": 4.894952251023192,
|
| 303 |
+
"grad_norm": 10.779870986938477,
|
| 304 |
+
"learning_rate": 1.6793478260869567e-05,
|
| 305 |
+
"loss": 0.8835,
|
| 306 |
+
"step": 900
|
| 307 |
+
},
|
| 308 |
+
{
|
| 309 |
+
"epoch": 5.0,
|
| 310 |
+
"eval_accuracy": 0.7941176470588235,
|
| 311 |
+
"eval_loss": 0.6966820359230042,
|
| 312 |
+
"eval_macro_f1": 0.7932602331198986,
|
| 313 |
+
"eval_runtime": 0.7933,
|
| 314 |
+
"eval_samples_per_second": 942.858,
|
| 315 |
+
"eval_steps_per_second": 59.244,
|
| 316 |
+
"eval_weighted_f1": 0.794286227392676,
|
| 317 |
+
"step": 920
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 5.0272851296043655,
|
| 321 |
+
"grad_norm": 47.47770309448242,
|
| 322 |
+
"learning_rate": 1.664251207729469e-05,
|
| 323 |
+
"loss": 1.0186,
|
| 324 |
+
"step": 925
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 5.163710777626194,
|
| 328 |
+
"grad_norm": 60.878822326660156,
|
| 329 |
+
"learning_rate": 1.649154589371981e-05,
|
| 330 |
+
"loss": 0.819,
|
| 331 |
+
"step": 950
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 5.300136425648022,
|
| 335 |
+
"grad_norm": 118.20897674560547,
|
| 336 |
+
"learning_rate": 1.634057971014493e-05,
|
| 337 |
+
"loss": 0.6266,
|
| 338 |
+
"step": 975
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 5.43656207366985,
|
| 342 |
+
"grad_norm": 43.0,
|
| 343 |
+
"learning_rate": 1.618961352657005e-05,
|
| 344 |
+
"loss": 0.5908,
|
| 345 |
+
"step": 1000
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 5.572987721691678,
|
| 349 |
+
"grad_norm": 53.79621124267578,
|
| 350 |
+
"learning_rate": 1.603864734299517e-05,
|
| 351 |
+
"loss": 0.627,
|
| 352 |
+
"step": 1025
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 5.709413369713506,
|
| 356 |
+
"grad_norm": 55.738792419433594,
|
| 357 |
+
"learning_rate": 1.5887681159420293e-05,
|
| 358 |
+
"loss": 0.7622,
|
| 359 |
+
"step": 1050
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 5.845839017735334,
|
| 363 |
+
"grad_norm": 6.537409782409668,
|
| 364 |
+
"learning_rate": 1.5736714975845415e-05,
|
| 365 |
+
"loss": 0.6851,
|
| 366 |
+
"step": 1075
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 5.982264665757162,
|
| 370 |
+
"grad_norm": 23.49893569946289,
|
| 371 |
+
"learning_rate": 1.5585748792270533e-05,
|
| 372 |
+
"loss": 0.5851,
|
| 373 |
+
"step": 1100
|
| 374 |
+
},
|
| 375 |
+
{
|
| 376 |
+
"epoch": 6.0,
|
| 377 |
+
"eval_accuracy": 0.8008021390374331,
|
| 378 |
+
"eval_loss": 0.72493976354599,
|
| 379 |
+
"eval_macro_f1": 0.801877684478554,
|
| 380 |
+
"eval_runtime": 0.7989,
|
| 381 |
+
"eval_samples_per_second": 936.233,
|
| 382 |
+
"eval_steps_per_second": 58.827,
|
| 383 |
+
"eval_weighted_f1": 0.8022964360701532,
|
| 384 |
+
"step": 1104
|
| 385 |
+
}
|
| 386 |
+
],
|
| 387 |
+
"logging_steps": 25,
|
| 388 |
+
"max_steps": 3680,
|
| 389 |
+
"num_input_tokens_seen": 0,
|
| 390 |
+
"num_train_epochs": 20,
|
| 391 |
+
"save_steps": 500,
|
| 392 |
+
"stateful_callbacks": {
|
| 393 |
+
"EarlyStoppingCallback": {
|
| 394 |
+
"args": {
|
| 395 |
+
"early_stopping_patience": 3,
|
| 396 |
+
"early_stopping_threshold": 0.0
|
| 397 |
+
},
|
| 398 |
+
"attributes": {
|
| 399 |
+
"early_stopping_patience_counter": 3
|
| 400 |
+
}
|
| 401 |
+
},
|
| 402 |
+
"TrainerControl": {
|
| 403 |
+
"args": {
|
| 404 |
+
"should_epoch_stop": false,
|
| 405 |
+
"should_evaluate": false,
|
| 406 |
+
"should_log": false,
|
| 407 |
+
"should_save": true,
|
| 408 |
+
"should_training_stop": true
|
| 409 |
+
},
|
| 410 |
+
"attributes": {}
|
| 411 |
+
}
|
| 412 |
+
},
|
| 413 |
+
"total_flos": 0.0,
|
| 414 |
+
"train_batch_size": 8,
|
| 415 |
+
"trial_name": null,
|
| 416 |
+
"trial_params": null
|
| 417 |
+
}
|
ood-vifactcheck-bamibert-topic-domain-06-s43-20260827T195550Z-3858bb98/trainer/checkpoint-552/special_tokens_map.json
ADDED
|
@@ -0,0 +1,51 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": {
|
| 3 |
+
"content": "<s>",
|
| 4 |
+
"lstrip": false,
|
| 5 |
+
"normalized": false,
|
| 6 |
+
"rstrip": false,
|
| 7 |
+
"single_word": false
|
| 8 |
+
},
|
| 9 |
+
"cls_token": {
|
| 10 |
+
"content": "<s>",
|
| 11 |
+
"lstrip": false,
|
| 12 |
+
"normalized": false,
|
| 13 |
+
"rstrip": false,
|
| 14 |
+
"single_word": false
|
| 15 |
+
},
|
| 16 |
+
"eos_token": {
|
| 17 |
+
"content": "</s>",
|
| 18 |
+
"lstrip": false,
|
| 19 |
+
"normalized": false,
|
| 20 |
+
"rstrip": false,
|
| 21 |
+
"single_word": false
|
| 22 |
+
},
|
| 23 |
+
"mask_token": {
|
| 24 |
+
"content": "<mask>",
|
| 25 |
+
"lstrip": false,
|
| 26 |
+
"normalized": false,
|
| 27 |
+
"rstrip": false,
|
| 28 |
+
"single_word": false
|
| 29 |
+
},
|
| 30 |
+
"pad_token": {
|
| 31 |
+
"content": "<pad>",
|
| 32 |
+
"lstrip": false,
|
| 33 |
+
"normalized": false,
|
| 34 |
+
"rstrip": false,
|
| 35 |
+
"single_word": false
|
| 36 |
+
},
|
| 37 |
+
"sep_token": {
|
| 38 |
+
"content": "</s>",
|
| 39 |
+
"lstrip": false,
|
| 40 |
+
"normalized": false,
|
| 41 |
+
"rstrip": false,
|
| 42 |
+
"single_word": false
|
| 43 |
+
},
|
| 44 |
+
"unk_token": {
|
| 45 |
+
"content": "<unk>",
|
| 46 |
+
"lstrip": false,
|
| 47 |
+
"normalized": false,
|
| 48 |
+
"rstrip": false,
|
| 49 |
+
"single_word": false
|
| 50 |
+
}
|
| 51 |
+
}
|
ood-vifactcheck-bamibert-topic-domain-06-s43-20260827T195550Z-3858bb98/trainer/checkpoint-552/tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-bamibert-topic-domain-06-s43-20260827T195550Z-3858bb98/trainer/checkpoint-552/tokenizer_config.json
ADDED
|
@@ -0,0 +1,55 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"added_tokens_decoder": {
|
| 3 |
+
"0": {
|
| 4 |
+
"content": "<s>",
|
| 5 |
+
"lstrip": false,
|
| 6 |
+
"normalized": false,
|
| 7 |
+
"rstrip": false,
|
| 8 |
+
"single_word": false,
|
| 9 |
+
"special": true
|
| 10 |
+
},
|
| 11 |
+
"1": {
|
| 12 |
+
"content": "<pad>",
|
| 13 |
+
"lstrip": false,
|
| 14 |
+
"normalized": false,
|
| 15 |
+
"rstrip": false,
|
| 16 |
+
"single_word": false,
|
| 17 |
+
"special": true
|
| 18 |
+
},
|
| 19 |
+
"2": {
|
| 20 |
+
"content": "</s>",
|
| 21 |
+
"lstrip": false,
|
| 22 |
+
"normalized": false,
|
| 23 |
+
"rstrip": false,
|
| 24 |
+
"single_word": false,
|
| 25 |
+
"special": true
|
| 26 |
+
},
|
| 27 |
+
"3": {
|
| 28 |
+
"content": "<unk>",
|
| 29 |
+
"lstrip": false,
|
| 30 |
+
"normalized": false,
|
| 31 |
+
"rstrip": false,
|
| 32 |
+
"single_word": false,
|
| 33 |
+
"special": true
|
| 34 |
+
},
|
| 35 |
+
"20480": {
|
| 36 |
+
"content": "<mask>",
|
| 37 |
+
"lstrip": false,
|
| 38 |
+
"normalized": false,
|
| 39 |
+
"rstrip": false,
|
| 40 |
+
"single_word": false,
|
| 41 |
+
"special": true
|
| 42 |
+
}
|
| 43 |
+
},
|
| 44 |
+
"bos_token": "<s>",
|
| 45 |
+
"clean_up_tokenization_spaces": false,
|
| 46 |
+
"cls_token": "<s>",
|
| 47 |
+
"eos_token": "</s>",
|
| 48 |
+
"extra_special_tokens": {},
|
| 49 |
+
"mask_token": "<mask>",
|
| 50 |
+
"model_max_length": 2048,
|
| 51 |
+
"pad_token": "<pad>",
|
| 52 |
+
"sep_token": "</s>",
|
| 53 |
+
"tokenizer_class": "XLMRobertaTokenizerFast",
|
| 54 |
+
"unk_token": "<unk>"
|
| 55 |
+
}
|
ood-vifactcheck-bamibert-topic-domain-06-s43-20260827T195550Z-3858bb98/trainer/checkpoint-552/trainer_state.json
ADDED
|
@@ -0,0 +1,230 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 552,
|
| 3 |
+
"best_metric": 0.8170391895457513,
|
| 4 |
+
"best_model_checkpoint": "/root/VNEncoder/results/checkpoints/ood-vifactcheck-bamibert-topic-domain-06-s43-20260827T195550Z-3858bb98/trainer/checkpoint-552",
|
| 5 |
+
"epoch": 3.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 552,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.1364256480218281,
|
| 14 |
+
"grad_norm": 13.91116714477539,
|
| 15 |
+
"learning_rate": 1.3043478260869566e-06,
|
| 16 |
+
"loss": 4.4188,
|
| 17 |
+
"step": 25
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.2728512960436562,
|
| 21 |
+
"grad_norm": 13.288341522216797,
|
| 22 |
+
"learning_rate": 2.6630434782608698e-06,
|
| 23 |
+
"loss": 4.3869,
|
| 24 |
+
"step": 50
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.4092769440654843,
|
| 28 |
+
"grad_norm": 10.405351638793945,
|
| 29 |
+
"learning_rate": 4.021739130434783e-06,
|
| 30 |
+
"loss": 4.4184,
|
| 31 |
+
"step": 75
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.5457025920873124,
|
| 35 |
+
"grad_norm": 13.788087844848633,
|
| 36 |
+
"learning_rate": 5.380434782608695e-06,
|
| 37 |
+
"loss": 4.3772,
|
| 38 |
+
"step": 100
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.6821282401091405,
|
| 42 |
+
"grad_norm": 18.209796905517578,
|
| 43 |
+
"learning_rate": 6.739130434782609e-06,
|
| 44 |
+
"loss": 4.3307,
|
| 45 |
+
"step": 125
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.8185538881309686,
|
| 49 |
+
"grad_norm": 21.05099105834961,
|
| 50 |
+
"learning_rate": 8.097826086956523e-06,
|
| 51 |
+
"loss": 4.3437,
|
| 52 |
+
"step": 150
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"epoch": 0.9549795361527967,
|
| 56 |
+
"grad_norm": 15.863072395324707,
|
| 57 |
+
"learning_rate": 9.456521739130436e-06,
|
| 58 |
+
"loss": 4.2182,
|
| 59 |
+
"step": 175
|
| 60 |
+
},
|
| 61 |
+
{
|
| 62 |
+
"epoch": 1.0,
|
| 63 |
+
"eval_accuracy": 0.49732620320855614,
|
| 64 |
+
"eval_loss": 1.0000888109207153,
|
| 65 |
+
"eval_macro_f1": 0.4271992952816907,
|
| 66 |
+
"eval_runtime": 0.8087,
|
| 67 |
+
"eval_samples_per_second": 924.95,
|
| 68 |
+
"eval_steps_per_second": 58.118,
|
| 69 |
+
"eval_weighted_f1": 0.4321530862066507,
|
| 70 |
+
"step": 184
|
| 71 |
+
},
|
| 72 |
+
{
|
| 73 |
+
"epoch": 1.08731241473397,
|
| 74 |
+
"grad_norm": 21.70718765258789,
|
| 75 |
+
"learning_rate": 1.0815217391304349e-05,
|
| 76 |
+
"loss": 3.9352,
|
| 77 |
+
"step": 200
|
| 78 |
+
},
|
| 79 |
+
{
|
| 80 |
+
"epoch": 1.223738062755798,
|
| 81 |
+
"grad_norm": 70.66429138183594,
|
| 82 |
+
"learning_rate": 1.2173913043478263e-05,
|
| 83 |
+
"loss": 3.7315,
|
| 84 |
+
"step": 225
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 1.360163710777626,
|
| 88 |
+
"grad_norm": 82.18575286865234,
|
| 89 |
+
"learning_rate": 1.3532608695652174e-05,
|
| 90 |
+
"loss": 3.3856,
|
| 91 |
+
"step": 250
|
| 92 |
+
},
|
| 93 |
+
{
|
| 94 |
+
"epoch": 1.4965893587994543,
|
| 95 |
+
"grad_norm": 72.28233337402344,
|
| 96 |
+
"learning_rate": 1.4891304347826087e-05,
|
| 97 |
+
"loss": 3.109,
|
| 98 |
+
"step": 275
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 1.6330150068212825,
|
| 102 |
+
"grad_norm": 38.895118713378906,
|
| 103 |
+
"learning_rate": 1.6250000000000002e-05,
|
| 104 |
+
"loss": 3.0423,
|
| 105 |
+
"step": 300
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 1.7694406548431105,
|
| 109 |
+
"grad_norm": 46.553977966308594,
|
| 110 |
+
"learning_rate": 1.7608695652173915e-05,
|
| 111 |
+
"loss": 2.6991,
|
| 112 |
+
"step": 325
|
| 113 |
+
},
|
| 114 |
+
{
|
| 115 |
+
"epoch": 1.9058663028649385,
|
| 116 |
+
"grad_norm": 73.34980773925781,
|
| 117 |
+
"learning_rate": 1.8967391304347828e-05,
|
| 118 |
+
"loss": 2.4286,
|
| 119 |
+
"step": 350
|
| 120 |
+
},
|
| 121 |
+
{
|
| 122 |
+
"epoch": 2.0,
|
| 123 |
+
"eval_accuracy": 0.6818181818181818,
|
| 124 |
+
"eval_loss": 0.7317593097686768,
|
| 125 |
+
"eval_macro_f1": 0.6724501069918717,
|
| 126 |
+
"eval_runtime": 0.8509,
|
| 127 |
+
"eval_samples_per_second": 879.105,
|
| 128 |
+
"eval_steps_per_second": 55.238,
|
| 129 |
+
"eval_weighted_f1": 0.6733933760795694,
|
| 130 |
+
"step": 368
|
| 131 |
+
},
|
| 132 |
+
{
|
| 133 |
+
"epoch": 2.038199181446112,
|
| 134 |
+
"grad_norm": 40.529052734375,
|
| 135 |
+
"learning_rate": 1.9963768115942032e-05,
|
| 136 |
+
"loss": 2.4305,
|
| 137 |
+
"step": 375
|
| 138 |
+
},
|
| 139 |
+
{
|
| 140 |
+
"epoch": 2.17462482946794,
|
| 141 |
+
"grad_norm": 41.45320510864258,
|
| 142 |
+
"learning_rate": 1.9812801932367153e-05,
|
| 143 |
+
"loss": 2.1477,
|
| 144 |
+
"step": 400
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 2.311050477489768,
|
| 148 |
+
"grad_norm": 29.474023818969727,
|
| 149 |
+
"learning_rate": 1.966183574879227e-05,
|
| 150 |
+
"loss": 2.0595,
|
| 151 |
+
"step": 425
|
| 152 |
+
},
|
| 153 |
+
{
|
| 154 |
+
"epoch": 2.447476125511596,
|
| 155 |
+
"grad_norm": 38.42793655395508,
|
| 156 |
+
"learning_rate": 1.9510869565217393e-05,
|
| 157 |
+
"loss": 1.9622,
|
| 158 |
+
"step": 450
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 2.5839017735334244,
|
| 162 |
+
"grad_norm": 32.58100128173828,
|
| 163 |
+
"learning_rate": 1.9359903381642515e-05,
|
| 164 |
+
"loss": 1.9382,
|
| 165 |
+
"step": 475
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 2.720327421555252,
|
| 169 |
+
"grad_norm": 34.730350494384766,
|
| 170 |
+
"learning_rate": 1.9208937198067633e-05,
|
| 171 |
+
"loss": 2.1929,
|
| 172 |
+
"step": 500
|
| 173 |
+
},
|
| 174 |
+
{
|
| 175 |
+
"epoch": 2.8567530695770804,
|
| 176 |
+
"grad_norm": 31.687461853027344,
|
| 177 |
+
"learning_rate": 1.9057971014492754e-05,
|
| 178 |
+
"loss": 1.9919,
|
| 179 |
+
"step": 525
|
| 180 |
+
},
|
| 181 |
+
{
|
| 182 |
+
"epoch": 2.9931787175989086,
|
| 183 |
+
"grad_norm": 33.949790954589844,
|
| 184 |
+
"learning_rate": 1.8907004830917876e-05,
|
| 185 |
+
"loss": 2.1664,
|
| 186 |
+
"step": 550
|
| 187 |
+
},
|
| 188 |
+
{
|
| 189 |
+
"epoch": 3.0,
|
| 190 |
+
"eval_accuracy": 0.81951871657754,
|
| 191 |
+
"eval_loss": 0.49579721689224243,
|
| 192 |
+
"eval_macro_f1": 0.8170391895457513,
|
| 193 |
+
"eval_runtime": 0.798,
|
| 194 |
+
"eval_samples_per_second": 937.352,
|
| 195 |
+
"eval_steps_per_second": 58.898,
|
| 196 |
+
"eval_weighted_f1": 0.8176412342953566,
|
| 197 |
+
"step": 552
|
| 198 |
+
}
|
| 199 |
+
],
|
| 200 |
+
"logging_steps": 25,
|
| 201 |
+
"max_steps": 3680,
|
| 202 |
+
"num_input_tokens_seen": 0,
|
| 203 |
+
"num_train_epochs": 20,
|
| 204 |
+
"save_steps": 500,
|
| 205 |
+
"stateful_callbacks": {
|
| 206 |
+
"EarlyStoppingCallback": {
|
| 207 |
+
"args": {
|
| 208 |
+
"early_stopping_patience": 3,
|
| 209 |
+
"early_stopping_threshold": 0.0
|
| 210 |
+
},
|
| 211 |
+
"attributes": {
|
| 212 |
+
"early_stopping_patience_counter": 0
|
| 213 |
+
}
|
| 214 |
+
},
|
| 215 |
+
"TrainerControl": {
|
| 216 |
+
"args": {
|
| 217 |
+
"should_epoch_stop": false,
|
| 218 |
+
"should_evaluate": false,
|
| 219 |
+
"should_log": false,
|
| 220 |
+
"should_save": true,
|
| 221 |
+
"should_training_stop": false
|
| 222 |
+
},
|
| 223 |
+
"attributes": {}
|
| 224 |
+
}
|
| 225 |
+
},
|
| 226 |
+
"total_flos": 0.0,
|
| 227 |
+
"train_batch_size": 8,
|
| 228 |
+
"trial_name": null,
|
| 229 |
+
"trial_params": null
|
| 230 |
+
}
|
ood-vifactcheck-bamibert-topic-domain-07-s43-20260827T195828Z-3858bb98/trainer/checkpoint-2167/special_tokens_map.json
ADDED
|
@@ -0,0 +1,51 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": {
|
| 3 |
+
"content": "<s>",
|
| 4 |
+
"lstrip": false,
|
| 5 |
+
"normalized": false,
|
| 6 |
+
"rstrip": false,
|
| 7 |
+
"single_word": false
|
| 8 |
+
},
|
| 9 |
+
"cls_token": {
|
| 10 |
+
"content": "<s>",
|
| 11 |
+
"lstrip": false,
|
| 12 |
+
"normalized": false,
|
| 13 |
+
"rstrip": false,
|
| 14 |
+
"single_word": false
|
| 15 |
+
},
|
| 16 |
+
"eos_token": {
|
| 17 |
+
"content": "</s>",
|
| 18 |
+
"lstrip": false,
|
| 19 |
+
"normalized": false,
|
| 20 |
+
"rstrip": false,
|
| 21 |
+
"single_word": false
|
| 22 |
+
},
|
| 23 |
+
"mask_token": {
|
| 24 |
+
"content": "<mask>",
|
| 25 |
+
"lstrip": false,
|
| 26 |
+
"normalized": false,
|
| 27 |
+
"rstrip": false,
|
| 28 |
+
"single_word": false
|
| 29 |
+
},
|
| 30 |
+
"pad_token": {
|
| 31 |
+
"content": "<pad>",
|
| 32 |
+
"lstrip": false,
|
| 33 |
+
"normalized": false,
|
| 34 |
+
"rstrip": false,
|
| 35 |
+
"single_word": false
|
| 36 |
+
},
|
| 37 |
+
"sep_token": {
|
| 38 |
+
"content": "</s>",
|
| 39 |
+
"lstrip": false,
|
| 40 |
+
"normalized": false,
|
| 41 |
+
"rstrip": false,
|
| 42 |
+
"single_word": false
|
| 43 |
+
},
|
| 44 |
+
"unk_token": {
|
| 45 |
+
"content": "<unk>",
|
| 46 |
+
"lstrip": false,
|
| 47 |
+
"normalized": false,
|
| 48 |
+
"rstrip": false,
|
| 49 |
+
"single_word": false
|
| 50 |
+
}
|
| 51 |
+
}
|
ood-vifactcheck-bamibert-topic-domain-07-s43-20260827T195828Z-3858bb98/trainer/checkpoint-2167/tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-bamibert-topic-domain-07-s43-20260827T195828Z-3858bb98/trainer/checkpoint-2167/tokenizer_config.json
ADDED
|
@@ -0,0 +1,55 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"added_tokens_decoder": {
|
| 3 |
+
"0": {
|
| 4 |
+
"content": "<s>",
|
| 5 |
+
"lstrip": false,
|
| 6 |
+
"normalized": false,
|
| 7 |
+
"rstrip": false,
|
| 8 |
+
"single_word": false,
|
| 9 |
+
"special": true
|
| 10 |
+
},
|
| 11 |
+
"1": {
|
| 12 |
+
"content": "<pad>",
|
| 13 |
+
"lstrip": false,
|
| 14 |
+
"normalized": false,
|
| 15 |
+
"rstrip": false,
|
| 16 |
+
"single_word": false,
|
| 17 |
+
"special": true
|
| 18 |
+
},
|
| 19 |
+
"2": {
|
| 20 |
+
"content": "</s>",
|
| 21 |
+
"lstrip": false,
|
| 22 |
+
"normalized": false,
|
| 23 |
+
"rstrip": false,
|
| 24 |
+
"single_word": false,
|
| 25 |
+
"special": true
|
| 26 |
+
},
|
| 27 |
+
"3": {
|
| 28 |
+
"content": "<unk>",
|
| 29 |
+
"lstrip": false,
|
| 30 |
+
"normalized": false,
|
| 31 |
+
"rstrip": false,
|
| 32 |
+
"single_word": false,
|
| 33 |
+
"special": true
|
| 34 |
+
},
|
| 35 |
+
"20480": {
|
| 36 |
+
"content": "<mask>",
|
| 37 |
+
"lstrip": false,
|
| 38 |
+
"normalized": false,
|
| 39 |
+
"rstrip": false,
|
| 40 |
+
"single_word": false,
|
| 41 |
+
"special": true
|
| 42 |
+
}
|
| 43 |
+
},
|
| 44 |
+
"bos_token": "<s>",
|
| 45 |
+
"clean_up_tokenization_spaces": false,
|
| 46 |
+
"cls_token": "<s>",
|
| 47 |
+
"eos_token": "</s>",
|
| 48 |
+
"extra_special_tokens": {},
|
| 49 |
+
"mask_token": "<mask>",
|
| 50 |
+
"model_max_length": 2048,
|
| 51 |
+
"pad_token": "<pad>",
|
| 52 |
+
"sep_token": "</s>",
|
| 53 |
+
"tokenizer_class": "XLMRobertaTokenizerFast",
|
| 54 |
+
"unk_token": "<unk>"
|
| 55 |
+
}
|
ood-vifactcheck-bamibert-topic-domain-07-s43-20260827T195828Z-3858bb98/trainer/checkpoint-2167/trainer_state.json
ADDED
|
@@ -0,0 +1,766 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 2167,
|
| 3 |
+
"best_metric": 0.8423839833625967,
|
| 4 |
+
"best_model_checkpoint": "/root/VNEncoder/results/checkpoints/ood-vifactcheck-bamibert-topic-domain-07-s43-20260827T195828Z-3858bb98/trainer/checkpoint-2167",
|
| 5 |
+
"epoch": 11.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 2167,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.12706480304955528,
|
| 14 |
+
"grad_norm": 29.48038673400879,
|
| 15 |
+
"learning_rate": 1.218274111675127e-06,
|
| 16 |
+
"loss": 4.4267,
|
| 17 |
+
"step": 25
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.25412960609911056,
|
| 21 |
+
"grad_norm": 14.88051700592041,
|
| 22 |
+
"learning_rate": 2.487309644670051e-06,
|
| 23 |
+
"loss": 4.4197,
|
| 24 |
+
"step": 50
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.3811944091486658,
|
| 28 |
+
"grad_norm": 6.070577144622803,
|
| 29 |
+
"learning_rate": 3.756345177664975e-06,
|
| 30 |
+
"loss": 4.4149,
|
| 31 |
+
"step": 75
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.5082592121982211,
|
| 35 |
+
"grad_norm": 7.696415901184082,
|
| 36 |
+
"learning_rate": 5.025380710659899e-06,
|
| 37 |
+
"loss": 4.3894,
|
| 38 |
+
"step": 100
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.6353240152477764,
|
| 42 |
+
"grad_norm": 52.55864715576172,
|
| 43 |
+
"learning_rate": 6.2944162436548225e-06,
|
| 44 |
+
"loss": 4.3573,
|
| 45 |
+
"step": 125
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.7623888182973316,
|
| 49 |
+
"grad_norm": 9.117968559265137,
|
| 50 |
+
"learning_rate": 7.563451776649747e-06,
|
| 51 |
+
"loss": 4.3132,
|
| 52 |
+
"step": 150
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"epoch": 0.8894536213468869,
|
| 56 |
+
"grad_norm": 14.240519523620605,
|
| 57 |
+
"learning_rate": 8.832487309644671e-06,
|
| 58 |
+
"loss": 4.1852,
|
| 59 |
+
"step": 175
|
| 60 |
+
},
|
| 61 |
+
{
|
| 62 |
+
"epoch": 1.0,
|
| 63 |
+
"eval_accuracy": 0.5022421524663677,
|
| 64 |
+
"eval_loss": 1.0154396295547485,
|
| 65 |
+
"eval_macro_f1": 0.40034331327123024,
|
| 66 |
+
"eval_runtime": 0.7355,
|
| 67 |
+
"eval_samples_per_second": 909.622,
|
| 68 |
+
"eval_steps_per_second": 57.106,
|
| 69 |
+
"eval_weighted_f1": 0.4039642637736501,
|
| 70 |
+
"step": 197
|
| 71 |
+
},
|
| 72 |
+
{
|
| 73 |
+
"epoch": 1.0152477763659467,
|
| 74 |
+
"grad_norm": 30.81084442138672,
|
| 75 |
+
"learning_rate": 1.0101522842639596e-05,
|
| 76 |
+
"loss": 4.0987,
|
| 77 |
+
"step": 200
|
| 78 |
+
},
|
| 79 |
+
{
|
| 80 |
+
"epoch": 1.1423125794155018,
|
| 81 |
+
"grad_norm": 37.35525131225586,
|
| 82 |
+
"learning_rate": 1.1370558375634519e-05,
|
| 83 |
+
"loss": 3.7559,
|
| 84 |
+
"step": 225
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 1.2693773824650572,
|
| 88 |
+
"grad_norm": 49.9312858581543,
|
| 89 |
+
"learning_rate": 1.2639593908629442e-05,
|
| 90 |
+
"loss": 3.433,
|
| 91 |
+
"step": 250
|
| 92 |
+
},
|
| 93 |
+
{
|
| 94 |
+
"epoch": 1.3964421855146125,
|
| 95 |
+
"grad_norm": 41.576908111572266,
|
| 96 |
+
"learning_rate": 1.3908629441624366e-05,
|
| 97 |
+
"loss": 2.9848,
|
| 98 |
+
"step": 275
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 1.5235069885641677,
|
| 102 |
+
"grad_norm": 38.61850357055664,
|
| 103 |
+
"learning_rate": 1.5177664974619291e-05,
|
| 104 |
+
"loss": 2.9432,
|
| 105 |
+
"step": 300
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 1.650571791613723,
|
| 109 |
+
"grad_norm": 55.122806549072266,
|
| 110 |
+
"learning_rate": 1.6446700507614214e-05,
|
| 111 |
+
"loss": 2.5225,
|
| 112 |
+
"step": 325
|
| 113 |
+
},
|
| 114 |
+
{
|
| 115 |
+
"epoch": 1.7776365946632784,
|
| 116 |
+
"grad_norm": 41.19192886352539,
|
| 117 |
+
"learning_rate": 1.771573604060914e-05,
|
| 118 |
+
"loss": 2.4043,
|
| 119 |
+
"step": 350
|
| 120 |
+
},
|
| 121 |
+
{
|
| 122 |
+
"epoch": 1.9047013977128335,
|
| 123 |
+
"grad_norm": 34.1766471862793,
|
| 124 |
+
"learning_rate": 1.8984771573604063e-05,
|
| 125 |
+
"loss": 2.6599,
|
| 126 |
+
"step": 375
|
| 127 |
+
},
|
| 128 |
+
{
|
| 129 |
+
"epoch": 2.0,
|
| 130 |
+
"eval_accuracy": 0.7877428998505231,
|
| 131 |
+
"eval_loss": 0.5306252241134644,
|
| 132 |
+
"eval_macro_f1": 0.7863585014643224,
|
| 133 |
+
"eval_runtime": 0.7237,
|
| 134 |
+
"eval_samples_per_second": 924.413,
|
| 135 |
+
"eval_steps_per_second": 58.035,
|
| 136 |
+
"eval_weighted_f1": 0.7866276798742975,
|
| 137 |
+
"step": 394
|
| 138 |
+
},
|
| 139 |
+
{
|
| 140 |
+
"epoch": 2.0304955527318933,
|
| 141 |
+
"grad_norm": 46.5556526184082,
|
| 142 |
+
"learning_rate": 1.997179921037789e-05,
|
| 143 |
+
"loss": 2.1544,
|
| 144 |
+
"step": 400
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 2.1575603557814484,
|
| 148 |
+
"grad_norm": 55.74009704589844,
|
| 149 |
+
"learning_rate": 1.9830795262267346e-05,
|
| 150 |
+
"loss": 2.0235,
|
| 151 |
+
"step": 425
|
| 152 |
+
},
|
| 153 |
+
{
|
| 154 |
+
"epoch": 2.2846251588310036,
|
| 155 |
+
"grad_norm": 31.618104934692383,
|
| 156 |
+
"learning_rate": 1.96897913141568e-05,
|
| 157 |
+
"loss": 2.1271,
|
| 158 |
+
"step": 450
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 2.411689961880559,
|
| 162 |
+
"grad_norm": 46.47744369506836,
|
| 163 |
+
"learning_rate": 1.954878736604625e-05,
|
| 164 |
+
"loss": 1.871,
|
| 165 |
+
"step": 475
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 2.5387547649301143,
|
| 169 |
+
"grad_norm": 58.39691925048828,
|
| 170 |
+
"learning_rate": 1.9407783417935703e-05,
|
| 171 |
+
"loss": 1.9686,
|
| 172 |
+
"step": 500
|
| 173 |
+
},
|
| 174 |
+
{
|
| 175 |
+
"epoch": 2.6658195679796695,
|
| 176 |
+
"grad_norm": 36.97040557861328,
|
| 177 |
+
"learning_rate": 1.9266779469825156e-05,
|
| 178 |
+
"loss": 2.0946,
|
| 179 |
+
"step": 525
|
| 180 |
+
},
|
| 181 |
+
{
|
| 182 |
+
"epoch": 2.792884371029225,
|
| 183 |
+
"grad_norm": 82.56703186035156,
|
| 184 |
+
"learning_rate": 1.912577552171461e-05,
|
| 185 |
+
"loss": 1.7246,
|
| 186 |
+
"step": 550
|
| 187 |
+
},
|
| 188 |
+
{
|
| 189 |
+
"epoch": 2.91994917407878,
|
| 190 |
+
"grad_norm": 57.654666900634766,
|
| 191 |
+
"learning_rate": 1.8984771573604063e-05,
|
| 192 |
+
"loss": 1.8446,
|
| 193 |
+
"step": 575
|
| 194 |
+
},
|
| 195 |
+
{
|
| 196 |
+
"epoch": 3.0,
|
| 197 |
+
"eval_accuracy": 0.8116591928251121,
|
| 198 |
+
"eval_loss": 0.5008328557014465,
|
| 199 |
+
"eval_macro_f1": 0.8100286300849188,
|
| 200 |
+
"eval_runtime": 0.7355,
|
| 201 |
+
"eval_samples_per_second": 909.561,
|
| 202 |
+
"eval_steps_per_second": 57.102,
|
| 203 |
+
"eval_weighted_f1": 0.810117109840787,
|
| 204 |
+
"step": 591
|
| 205 |
+
},
|
| 206 |
+
{
|
| 207 |
+
"epoch": 3.04574332909784,
|
| 208 |
+
"grad_norm": 49.310089111328125,
|
| 209 |
+
"learning_rate": 1.8843767625493517e-05,
|
| 210 |
+
"loss": 1.4955,
|
| 211 |
+
"step": 600
|
| 212 |
+
},
|
| 213 |
+
{
|
| 214 |
+
"epoch": 3.172808132147395,
|
| 215 |
+
"grad_norm": 49.757423400878906,
|
| 216 |
+
"learning_rate": 1.870276367738297e-05,
|
| 217 |
+
"loss": 1.2614,
|
| 218 |
+
"step": 625
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 3.2998729351969507,
|
| 222 |
+
"grad_norm": 50.10912322998047,
|
| 223 |
+
"learning_rate": 1.856175972927242e-05,
|
| 224 |
+
"loss": 1.2994,
|
| 225 |
+
"step": 650
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"epoch": 3.426937738246506,
|
| 229 |
+
"grad_norm": 28.961101531982422,
|
| 230 |
+
"learning_rate": 1.8420755781161873e-05,
|
| 231 |
+
"loss": 1.2444,
|
| 232 |
+
"step": 675
|
| 233 |
+
},
|
| 234 |
+
{
|
| 235 |
+
"epoch": 3.554002541296061,
|
| 236 |
+
"grad_norm": 118.08051300048828,
|
| 237 |
+
"learning_rate": 1.8279751833051327e-05,
|
| 238 |
+
"loss": 1.4413,
|
| 239 |
+
"step": 700
|
| 240 |
+
},
|
| 241 |
+
{
|
| 242 |
+
"epoch": 3.681067344345616,
|
| 243 |
+
"grad_norm": 36.147132873535156,
|
| 244 |
+
"learning_rate": 1.813874788494078e-05,
|
| 245 |
+
"loss": 1.2326,
|
| 246 |
+
"step": 725
|
| 247 |
+
},
|
| 248 |
+
{
|
| 249 |
+
"epoch": 3.8081321473951717,
|
| 250 |
+
"grad_norm": 46.774375915527344,
|
| 251 |
+
"learning_rate": 1.7997743936830234e-05,
|
| 252 |
+
"loss": 1.4243,
|
| 253 |
+
"step": 750
|
| 254 |
+
},
|
| 255 |
+
{
|
| 256 |
+
"epoch": 3.935196950444727,
|
| 257 |
+
"grad_norm": 37.9736442565918,
|
| 258 |
+
"learning_rate": 1.7856739988719684e-05,
|
| 259 |
+
"loss": 1.4571,
|
| 260 |
+
"step": 775
|
| 261 |
+
},
|
| 262 |
+
{
|
| 263 |
+
"epoch": 4.0,
|
| 264 |
+
"eval_accuracy": 0.8131539611360239,
|
| 265 |
+
"eval_loss": 0.4768521785736084,
|
| 266 |
+
"eval_macro_f1": 0.813189660695599,
|
| 267 |
+
"eval_runtime": 0.7501,
|
| 268 |
+
"eval_samples_per_second": 891.87,
|
| 269 |
+
"eval_steps_per_second": 55.992,
|
| 270 |
+
"eval_weighted_f1": 0.8132918645535413,
|
| 271 |
+
"step": 788
|
| 272 |
+
},
|
| 273 |
+
{
|
| 274 |
+
"epoch": 4.060991105463787,
|
| 275 |
+
"grad_norm": 34.35706329345703,
|
| 276 |
+
"learning_rate": 1.771573604060914e-05,
|
| 277 |
+
"loss": 1.1106,
|
| 278 |
+
"step": 800
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 4.188055908513342,
|
| 282 |
+
"grad_norm": 63.45062255859375,
|
| 283 |
+
"learning_rate": 1.7574732092498594e-05,
|
| 284 |
+
"loss": 0.8681,
|
| 285 |
+
"step": 825
|
| 286 |
+
},
|
| 287 |
+
{
|
| 288 |
+
"epoch": 4.315120711562897,
|
| 289 |
+
"grad_norm": 32.65141677856445,
|
| 290 |
+
"learning_rate": 1.7433728144388044e-05,
|
| 291 |
+
"loss": 0.873,
|
| 292 |
+
"step": 850
|
| 293 |
+
},
|
| 294 |
+
{
|
| 295 |
+
"epoch": 4.4421855146124525,
|
| 296 |
+
"grad_norm": 58.85293960571289,
|
| 297 |
+
"learning_rate": 1.7292724196277497e-05,
|
| 298 |
+
"loss": 0.901,
|
| 299 |
+
"step": 875
|
| 300 |
+
},
|
| 301 |
+
{
|
| 302 |
+
"epoch": 4.569250317662007,
|
| 303 |
+
"grad_norm": 29.3676815032959,
|
| 304 |
+
"learning_rate": 1.715172024816695e-05,
|
| 305 |
+
"loss": 0.9577,
|
| 306 |
+
"step": 900
|
| 307 |
+
},
|
| 308 |
+
{
|
| 309 |
+
"epoch": 4.696315120711563,
|
| 310 |
+
"grad_norm": 64.39431762695312,
|
| 311 |
+
"learning_rate": 1.7010716300056404e-05,
|
| 312 |
+
"loss": 0.8218,
|
| 313 |
+
"step": 925
|
| 314 |
+
},
|
| 315 |
+
{
|
| 316 |
+
"epoch": 4.823379923761118,
|
| 317 |
+
"grad_norm": 66.92719268798828,
|
| 318 |
+
"learning_rate": 1.6869712351945854e-05,
|
| 319 |
+
"loss": 1.1205,
|
| 320 |
+
"step": 950
|
| 321 |
+
},
|
| 322 |
+
{
|
| 323 |
+
"epoch": 4.950444726810673,
|
| 324 |
+
"grad_norm": 21.281335830688477,
|
| 325 |
+
"learning_rate": 1.672870840383531e-05,
|
| 326 |
+
"loss": 0.8389,
|
| 327 |
+
"step": 975
|
| 328 |
+
},
|
| 329 |
+
{
|
| 330 |
+
"epoch": 5.0,
|
| 331 |
+
"eval_accuracy": 0.8221225710014948,
|
| 332 |
+
"eval_loss": 0.5535773038864136,
|
| 333 |
+
"eval_macro_f1": 0.8226452050606753,
|
| 334 |
+
"eval_runtime": 0.7234,
|
| 335 |
+
"eval_samples_per_second": 924.812,
|
| 336 |
+
"eval_steps_per_second": 58.06,
|
| 337 |
+
"eval_weighted_f1": 0.8227244405613038,
|
| 338 |
+
"step": 985
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 5.076238881829733,
|
| 342 |
+
"grad_norm": 31.218128204345703,
|
| 343 |
+
"learning_rate": 1.6587704455724764e-05,
|
| 344 |
+
"loss": 0.7517,
|
| 345 |
+
"step": 1000
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 5.203303684879288,
|
| 349 |
+
"grad_norm": 81.97185516357422,
|
| 350 |
+
"learning_rate": 1.6446700507614214e-05,
|
| 351 |
+
"loss": 0.416,
|
| 352 |
+
"step": 1025
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 5.330368487928844,
|
| 356 |
+
"grad_norm": 44.270484924316406,
|
| 357 |
+
"learning_rate": 1.6305696559503667e-05,
|
| 358 |
+
"loss": 0.6863,
|
| 359 |
+
"step": 1050
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 5.457433290978399,
|
| 363 |
+
"grad_norm": 116.3810806274414,
|
| 364 |
+
"learning_rate": 1.616469261139312e-05,
|
| 365 |
+
"loss": 0.588,
|
| 366 |
+
"step": 1075
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 5.584498094027954,
|
| 370 |
+
"grad_norm": 52.35886764526367,
|
| 371 |
+
"learning_rate": 1.6023688663282574e-05,
|
| 372 |
+
"loss": 0.8374,
|
| 373 |
+
"step": 1100
|
| 374 |
+
},
|
| 375 |
+
{
|
| 376 |
+
"epoch": 5.71156289707751,
|
| 377 |
+
"grad_norm": 29.787914276123047,
|
| 378 |
+
"learning_rate": 1.5882684715172024e-05,
|
| 379 |
+
"loss": 0.618,
|
| 380 |
+
"step": 1125
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 5.838627700127065,
|
| 384 |
+
"grad_norm": 57.911075592041016,
|
| 385 |
+
"learning_rate": 1.5741680767061477e-05,
|
| 386 |
+
"loss": 0.6089,
|
| 387 |
+
"step": 1150
|
| 388 |
+
},
|
| 389 |
+
{
|
| 390 |
+
"epoch": 5.96569250317662,
|
| 391 |
+
"grad_norm": 19.362987518310547,
|
| 392 |
+
"learning_rate": 1.5600676818950934e-05,
|
| 393 |
+
"loss": 0.7111,
|
| 394 |
+
"step": 1175
|
| 395 |
+
},
|
| 396 |
+
{
|
| 397 |
+
"epoch": 6.0,
|
| 398 |
+
"eval_accuracy": 0.8266068759342302,
|
| 399 |
+
"eval_loss": 0.5884621143341064,
|
| 400 |
+
"eval_macro_f1": 0.8269879429990837,
|
| 401 |
+
"eval_runtime": 0.7328,
|
| 402 |
+
"eval_samples_per_second": 912.931,
|
| 403 |
+
"eval_steps_per_second": 57.314,
|
| 404 |
+
"eval_weighted_f1": 0.8270232273444624,
|
| 405 |
+
"step": 1182
|
| 406 |
+
},
|
| 407 |
+
{
|
| 408 |
+
"epoch": 6.09148665819568,
|
| 409 |
+
"grad_norm": 21.163572311401367,
|
| 410 |
+
"learning_rate": 1.5459672870840384e-05,
|
| 411 |
+
"loss": 0.4341,
|
| 412 |
+
"step": 1200
|
| 413 |
+
},
|
| 414 |
+
{
|
| 415 |
+
"epoch": 6.2185514612452355,
|
| 416 |
+
"grad_norm": 81.59049987792969,
|
| 417 |
+
"learning_rate": 1.5318668922729838e-05,
|
| 418 |
+
"loss": 0.4025,
|
| 419 |
+
"step": 1225
|
| 420 |
+
},
|
| 421 |
+
{
|
| 422 |
+
"epoch": 6.34561626429479,
|
| 423 |
+
"grad_norm": 25.52253532409668,
|
| 424 |
+
"learning_rate": 1.5177664974619291e-05,
|
| 425 |
+
"loss": 0.3291,
|
| 426 |
+
"step": 1250
|
| 427 |
+
},
|
| 428 |
+
{
|
| 429 |
+
"epoch": 6.472681067344346,
|
| 430 |
+
"grad_norm": 3.302086591720581,
|
| 431 |
+
"learning_rate": 1.5036661026508743e-05,
|
| 432 |
+
"loss": 0.3049,
|
| 433 |
+
"step": 1275
|
| 434 |
+
},
|
| 435 |
+
{
|
| 436 |
+
"epoch": 6.599745870393901,
|
| 437 |
+
"grad_norm": 17.74028205871582,
|
| 438 |
+
"learning_rate": 1.4895657078398196e-05,
|
| 439 |
+
"loss": 0.5148,
|
| 440 |
+
"step": 1300
|
| 441 |
+
},
|
| 442 |
+
{
|
| 443 |
+
"epoch": 6.726810673443456,
|
| 444 |
+
"grad_norm": 7.377353191375732,
|
| 445 |
+
"learning_rate": 1.4754653130287648e-05,
|
| 446 |
+
"loss": 0.4865,
|
| 447 |
+
"step": 1325
|
| 448 |
+
},
|
| 449 |
+
{
|
| 450 |
+
"epoch": 6.853875476493012,
|
| 451 |
+
"grad_norm": 17.74429702758789,
|
| 452 |
+
"learning_rate": 1.4613649182177103e-05,
|
| 453 |
+
"loss": 0.5307,
|
| 454 |
+
"step": 1350
|
| 455 |
+
},
|
| 456 |
+
{
|
| 457 |
+
"epoch": 6.980940279542566,
|
| 458 |
+
"grad_norm": 77.65316772460938,
|
| 459 |
+
"learning_rate": 1.4472645234066556e-05,
|
| 460 |
+
"loss": 0.4986,
|
| 461 |
+
"step": 1375
|
| 462 |
+
},
|
| 463 |
+
{
|
| 464 |
+
"epoch": 7.0,
|
| 465 |
+
"eval_accuracy": 0.8310911808669657,
|
| 466 |
+
"eval_loss": 0.7535898685455322,
|
| 467 |
+
"eval_macro_f1": 0.8296475120561762,
|
| 468 |
+
"eval_runtime": 0.7467,
|
| 469 |
+
"eval_samples_per_second": 895.984,
|
| 470 |
+
"eval_steps_per_second": 56.25,
|
| 471 |
+
"eval_weighted_f1": 0.8295744703501335,
|
| 472 |
+
"step": 1379
|
| 473 |
+
},
|
| 474 |
+
{
|
| 475 |
+
"epoch": 7.106734434561626,
|
| 476 |
+
"grad_norm": 47.726463317871094,
|
| 477 |
+
"learning_rate": 1.4331641285956008e-05,
|
| 478 |
+
"loss": 0.2898,
|
| 479 |
+
"step": 1400
|
| 480 |
+
},
|
| 481 |
+
{
|
| 482 |
+
"epoch": 7.233799237611182,
|
| 483 |
+
"grad_norm": 127.74728393554688,
|
| 484 |
+
"learning_rate": 1.4190637337845461e-05,
|
| 485 |
+
"loss": 0.3503,
|
| 486 |
+
"step": 1425
|
| 487 |
+
},
|
| 488 |
+
{
|
| 489 |
+
"epoch": 7.360864040660737,
|
| 490 |
+
"grad_norm": 72.09273529052734,
|
| 491 |
+
"learning_rate": 1.4049633389734913e-05,
|
| 492 |
+
"loss": 0.3364,
|
| 493 |
+
"step": 1450
|
| 494 |
+
},
|
| 495 |
+
{
|
| 496 |
+
"epoch": 7.487928843710292,
|
| 497 |
+
"grad_norm": 65.77955627441406,
|
| 498 |
+
"learning_rate": 1.3908629441624366e-05,
|
| 499 |
+
"loss": 0.3278,
|
| 500 |
+
"step": 1475
|
| 501 |
+
},
|
| 502 |
+
{
|
| 503 |
+
"epoch": 7.614993646759848,
|
| 504 |
+
"grad_norm": 32.90453338623047,
|
| 505 |
+
"learning_rate": 1.3767625493513818e-05,
|
| 506 |
+
"loss": 0.3833,
|
| 507 |
+
"step": 1500
|
| 508 |
+
},
|
| 509 |
+
{
|
| 510 |
+
"epoch": 7.742058449809403,
|
| 511 |
+
"grad_norm": 248.89613342285156,
|
| 512 |
+
"learning_rate": 1.3626621545403271e-05,
|
| 513 |
+
"loss": 0.391,
|
| 514 |
+
"step": 1525
|
| 515 |
+
},
|
| 516 |
+
{
|
| 517 |
+
"epoch": 7.869123252858958,
|
| 518 |
+
"grad_norm": 202.55787658691406,
|
| 519 |
+
"learning_rate": 1.3485617597292726e-05,
|
| 520 |
+
"loss": 0.1825,
|
| 521 |
+
"step": 1550
|
| 522 |
+
},
|
| 523 |
+
{
|
| 524 |
+
"epoch": 7.9961880559085134,
|
| 525 |
+
"grad_norm": 57.24322509765625,
|
| 526 |
+
"learning_rate": 1.3344613649182178e-05,
|
| 527 |
+
"loss": 0.4829,
|
| 528 |
+
"step": 1575
|
| 529 |
+
},
|
| 530 |
+
{
|
| 531 |
+
"epoch": 8.0,
|
| 532 |
+
"eval_accuracy": 0.820627802690583,
|
| 533 |
+
"eval_loss": 0.9836516976356506,
|
| 534 |
+
"eval_macro_f1": 0.8213745088846777,
|
| 535 |
+
"eval_runtime": 0.728,
|
| 536 |
+
"eval_samples_per_second": 918.916,
|
| 537 |
+
"eval_steps_per_second": 57.69,
|
| 538 |
+
"eval_weighted_f1": 0.8213917255816803,
|
| 539 |
+
"step": 1576
|
| 540 |
+
},
|
| 541 |
+
{
|
| 542 |
+
"epoch": 8.121982210927573,
|
| 543 |
+
"grad_norm": 1.8322113752365112,
|
| 544 |
+
"learning_rate": 1.3203609701071631e-05,
|
| 545 |
+
"loss": 0.1437,
|
| 546 |
+
"step": 1600
|
| 547 |
+
},
|
| 548 |
+
{
|
| 549 |
+
"epoch": 8.249047013977128,
|
| 550 |
+
"grad_norm": 178.53717041015625,
|
| 551 |
+
"learning_rate": 1.3062605752961083e-05,
|
| 552 |
+
"loss": 0.2542,
|
| 553 |
+
"step": 1625
|
| 554 |
+
},
|
| 555 |
+
{
|
| 556 |
+
"epoch": 8.376111817026684,
|
| 557 |
+
"grad_norm": 157.9364471435547,
|
| 558 |
+
"learning_rate": 1.2921601804850537e-05,
|
| 559 |
+
"loss": 0.3173,
|
| 560 |
+
"step": 1650
|
| 561 |
+
},
|
| 562 |
+
{
|
| 563 |
+
"epoch": 8.503176620076239,
|
| 564 |
+
"grad_norm": 93.94173431396484,
|
| 565 |
+
"learning_rate": 1.278059785673999e-05,
|
| 566 |
+
"loss": 0.3372,
|
| 567 |
+
"step": 1675
|
| 568 |
+
},
|
| 569 |
+
{
|
| 570 |
+
"epoch": 8.630241423125794,
|
| 571 |
+
"grad_norm": 47.48844528198242,
|
| 572 |
+
"learning_rate": 1.2639593908629442e-05,
|
| 573 |
+
"loss": 0.3196,
|
| 574 |
+
"step": 1700
|
| 575 |
+
},
|
| 576 |
+
{
|
| 577 |
+
"epoch": 8.757306226175348,
|
| 578 |
+
"grad_norm": 0.2058953493833542,
|
| 579 |
+
"learning_rate": 1.2498589960518897e-05,
|
| 580 |
+
"loss": 0.3344,
|
| 581 |
+
"step": 1725
|
| 582 |
+
},
|
| 583 |
+
{
|
| 584 |
+
"epoch": 8.884371029224905,
|
| 585 |
+
"grad_norm": 0.08445502817630768,
|
| 586 |
+
"learning_rate": 1.2357586012408348e-05,
|
| 587 |
+
"loss": 0.3248,
|
| 588 |
+
"step": 1750
|
| 589 |
+
},
|
| 590 |
+
{
|
| 591 |
+
"epoch": 9.0,
|
| 592 |
+
"eval_accuracy": 0.8310911808669657,
|
| 593 |
+
"eval_loss": 0.9798506498336792,
|
| 594 |
+
"eval_macro_f1": 0.8315531652064146,
|
| 595 |
+
"eval_runtime": 0.7324,
|
| 596 |
+
"eval_samples_per_second": 913.443,
|
| 597 |
+
"eval_steps_per_second": 57.346,
|
| 598 |
+
"eval_weighted_f1": 0.8315146655390288,
|
| 599 |
+
"step": 1773
|
| 600 |
+
},
|
| 601 |
+
{
|
| 602 |
+
"epoch": 9.010165184243965,
|
| 603 |
+
"grad_norm": 107.16582489013672,
|
| 604 |
+
"learning_rate": 1.2216582064297802e-05,
|
| 605 |
+
"loss": 0.3178,
|
| 606 |
+
"step": 1775
|
| 607 |
+
},
|
| 608 |
+
{
|
| 609 |
+
"epoch": 9.13722998729352,
|
| 610 |
+
"grad_norm": 1.8582719564437866,
|
| 611 |
+
"learning_rate": 1.2075578116187253e-05,
|
| 612 |
+
"loss": 0.1502,
|
| 613 |
+
"step": 1800
|
| 614 |
+
},
|
| 615 |
+
{
|
| 616 |
+
"epoch": 9.264294790343074,
|
| 617 |
+
"grad_norm": 0.08437829464673996,
|
| 618 |
+
"learning_rate": 1.1934574168076707e-05,
|
| 619 |
+
"loss": 0.1494,
|
| 620 |
+
"step": 1825
|
| 621 |
+
},
|
| 622 |
+
{
|
| 623 |
+
"epoch": 9.39135959339263,
|
| 624 |
+
"grad_norm": 2.747943639755249,
|
| 625 |
+
"learning_rate": 1.179357021996616e-05,
|
| 626 |
+
"loss": 0.1686,
|
| 627 |
+
"step": 1850
|
| 628 |
+
},
|
| 629 |
+
{
|
| 630 |
+
"epoch": 9.518424396442185,
|
| 631 |
+
"grad_norm": 0.07697977870702744,
|
| 632 |
+
"learning_rate": 1.1652566271855612e-05,
|
| 633 |
+
"loss": 0.3106,
|
| 634 |
+
"step": 1875
|
| 635 |
+
},
|
| 636 |
+
{
|
| 637 |
+
"epoch": 9.64548919949174,
|
| 638 |
+
"grad_norm": 139.0595245361328,
|
| 639 |
+
"learning_rate": 1.1511562323745067e-05,
|
| 640 |
+
"loss": 0.3035,
|
| 641 |
+
"step": 1900
|
| 642 |
+
},
|
| 643 |
+
{
|
| 644 |
+
"epoch": 9.772554002541296,
|
| 645 |
+
"grad_norm": 0.6565935015678406,
|
| 646 |
+
"learning_rate": 1.1370558375634519e-05,
|
| 647 |
+
"loss": 0.1772,
|
| 648 |
+
"step": 1925
|
| 649 |
+
},
|
| 650 |
+
{
|
| 651 |
+
"epoch": 9.899618805590851,
|
| 652 |
+
"grad_norm": 0.24950720369815826,
|
| 653 |
+
"learning_rate": 1.1229554427523972e-05,
|
| 654 |
+
"loss": 0.2652,
|
| 655 |
+
"step": 1950
|
| 656 |
+
},
|
| 657 |
+
{
|
| 658 |
+
"epoch": 10.0,
|
| 659 |
+
"eval_accuracy": 0.8340807174887892,
|
| 660 |
+
"eval_loss": 1.067696213722229,
|
| 661 |
+
"eval_macro_f1": 0.8345702219772342,
|
| 662 |
+
"eval_runtime": 0.7361,
|
| 663 |
+
"eval_samples_per_second": 908.826,
|
| 664 |
+
"eval_steps_per_second": 57.056,
|
| 665 |
+
"eval_weighted_f1": 0.8345086622837158,
|
| 666 |
+
"step": 1970
|
| 667 |
+
},
|
| 668 |
+
{
|
| 669 |
+
"epoch": 10.025412960609911,
|
| 670 |
+
"grad_norm": 0.5611938238143921,
|
| 671 |
+
"learning_rate": 1.1088550479413425e-05,
|
| 672 |
+
"loss": 0.1058,
|
| 673 |
+
"step": 1975
|
| 674 |
+
},
|
| 675 |
+
{
|
| 676 |
+
"epoch": 10.152477763659466,
|
| 677 |
+
"grad_norm": 159.5753631591797,
|
| 678 |
+
"learning_rate": 1.0947546531302877e-05,
|
| 679 |
+
"loss": 0.3395,
|
| 680 |
+
"step": 2000
|
| 681 |
+
},
|
| 682 |
+
{
|
| 683 |
+
"epoch": 10.279542566709022,
|
| 684 |
+
"grad_norm": 0.11062340438365936,
|
| 685 |
+
"learning_rate": 1.080654258319233e-05,
|
| 686 |
+
"loss": 0.2162,
|
| 687 |
+
"step": 2025
|
| 688 |
+
},
|
| 689 |
+
{
|
| 690 |
+
"epoch": 10.406607369758577,
|
| 691 |
+
"grad_norm": 125.50062561035156,
|
| 692 |
+
"learning_rate": 1.0665538635081782e-05,
|
| 693 |
+
"loss": 0.2631,
|
| 694 |
+
"step": 2050
|
| 695 |
+
},
|
| 696 |
+
{
|
| 697 |
+
"epoch": 10.533672172808132,
|
| 698 |
+
"grad_norm": 0.30203333497047424,
|
| 699 |
+
"learning_rate": 1.0524534686971236e-05,
|
| 700 |
+
"loss": 0.2117,
|
| 701 |
+
"step": 2075
|
| 702 |
+
},
|
| 703 |
+
{
|
| 704 |
+
"epoch": 10.660736975857688,
|
| 705 |
+
"grad_norm": 0.17817461490631104,
|
| 706 |
+
"learning_rate": 1.038353073886069e-05,
|
| 707 |
+
"loss": 0.1672,
|
| 708 |
+
"step": 2100
|
| 709 |
+
},
|
| 710 |
+
{
|
| 711 |
+
"epoch": 10.787801778907243,
|
| 712 |
+
"grad_norm": 528.10791015625,
|
| 713 |
+
"learning_rate": 1.0242526790750142e-05,
|
| 714 |
+
"loss": 0.2332,
|
| 715 |
+
"step": 2125
|
| 716 |
+
},
|
| 717 |
+
{
|
| 718 |
+
"epoch": 10.914866581956797,
|
| 719 |
+
"grad_norm": 111.58641052246094,
|
| 720 |
+
"learning_rate": 1.0101522842639596e-05,
|
| 721 |
+
"loss": 0.1683,
|
| 722 |
+
"step": 2150
|
| 723 |
+
},
|
| 724 |
+
{
|
| 725 |
+
"epoch": 11.0,
|
| 726 |
+
"eval_accuracy": 0.8415545590433483,
|
| 727 |
+
"eval_loss": 1.0293971300125122,
|
| 728 |
+
"eval_macro_f1": 0.8423839833625967,
|
| 729 |
+
"eval_runtime": 0.7521,
|
| 730 |
+
"eval_samples_per_second": 889.519,
|
| 731 |
+
"eval_steps_per_second": 55.844,
|
| 732 |
+
"eval_weighted_f1": 0.8422944474828612,
|
| 733 |
+
"step": 2167
|
| 734 |
+
}
|
| 735 |
+
],
|
| 736 |
+
"logging_steps": 25,
|
| 737 |
+
"max_steps": 3940,
|
| 738 |
+
"num_input_tokens_seen": 0,
|
| 739 |
+
"num_train_epochs": 20,
|
| 740 |
+
"save_steps": 500,
|
| 741 |
+
"stateful_callbacks": {
|
| 742 |
+
"EarlyStoppingCallback": {
|
| 743 |
+
"args": {
|
| 744 |
+
"early_stopping_patience": 3,
|
| 745 |
+
"early_stopping_threshold": 0.0
|
| 746 |
+
},
|
| 747 |
+
"attributes": {
|
| 748 |
+
"early_stopping_patience_counter": 0
|
| 749 |
+
}
|
| 750 |
+
},
|
| 751 |
+
"TrainerControl": {
|
| 752 |
+
"args": {
|
| 753 |
+
"should_epoch_stop": false,
|
| 754 |
+
"should_evaluate": false,
|
| 755 |
+
"should_log": false,
|
| 756 |
+
"should_save": true,
|
| 757 |
+
"should_training_stop": false
|
| 758 |
+
},
|
| 759 |
+
"attributes": {}
|
| 760 |
+
}
|
| 761 |
+
},
|
| 762 |
+
"total_flos": 0.0,
|
| 763 |
+
"train_batch_size": 8,
|
| 764 |
+
"trial_name": null,
|
| 765 |
+
"trial_params": null
|
| 766 |
+
}
|
ood-vifactcheck-bamibert-topic-domain-07-s43-20260827T195828Z-3858bb98/trainer/checkpoint-2758/special_tokens_map.json
ADDED
|
@@ -0,0 +1,51 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": {
|
| 3 |
+
"content": "<s>",
|
| 4 |
+
"lstrip": false,
|
| 5 |
+
"normalized": false,
|
| 6 |
+
"rstrip": false,
|
| 7 |
+
"single_word": false
|
| 8 |
+
},
|
| 9 |
+
"cls_token": {
|
| 10 |
+
"content": "<s>",
|
| 11 |
+
"lstrip": false,
|
| 12 |
+
"normalized": false,
|
| 13 |
+
"rstrip": false,
|
| 14 |
+
"single_word": false
|
| 15 |
+
},
|
| 16 |
+
"eos_token": {
|
| 17 |
+
"content": "</s>",
|
| 18 |
+
"lstrip": false,
|
| 19 |
+
"normalized": false,
|
| 20 |
+
"rstrip": false,
|
| 21 |
+
"single_word": false
|
| 22 |
+
},
|
| 23 |
+
"mask_token": {
|
| 24 |
+
"content": "<mask>",
|
| 25 |
+
"lstrip": false,
|
| 26 |
+
"normalized": false,
|
| 27 |
+
"rstrip": false,
|
| 28 |
+
"single_word": false
|
| 29 |
+
},
|
| 30 |
+
"pad_token": {
|
| 31 |
+
"content": "<pad>",
|
| 32 |
+
"lstrip": false,
|
| 33 |
+
"normalized": false,
|
| 34 |
+
"rstrip": false,
|
| 35 |
+
"single_word": false
|
| 36 |
+
},
|
| 37 |
+
"sep_token": {
|
| 38 |
+
"content": "</s>",
|
| 39 |
+
"lstrip": false,
|
| 40 |
+
"normalized": false,
|
| 41 |
+
"rstrip": false,
|
| 42 |
+
"single_word": false
|
| 43 |
+
},
|
| 44 |
+
"unk_token": {
|
| 45 |
+
"content": "<unk>",
|
| 46 |
+
"lstrip": false,
|
| 47 |
+
"normalized": false,
|
| 48 |
+
"rstrip": false,
|
| 49 |
+
"single_word": false
|
| 50 |
+
}
|
| 51 |
+
}
|
ood-vifactcheck-bamibert-topic-domain-07-s43-20260827T195828Z-3858bb98/trainer/checkpoint-2758/tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-bamibert-topic-domain-07-s43-20260827T195828Z-3858bb98/trainer/checkpoint-2758/tokenizer_config.json
ADDED
|
@@ -0,0 +1,55 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"added_tokens_decoder": {
|
| 3 |
+
"0": {
|
| 4 |
+
"content": "<s>",
|
| 5 |
+
"lstrip": false,
|
| 6 |
+
"normalized": false,
|
| 7 |
+
"rstrip": false,
|
| 8 |
+
"single_word": false,
|
| 9 |
+
"special": true
|
| 10 |
+
},
|
| 11 |
+
"1": {
|
| 12 |
+
"content": "<pad>",
|
| 13 |
+
"lstrip": false,
|
| 14 |
+
"normalized": false,
|
| 15 |
+
"rstrip": false,
|
| 16 |
+
"single_word": false,
|
| 17 |
+
"special": true
|
| 18 |
+
},
|
| 19 |
+
"2": {
|
| 20 |
+
"content": "</s>",
|
| 21 |
+
"lstrip": false,
|
| 22 |
+
"normalized": false,
|
| 23 |
+
"rstrip": false,
|
| 24 |
+
"single_word": false,
|
| 25 |
+
"special": true
|
| 26 |
+
},
|
| 27 |
+
"3": {
|
| 28 |
+
"content": "<unk>",
|
| 29 |
+
"lstrip": false,
|
| 30 |
+
"normalized": false,
|
| 31 |
+
"rstrip": false,
|
| 32 |
+
"single_word": false,
|
| 33 |
+
"special": true
|
| 34 |
+
},
|
| 35 |
+
"20480": {
|
| 36 |
+
"content": "<mask>",
|
| 37 |
+
"lstrip": false,
|
| 38 |
+
"normalized": false,
|
| 39 |
+
"rstrip": false,
|
| 40 |
+
"single_word": false,
|
| 41 |
+
"special": true
|
| 42 |
+
}
|
| 43 |
+
},
|
| 44 |
+
"bos_token": "<s>",
|
| 45 |
+
"clean_up_tokenization_spaces": false,
|
| 46 |
+
"cls_token": "<s>",
|
| 47 |
+
"eos_token": "</s>",
|
| 48 |
+
"extra_special_tokens": {},
|
| 49 |
+
"mask_token": "<mask>",
|
| 50 |
+
"model_max_length": 2048,
|
| 51 |
+
"pad_token": "<pad>",
|
| 52 |
+
"sep_token": "</s>",
|
| 53 |
+
"tokenizer_class": "XLMRobertaTokenizerFast",
|
| 54 |
+
"unk_token": "<unk>"
|
| 55 |
+
}
|
ood-vifactcheck-bamibert-topic-domain-07-s43-20260827T195828Z-3858bb98/trainer/checkpoint-2758/trainer_state.json
ADDED
|
@@ -0,0 +1,967 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 2167,
|
| 3 |
+
"best_metric": 0.8423839833625967,
|
| 4 |
+
"best_model_checkpoint": "/root/VNEncoder/results/checkpoints/ood-vifactcheck-bamibert-topic-domain-07-s43-20260827T195828Z-3858bb98/trainer/checkpoint-2167",
|
| 5 |
+
"epoch": 14.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 2758,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.12706480304955528,
|
| 14 |
+
"grad_norm": 29.48038673400879,
|
| 15 |
+
"learning_rate": 1.218274111675127e-06,
|
| 16 |
+
"loss": 4.4267,
|
| 17 |
+
"step": 25
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.25412960609911056,
|
| 21 |
+
"grad_norm": 14.88051700592041,
|
| 22 |
+
"learning_rate": 2.487309644670051e-06,
|
| 23 |
+
"loss": 4.4197,
|
| 24 |
+
"step": 50
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.3811944091486658,
|
| 28 |
+
"grad_norm": 6.070577144622803,
|
| 29 |
+
"learning_rate": 3.756345177664975e-06,
|
| 30 |
+
"loss": 4.4149,
|
| 31 |
+
"step": 75
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.5082592121982211,
|
| 35 |
+
"grad_norm": 7.696415901184082,
|
| 36 |
+
"learning_rate": 5.025380710659899e-06,
|
| 37 |
+
"loss": 4.3894,
|
| 38 |
+
"step": 100
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.6353240152477764,
|
| 42 |
+
"grad_norm": 52.55864715576172,
|
| 43 |
+
"learning_rate": 6.2944162436548225e-06,
|
| 44 |
+
"loss": 4.3573,
|
| 45 |
+
"step": 125
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.7623888182973316,
|
| 49 |
+
"grad_norm": 9.117968559265137,
|
| 50 |
+
"learning_rate": 7.563451776649747e-06,
|
| 51 |
+
"loss": 4.3132,
|
| 52 |
+
"step": 150
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"epoch": 0.8894536213468869,
|
| 56 |
+
"grad_norm": 14.240519523620605,
|
| 57 |
+
"learning_rate": 8.832487309644671e-06,
|
| 58 |
+
"loss": 4.1852,
|
| 59 |
+
"step": 175
|
| 60 |
+
},
|
| 61 |
+
{
|
| 62 |
+
"epoch": 1.0,
|
| 63 |
+
"eval_accuracy": 0.5022421524663677,
|
| 64 |
+
"eval_loss": 1.0154396295547485,
|
| 65 |
+
"eval_macro_f1": 0.40034331327123024,
|
| 66 |
+
"eval_runtime": 0.7355,
|
| 67 |
+
"eval_samples_per_second": 909.622,
|
| 68 |
+
"eval_steps_per_second": 57.106,
|
| 69 |
+
"eval_weighted_f1": 0.4039642637736501,
|
| 70 |
+
"step": 197
|
| 71 |
+
},
|
| 72 |
+
{
|
| 73 |
+
"epoch": 1.0152477763659467,
|
| 74 |
+
"grad_norm": 30.81084442138672,
|
| 75 |
+
"learning_rate": 1.0101522842639596e-05,
|
| 76 |
+
"loss": 4.0987,
|
| 77 |
+
"step": 200
|
| 78 |
+
},
|
| 79 |
+
{
|
| 80 |
+
"epoch": 1.1423125794155018,
|
| 81 |
+
"grad_norm": 37.35525131225586,
|
| 82 |
+
"learning_rate": 1.1370558375634519e-05,
|
| 83 |
+
"loss": 3.7559,
|
| 84 |
+
"step": 225
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 1.2693773824650572,
|
| 88 |
+
"grad_norm": 49.9312858581543,
|
| 89 |
+
"learning_rate": 1.2639593908629442e-05,
|
| 90 |
+
"loss": 3.433,
|
| 91 |
+
"step": 250
|
| 92 |
+
},
|
| 93 |
+
{
|
| 94 |
+
"epoch": 1.3964421855146125,
|
| 95 |
+
"grad_norm": 41.576908111572266,
|
| 96 |
+
"learning_rate": 1.3908629441624366e-05,
|
| 97 |
+
"loss": 2.9848,
|
| 98 |
+
"step": 275
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 1.5235069885641677,
|
| 102 |
+
"grad_norm": 38.61850357055664,
|
| 103 |
+
"learning_rate": 1.5177664974619291e-05,
|
| 104 |
+
"loss": 2.9432,
|
| 105 |
+
"step": 300
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 1.650571791613723,
|
| 109 |
+
"grad_norm": 55.122806549072266,
|
| 110 |
+
"learning_rate": 1.6446700507614214e-05,
|
| 111 |
+
"loss": 2.5225,
|
| 112 |
+
"step": 325
|
| 113 |
+
},
|
| 114 |
+
{
|
| 115 |
+
"epoch": 1.7776365946632784,
|
| 116 |
+
"grad_norm": 41.19192886352539,
|
| 117 |
+
"learning_rate": 1.771573604060914e-05,
|
| 118 |
+
"loss": 2.4043,
|
| 119 |
+
"step": 350
|
| 120 |
+
},
|
| 121 |
+
{
|
| 122 |
+
"epoch": 1.9047013977128335,
|
| 123 |
+
"grad_norm": 34.1766471862793,
|
| 124 |
+
"learning_rate": 1.8984771573604063e-05,
|
| 125 |
+
"loss": 2.6599,
|
| 126 |
+
"step": 375
|
| 127 |
+
},
|
| 128 |
+
{
|
| 129 |
+
"epoch": 2.0,
|
| 130 |
+
"eval_accuracy": 0.7877428998505231,
|
| 131 |
+
"eval_loss": 0.5306252241134644,
|
| 132 |
+
"eval_macro_f1": 0.7863585014643224,
|
| 133 |
+
"eval_runtime": 0.7237,
|
| 134 |
+
"eval_samples_per_second": 924.413,
|
| 135 |
+
"eval_steps_per_second": 58.035,
|
| 136 |
+
"eval_weighted_f1": 0.7866276798742975,
|
| 137 |
+
"step": 394
|
| 138 |
+
},
|
| 139 |
+
{
|
| 140 |
+
"epoch": 2.0304955527318933,
|
| 141 |
+
"grad_norm": 46.5556526184082,
|
| 142 |
+
"learning_rate": 1.997179921037789e-05,
|
| 143 |
+
"loss": 2.1544,
|
| 144 |
+
"step": 400
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 2.1575603557814484,
|
| 148 |
+
"grad_norm": 55.74009704589844,
|
| 149 |
+
"learning_rate": 1.9830795262267346e-05,
|
| 150 |
+
"loss": 2.0235,
|
| 151 |
+
"step": 425
|
| 152 |
+
},
|
| 153 |
+
{
|
| 154 |
+
"epoch": 2.2846251588310036,
|
| 155 |
+
"grad_norm": 31.618104934692383,
|
| 156 |
+
"learning_rate": 1.96897913141568e-05,
|
| 157 |
+
"loss": 2.1271,
|
| 158 |
+
"step": 450
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 2.411689961880559,
|
| 162 |
+
"grad_norm": 46.47744369506836,
|
| 163 |
+
"learning_rate": 1.954878736604625e-05,
|
| 164 |
+
"loss": 1.871,
|
| 165 |
+
"step": 475
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 2.5387547649301143,
|
| 169 |
+
"grad_norm": 58.39691925048828,
|
| 170 |
+
"learning_rate": 1.9407783417935703e-05,
|
| 171 |
+
"loss": 1.9686,
|
| 172 |
+
"step": 500
|
| 173 |
+
},
|
| 174 |
+
{
|
| 175 |
+
"epoch": 2.6658195679796695,
|
| 176 |
+
"grad_norm": 36.97040557861328,
|
| 177 |
+
"learning_rate": 1.9266779469825156e-05,
|
| 178 |
+
"loss": 2.0946,
|
| 179 |
+
"step": 525
|
| 180 |
+
},
|
| 181 |
+
{
|
| 182 |
+
"epoch": 2.792884371029225,
|
| 183 |
+
"grad_norm": 82.56703186035156,
|
| 184 |
+
"learning_rate": 1.912577552171461e-05,
|
| 185 |
+
"loss": 1.7246,
|
| 186 |
+
"step": 550
|
| 187 |
+
},
|
| 188 |
+
{
|
| 189 |
+
"epoch": 2.91994917407878,
|
| 190 |
+
"grad_norm": 57.654666900634766,
|
| 191 |
+
"learning_rate": 1.8984771573604063e-05,
|
| 192 |
+
"loss": 1.8446,
|
| 193 |
+
"step": 575
|
| 194 |
+
},
|
| 195 |
+
{
|
| 196 |
+
"epoch": 3.0,
|
| 197 |
+
"eval_accuracy": 0.8116591928251121,
|
| 198 |
+
"eval_loss": 0.5008328557014465,
|
| 199 |
+
"eval_macro_f1": 0.8100286300849188,
|
| 200 |
+
"eval_runtime": 0.7355,
|
| 201 |
+
"eval_samples_per_second": 909.561,
|
| 202 |
+
"eval_steps_per_second": 57.102,
|
| 203 |
+
"eval_weighted_f1": 0.810117109840787,
|
| 204 |
+
"step": 591
|
| 205 |
+
},
|
| 206 |
+
{
|
| 207 |
+
"epoch": 3.04574332909784,
|
| 208 |
+
"grad_norm": 49.310089111328125,
|
| 209 |
+
"learning_rate": 1.8843767625493517e-05,
|
| 210 |
+
"loss": 1.4955,
|
| 211 |
+
"step": 600
|
| 212 |
+
},
|
| 213 |
+
{
|
| 214 |
+
"epoch": 3.172808132147395,
|
| 215 |
+
"grad_norm": 49.757423400878906,
|
| 216 |
+
"learning_rate": 1.870276367738297e-05,
|
| 217 |
+
"loss": 1.2614,
|
| 218 |
+
"step": 625
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 3.2998729351969507,
|
| 222 |
+
"grad_norm": 50.10912322998047,
|
| 223 |
+
"learning_rate": 1.856175972927242e-05,
|
| 224 |
+
"loss": 1.2994,
|
| 225 |
+
"step": 650
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"epoch": 3.426937738246506,
|
| 229 |
+
"grad_norm": 28.961101531982422,
|
| 230 |
+
"learning_rate": 1.8420755781161873e-05,
|
| 231 |
+
"loss": 1.2444,
|
| 232 |
+
"step": 675
|
| 233 |
+
},
|
| 234 |
+
{
|
| 235 |
+
"epoch": 3.554002541296061,
|
| 236 |
+
"grad_norm": 118.08051300048828,
|
| 237 |
+
"learning_rate": 1.8279751833051327e-05,
|
| 238 |
+
"loss": 1.4413,
|
| 239 |
+
"step": 700
|
| 240 |
+
},
|
| 241 |
+
{
|
| 242 |
+
"epoch": 3.681067344345616,
|
| 243 |
+
"grad_norm": 36.147132873535156,
|
| 244 |
+
"learning_rate": 1.813874788494078e-05,
|
| 245 |
+
"loss": 1.2326,
|
| 246 |
+
"step": 725
|
| 247 |
+
},
|
| 248 |
+
{
|
| 249 |
+
"epoch": 3.8081321473951717,
|
| 250 |
+
"grad_norm": 46.774375915527344,
|
| 251 |
+
"learning_rate": 1.7997743936830234e-05,
|
| 252 |
+
"loss": 1.4243,
|
| 253 |
+
"step": 750
|
| 254 |
+
},
|
| 255 |
+
{
|
| 256 |
+
"epoch": 3.935196950444727,
|
| 257 |
+
"grad_norm": 37.9736442565918,
|
| 258 |
+
"learning_rate": 1.7856739988719684e-05,
|
| 259 |
+
"loss": 1.4571,
|
| 260 |
+
"step": 775
|
| 261 |
+
},
|
| 262 |
+
{
|
| 263 |
+
"epoch": 4.0,
|
| 264 |
+
"eval_accuracy": 0.8131539611360239,
|
| 265 |
+
"eval_loss": 0.4768521785736084,
|
| 266 |
+
"eval_macro_f1": 0.813189660695599,
|
| 267 |
+
"eval_runtime": 0.7501,
|
| 268 |
+
"eval_samples_per_second": 891.87,
|
| 269 |
+
"eval_steps_per_second": 55.992,
|
| 270 |
+
"eval_weighted_f1": 0.8132918645535413,
|
| 271 |
+
"step": 788
|
| 272 |
+
},
|
| 273 |
+
{
|
| 274 |
+
"epoch": 4.060991105463787,
|
| 275 |
+
"grad_norm": 34.35706329345703,
|
| 276 |
+
"learning_rate": 1.771573604060914e-05,
|
| 277 |
+
"loss": 1.1106,
|
| 278 |
+
"step": 800
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 4.188055908513342,
|
| 282 |
+
"grad_norm": 63.45062255859375,
|
| 283 |
+
"learning_rate": 1.7574732092498594e-05,
|
| 284 |
+
"loss": 0.8681,
|
| 285 |
+
"step": 825
|
| 286 |
+
},
|
| 287 |
+
{
|
| 288 |
+
"epoch": 4.315120711562897,
|
| 289 |
+
"grad_norm": 32.65141677856445,
|
| 290 |
+
"learning_rate": 1.7433728144388044e-05,
|
| 291 |
+
"loss": 0.873,
|
| 292 |
+
"step": 850
|
| 293 |
+
},
|
| 294 |
+
{
|
| 295 |
+
"epoch": 4.4421855146124525,
|
| 296 |
+
"grad_norm": 58.85293960571289,
|
| 297 |
+
"learning_rate": 1.7292724196277497e-05,
|
| 298 |
+
"loss": 0.901,
|
| 299 |
+
"step": 875
|
| 300 |
+
},
|
| 301 |
+
{
|
| 302 |
+
"epoch": 4.569250317662007,
|
| 303 |
+
"grad_norm": 29.3676815032959,
|
| 304 |
+
"learning_rate": 1.715172024816695e-05,
|
| 305 |
+
"loss": 0.9577,
|
| 306 |
+
"step": 900
|
| 307 |
+
},
|
| 308 |
+
{
|
| 309 |
+
"epoch": 4.696315120711563,
|
| 310 |
+
"grad_norm": 64.39431762695312,
|
| 311 |
+
"learning_rate": 1.7010716300056404e-05,
|
| 312 |
+
"loss": 0.8218,
|
| 313 |
+
"step": 925
|
| 314 |
+
},
|
| 315 |
+
{
|
| 316 |
+
"epoch": 4.823379923761118,
|
| 317 |
+
"grad_norm": 66.92719268798828,
|
| 318 |
+
"learning_rate": 1.6869712351945854e-05,
|
| 319 |
+
"loss": 1.1205,
|
| 320 |
+
"step": 950
|
| 321 |
+
},
|
| 322 |
+
{
|
| 323 |
+
"epoch": 4.950444726810673,
|
| 324 |
+
"grad_norm": 21.281335830688477,
|
| 325 |
+
"learning_rate": 1.672870840383531e-05,
|
| 326 |
+
"loss": 0.8389,
|
| 327 |
+
"step": 975
|
| 328 |
+
},
|
| 329 |
+
{
|
| 330 |
+
"epoch": 5.0,
|
| 331 |
+
"eval_accuracy": 0.8221225710014948,
|
| 332 |
+
"eval_loss": 0.5535773038864136,
|
| 333 |
+
"eval_macro_f1": 0.8226452050606753,
|
| 334 |
+
"eval_runtime": 0.7234,
|
| 335 |
+
"eval_samples_per_second": 924.812,
|
| 336 |
+
"eval_steps_per_second": 58.06,
|
| 337 |
+
"eval_weighted_f1": 0.8227244405613038,
|
| 338 |
+
"step": 985
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 5.076238881829733,
|
| 342 |
+
"grad_norm": 31.218128204345703,
|
| 343 |
+
"learning_rate": 1.6587704455724764e-05,
|
| 344 |
+
"loss": 0.7517,
|
| 345 |
+
"step": 1000
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 5.203303684879288,
|
| 349 |
+
"grad_norm": 81.97185516357422,
|
| 350 |
+
"learning_rate": 1.6446700507614214e-05,
|
| 351 |
+
"loss": 0.416,
|
| 352 |
+
"step": 1025
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 5.330368487928844,
|
| 356 |
+
"grad_norm": 44.270484924316406,
|
| 357 |
+
"learning_rate": 1.6305696559503667e-05,
|
| 358 |
+
"loss": 0.6863,
|
| 359 |
+
"step": 1050
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 5.457433290978399,
|
| 363 |
+
"grad_norm": 116.3810806274414,
|
| 364 |
+
"learning_rate": 1.616469261139312e-05,
|
| 365 |
+
"loss": 0.588,
|
| 366 |
+
"step": 1075
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 5.584498094027954,
|
| 370 |
+
"grad_norm": 52.35886764526367,
|
| 371 |
+
"learning_rate": 1.6023688663282574e-05,
|
| 372 |
+
"loss": 0.8374,
|
| 373 |
+
"step": 1100
|
| 374 |
+
},
|
| 375 |
+
{
|
| 376 |
+
"epoch": 5.71156289707751,
|
| 377 |
+
"grad_norm": 29.787914276123047,
|
| 378 |
+
"learning_rate": 1.5882684715172024e-05,
|
| 379 |
+
"loss": 0.618,
|
| 380 |
+
"step": 1125
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 5.838627700127065,
|
| 384 |
+
"grad_norm": 57.911075592041016,
|
| 385 |
+
"learning_rate": 1.5741680767061477e-05,
|
| 386 |
+
"loss": 0.6089,
|
| 387 |
+
"step": 1150
|
| 388 |
+
},
|
| 389 |
+
{
|
| 390 |
+
"epoch": 5.96569250317662,
|
| 391 |
+
"grad_norm": 19.362987518310547,
|
| 392 |
+
"learning_rate": 1.5600676818950934e-05,
|
| 393 |
+
"loss": 0.7111,
|
| 394 |
+
"step": 1175
|
| 395 |
+
},
|
| 396 |
+
{
|
| 397 |
+
"epoch": 6.0,
|
| 398 |
+
"eval_accuracy": 0.8266068759342302,
|
| 399 |
+
"eval_loss": 0.5884621143341064,
|
| 400 |
+
"eval_macro_f1": 0.8269879429990837,
|
| 401 |
+
"eval_runtime": 0.7328,
|
| 402 |
+
"eval_samples_per_second": 912.931,
|
| 403 |
+
"eval_steps_per_second": 57.314,
|
| 404 |
+
"eval_weighted_f1": 0.8270232273444624,
|
| 405 |
+
"step": 1182
|
| 406 |
+
},
|
| 407 |
+
{
|
| 408 |
+
"epoch": 6.09148665819568,
|
| 409 |
+
"grad_norm": 21.163572311401367,
|
| 410 |
+
"learning_rate": 1.5459672870840384e-05,
|
| 411 |
+
"loss": 0.4341,
|
| 412 |
+
"step": 1200
|
| 413 |
+
},
|
| 414 |
+
{
|
| 415 |
+
"epoch": 6.2185514612452355,
|
| 416 |
+
"grad_norm": 81.59049987792969,
|
| 417 |
+
"learning_rate": 1.5318668922729838e-05,
|
| 418 |
+
"loss": 0.4025,
|
| 419 |
+
"step": 1225
|
| 420 |
+
},
|
| 421 |
+
{
|
| 422 |
+
"epoch": 6.34561626429479,
|
| 423 |
+
"grad_norm": 25.52253532409668,
|
| 424 |
+
"learning_rate": 1.5177664974619291e-05,
|
| 425 |
+
"loss": 0.3291,
|
| 426 |
+
"step": 1250
|
| 427 |
+
},
|
| 428 |
+
{
|
| 429 |
+
"epoch": 6.472681067344346,
|
| 430 |
+
"grad_norm": 3.302086591720581,
|
| 431 |
+
"learning_rate": 1.5036661026508743e-05,
|
| 432 |
+
"loss": 0.3049,
|
| 433 |
+
"step": 1275
|
| 434 |
+
},
|
| 435 |
+
{
|
| 436 |
+
"epoch": 6.599745870393901,
|
| 437 |
+
"grad_norm": 17.74028205871582,
|
| 438 |
+
"learning_rate": 1.4895657078398196e-05,
|
| 439 |
+
"loss": 0.5148,
|
| 440 |
+
"step": 1300
|
| 441 |
+
},
|
| 442 |
+
{
|
| 443 |
+
"epoch": 6.726810673443456,
|
| 444 |
+
"grad_norm": 7.377353191375732,
|
| 445 |
+
"learning_rate": 1.4754653130287648e-05,
|
| 446 |
+
"loss": 0.4865,
|
| 447 |
+
"step": 1325
|
| 448 |
+
},
|
| 449 |
+
{
|
| 450 |
+
"epoch": 6.853875476493012,
|
| 451 |
+
"grad_norm": 17.74429702758789,
|
| 452 |
+
"learning_rate": 1.4613649182177103e-05,
|
| 453 |
+
"loss": 0.5307,
|
| 454 |
+
"step": 1350
|
| 455 |
+
},
|
| 456 |
+
{
|
| 457 |
+
"epoch": 6.980940279542566,
|
| 458 |
+
"grad_norm": 77.65316772460938,
|
| 459 |
+
"learning_rate": 1.4472645234066556e-05,
|
| 460 |
+
"loss": 0.4986,
|
| 461 |
+
"step": 1375
|
| 462 |
+
},
|
| 463 |
+
{
|
| 464 |
+
"epoch": 7.0,
|
| 465 |
+
"eval_accuracy": 0.8310911808669657,
|
| 466 |
+
"eval_loss": 0.7535898685455322,
|
| 467 |
+
"eval_macro_f1": 0.8296475120561762,
|
| 468 |
+
"eval_runtime": 0.7467,
|
| 469 |
+
"eval_samples_per_second": 895.984,
|
| 470 |
+
"eval_steps_per_second": 56.25,
|
| 471 |
+
"eval_weighted_f1": 0.8295744703501335,
|
| 472 |
+
"step": 1379
|
| 473 |
+
},
|
| 474 |
+
{
|
| 475 |
+
"epoch": 7.106734434561626,
|
| 476 |
+
"grad_norm": 47.726463317871094,
|
| 477 |
+
"learning_rate": 1.4331641285956008e-05,
|
| 478 |
+
"loss": 0.2898,
|
| 479 |
+
"step": 1400
|
| 480 |
+
},
|
| 481 |
+
{
|
| 482 |
+
"epoch": 7.233799237611182,
|
| 483 |
+
"grad_norm": 127.74728393554688,
|
| 484 |
+
"learning_rate": 1.4190637337845461e-05,
|
| 485 |
+
"loss": 0.3503,
|
| 486 |
+
"step": 1425
|
| 487 |
+
},
|
| 488 |
+
{
|
| 489 |
+
"epoch": 7.360864040660737,
|
| 490 |
+
"grad_norm": 72.09273529052734,
|
| 491 |
+
"learning_rate": 1.4049633389734913e-05,
|
| 492 |
+
"loss": 0.3364,
|
| 493 |
+
"step": 1450
|
| 494 |
+
},
|
| 495 |
+
{
|
| 496 |
+
"epoch": 7.487928843710292,
|
| 497 |
+
"grad_norm": 65.77955627441406,
|
| 498 |
+
"learning_rate": 1.3908629441624366e-05,
|
| 499 |
+
"loss": 0.3278,
|
| 500 |
+
"step": 1475
|
| 501 |
+
},
|
| 502 |
+
{
|
| 503 |
+
"epoch": 7.614993646759848,
|
| 504 |
+
"grad_norm": 32.90453338623047,
|
| 505 |
+
"learning_rate": 1.3767625493513818e-05,
|
| 506 |
+
"loss": 0.3833,
|
| 507 |
+
"step": 1500
|
| 508 |
+
},
|
| 509 |
+
{
|
| 510 |
+
"epoch": 7.742058449809403,
|
| 511 |
+
"grad_norm": 248.89613342285156,
|
| 512 |
+
"learning_rate": 1.3626621545403271e-05,
|
| 513 |
+
"loss": 0.391,
|
| 514 |
+
"step": 1525
|
| 515 |
+
},
|
| 516 |
+
{
|
| 517 |
+
"epoch": 7.869123252858958,
|
| 518 |
+
"grad_norm": 202.55787658691406,
|
| 519 |
+
"learning_rate": 1.3485617597292726e-05,
|
| 520 |
+
"loss": 0.1825,
|
| 521 |
+
"step": 1550
|
| 522 |
+
},
|
| 523 |
+
{
|
| 524 |
+
"epoch": 7.9961880559085134,
|
| 525 |
+
"grad_norm": 57.24322509765625,
|
| 526 |
+
"learning_rate": 1.3344613649182178e-05,
|
| 527 |
+
"loss": 0.4829,
|
| 528 |
+
"step": 1575
|
| 529 |
+
},
|
| 530 |
+
{
|
| 531 |
+
"epoch": 8.0,
|
| 532 |
+
"eval_accuracy": 0.820627802690583,
|
| 533 |
+
"eval_loss": 0.9836516976356506,
|
| 534 |
+
"eval_macro_f1": 0.8213745088846777,
|
| 535 |
+
"eval_runtime": 0.728,
|
| 536 |
+
"eval_samples_per_second": 918.916,
|
| 537 |
+
"eval_steps_per_second": 57.69,
|
| 538 |
+
"eval_weighted_f1": 0.8213917255816803,
|
| 539 |
+
"step": 1576
|
| 540 |
+
},
|
| 541 |
+
{
|
| 542 |
+
"epoch": 8.121982210927573,
|
| 543 |
+
"grad_norm": 1.8322113752365112,
|
| 544 |
+
"learning_rate": 1.3203609701071631e-05,
|
| 545 |
+
"loss": 0.1437,
|
| 546 |
+
"step": 1600
|
| 547 |
+
},
|
| 548 |
+
{
|
| 549 |
+
"epoch": 8.249047013977128,
|
| 550 |
+
"grad_norm": 178.53717041015625,
|
| 551 |
+
"learning_rate": 1.3062605752961083e-05,
|
| 552 |
+
"loss": 0.2542,
|
| 553 |
+
"step": 1625
|
| 554 |
+
},
|
| 555 |
+
{
|
| 556 |
+
"epoch": 8.376111817026684,
|
| 557 |
+
"grad_norm": 157.9364471435547,
|
| 558 |
+
"learning_rate": 1.2921601804850537e-05,
|
| 559 |
+
"loss": 0.3173,
|
| 560 |
+
"step": 1650
|
| 561 |
+
},
|
| 562 |
+
{
|
| 563 |
+
"epoch": 8.503176620076239,
|
| 564 |
+
"grad_norm": 93.94173431396484,
|
| 565 |
+
"learning_rate": 1.278059785673999e-05,
|
| 566 |
+
"loss": 0.3372,
|
| 567 |
+
"step": 1675
|
| 568 |
+
},
|
| 569 |
+
{
|
| 570 |
+
"epoch": 8.630241423125794,
|
| 571 |
+
"grad_norm": 47.48844528198242,
|
| 572 |
+
"learning_rate": 1.2639593908629442e-05,
|
| 573 |
+
"loss": 0.3196,
|
| 574 |
+
"step": 1700
|
| 575 |
+
},
|
| 576 |
+
{
|
| 577 |
+
"epoch": 8.757306226175348,
|
| 578 |
+
"grad_norm": 0.2058953493833542,
|
| 579 |
+
"learning_rate": 1.2498589960518897e-05,
|
| 580 |
+
"loss": 0.3344,
|
| 581 |
+
"step": 1725
|
| 582 |
+
},
|
| 583 |
+
{
|
| 584 |
+
"epoch": 8.884371029224905,
|
| 585 |
+
"grad_norm": 0.08445502817630768,
|
| 586 |
+
"learning_rate": 1.2357586012408348e-05,
|
| 587 |
+
"loss": 0.3248,
|
| 588 |
+
"step": 1750
|
| 589 |
+
},
|
| 590 |
+
{
|
| 591 |
+
"epoch": 9.0,
|
| 592 |
+
"eval_accuracy": 0.8310911808669657,
|
| 593 |
+
"eval_loss": 0.9798506498336792,
|
| 594 |
+
"eval_macro_f1": 0.8315531652064146,
|
| 595 |
+
"eval_runtime": 0.7324,
|
| 596 |
+
"eval_samples_per_second": 913.443,
|
| 597 |
+
"eval_steps_per_second": 57.346,
|
| 598 |
+
"eval_weighted_f1": 0.8315146655390288,
|
| 599 |
+
"step": 1773
|
| 600 |
+
},
|
| 601 |
+
{
|
| 602 |
+
"epoch": 9.010165184243965,
|
| 603 |
+
"grad_norm": 107.16582489013672,
|
| 604 |
+
"learning_rate": 1.2216582064297802e-05,
|
| 605 |
+
"loss": 0.3178,
|
| 606 |
+
"step": 1775
|
| 607 |
+
},
|
| 608 |
+
{
|
| 609 |
+
"epoch": 9.13722998729352,
|
| 610 |
+
"grad_norm": 1.8582719564437866,
|
| 611 |
+
"learning_rate": 1.2075578116187253e-05,
|
| 612 |
+
"loss": 0.1502,
|
| 613 |
+
"step": 1800
|
| 614 |
+
},
|
| 615 |
+
{
|
| 616 |
+
"epoch": 9.264294790343074,
|
| 617 |
+
"grad_norm": 0.08437829464673996,
|
| 618 |
+
"learning_rate": 1.1934574168076707e-05,
|
| 619 |
+
"loss": 0.1494,
|
| 620 |
+
"step": 1825
|
| 621 |
+
},
|
| 622 |
+
{
|
| 623 |
+
"epoch": 9.39135959339263,
|
| 624 |
+
"grad_norm": 2.747943639755249,
|
| 625 |
+
"learning_rate": 1.179357021996616e-05,
|
| 626 |
+
"loss": 0.1686,
|
| 627 |
+
"step": 1850
|
| 628 |
+
},
|
| 629 |
+
{
|
| 630 |
+
"epoch": 9.518424396442185,
|
| 631 |
+
"grad_norm": 0.07697977870702744,
|
| 632 |
+
"learning_rate": 1.1652566271855612e-05,
|
| 633 |
+
"loss": 0.3106,
|
| 634 |
+
"step": 1875
|
| 635 |
+
},
|
| 636 |
+
{
|
| 637 |
+
"epoch": 9.64548919949174,
|
| 638 |
+
"grad_norm": 139.0595245361328,
|
| 639 |
+
"learning_rate": 1.1511562323745067e-05,
|
| 640 |
+
"loss": 0.3035,
|
| 641 |
+
"step": 1900
|
| 642 |
+
},
|
| 643 |
+
{
|
| 644 |
+
"epoch": 9.772554002541296,
|
| 645 |
+
"grad_norm": 0.6565935015678406,
|
| 646 |
+
"learning_rate": 1.1370558375634519e-05,
|
| 647 |
+
"loss": 0.1772,
|
| 648 |
+
"step": 1925
|
| 649 |
+
},
|
| 650 |
+
{
|
| 651 |
+
"epoch": 9.899618805590851,
|
| 652 |
+
"grad_norm": 0.24950720369815826,
|
| 653 |
+
"learning_rate": 1.1229554427523972e-05,
|
| 654 |
+
"loss": 0.2652,
|
| 655 |
+
"step": 1950
|
| 656 |
+
},
|
| 657 |
+
{
|
| 658 |
+
"epoch": 10.0,
|
| 659 |
+
"eval_accuracy": 0.8340807174887892,
|
| 660 |
+
"eval_loss": 1.067696213722229,
|
| 661 |
+
"eval_macro_f1": 0.8345702219772342,
|
| 662 |
+
"eval_runtime": 0.7361,
|
| 663 |
+
"eval_samples_per_second": 908.826,
|
| 664 |
+
"eval_steps_per_second": 57.056,
|
| 665 |
+
"eval_weighted_f1": 0.8345086622837158,
|
| 666 |
+
"step": 1970
|
| 667 |
+
},
|
| 668 |
+
{
|
| 669 |
+
"epoch": 10.025412960609911,
|
| 670 |
+
"grad_norm": 0.5611938238143921,
|
| 671 |
+
"learning_rate": 1.1088550479413425e-05,
|
| 672 |
+
"loss": 0.1058,
|
| 673 |
+
"step": 1975
|
| 674 |
+
},
|
| 675 |
+
{
|
| 676 |
+
"epoch": 10.152477763659466,
|
| 677 |
+
"grad_norm": 159.5753631591797,
|
| 678 |
+
"learning_rate": 1.0947546531302877e-05,
|
| 679 |
+
"loss": 0.3395,
|
| 680 |
+
"step": 2000
|
| 681 |
+
},
|
| 682 |
+
{
|
| 683 |
+
"epoch": 10.279542566709022,
|
| 684 |
+
"grad_norm": 0.11062340438365936,
|
| 685 |
+
"learning_rate": 1.080654258319233e-05,
|
| 686 |
+
"loss": 0.2162,
|
| 687 |
+
"step": 2025
|
| 688 |
+
},
|
| 689 |
+
{
|
| 690 |
+
"epoch": 10.406607369758577,
|
| 691 |
+
"grad_norm": 125.50062561035156,
|
| 692 |
+
"learning_rate": 1.0665538635081782e-05,
|
| 693 |
+
"loss": 0.2631,
|
| 694 |
+
"step": 2050
|
| 695 |
+
},
|
| 696 |
+
{
|
| 697 |
+
"epoch": 10.533672172808132,
|
| 698 |
+
"grad_norm": 0.30203333497047424,
|
| 699 |
+
"learning_rate": 1.0524534686971236e-05,
|
| 700 |
+
"loss": 0.2117,
|
| 701 |
+
"step": 2075
|
| 702 |
+
},
|
| 703 |
+
{
|
| 704 |
+
"epoch": 10.660736975857688,
|
| 705 |
+
"grad_norm": 0.17817461490631104,
|
| 706 |
+
"learning_rate": 1.038353073886069e-05,
|
| 707 |
+
"loss": 0.1672,
|
| 708 |
+
"step": 2100
|
| 709 |
+
},
|
| 710 |
+
{
|
| 711 |
+
"epoch": 10.787801778907243,
|
| 712 |
+
"grad_norm": 528.10791015625,
|
| 713 |
+
"learning_rate": 1.0242526790750142e-05,
|
| 714 |
+
"loss": 0.2332,
|
| 715 |
+
"step": 2125
|
| 716 |
+
},
|
| 717 |
+
{
|
| 718 |
+
"epoch": 10.914866581956797,
|
| 719 |
+
"grad_norm": 111.58641052246094,
|
| 720 |
+
"learning_rate": 1.0101522842639596e-05,
|
| 721 |
+
"loss": 0.1683,
|
| 722 |
+
"step": 2150
|
| 723 |
+
},
|
| 724 |
+
{
|
| 725 |
+
"epoch": 11.0,
|
| 726 |
+
"eval_accuracy": 0.8415545590433483,
|
| 727 |
+
"eval_loss": 1.0293971300125122,
|
| 728 |
+
"eval_macro_f1": 0.8423839833625967,
|
| 729 |
+
"eval_runtime": 0.7521,
|
| 730 |
+
"eval_samples_per_second": 889.519,
|
| 731 |
+
"eval_steps_per_second": 55.844,
|
| 732 |
+
"eval_weighted_f1": 0.8422944474828612,
|
| 733 |
+
"step": 2167
|
| 734 |
+
},
|
| 735 |
+
{
|
| 736 |
+
"epoch": 11.040660736975857,
|
| 737 |
+
"grad_norm": 0.059737857431173325,
|
| 738 |
+
"learning_rate": 9.960518894529047e-06,
|
| 739 |
+
"loss": 0.2153,
|
| 740 |
+
"step": 2175
|
| 741 |
+
},
|
| 742 |
+
{
|
| 743 |
+
"epoch": 11.167725540025414,
|
| 744 |
+
"grad_norm": 4.420228481292725,
|
| 745 |
+
"learning_rate": 9.8195149464185e-06,
|
| 746 |
+
"loss": 0.1672,
|
| 747 |
+
"step": 2200
|
| 748 |
+
},
|
| 749 |
+
{
|
| 750 |
+
"epoch": 11.294790343074968,
|
| 751 |
+
"grad_norm": 1.0231783390045166,
|
| 752 |
+
"learning_rate": 9.678510998307954e-06,
|
| 753 |
+
"loss": 0.2809,
|
| 754 |
+
"step": 2225
|
| 755 |
+
},
|
| 756 |
+
{
|
| 757 |
+
"epoch": 11.421855146124523,
|
| 758 |
+
"grad_norm": 131.25509643554688,
|
| 759 |
+
"learning_rate": 9.537507050197406e-06,
|
| 760 |
+
"loss": 0.1128,
|
| 761 |
+
"step": 2250
|
| 762 |
+
},
|
| 763 |
+
{
|
| 764 |
+
"epoch": 11.54891994917408,
|
| 765 |
+
"grad_norm": 0.05244371294975281,
|
| 766 |
+
"learning_rate": 9.39650310208686e-06,
|
| 767 |
+
"loss": 0.1769,
|
| 768 |
+
"step": 2275
|
| 769 |
+
},
|
| 770 |
+
{
|
| 771 |
+
"epoch": 11.675984752223634,
|
| 772 |
+
"grad_norm": 0.0427590049803257,
|
| 773 |
+
"learning_rate": 9.255499153976311e-06,
|
| 774 |
+
"loss": 0.038,
|
| 775 |
+
"step": 2300
|
| 776 |
+
},
|
| 777 |
+
{
|
| 778 |
+
"epoch": 11.803049555273189,
|
| 779 |
+
"grad_norm": 166.7871551513672,
|
| 780 |
+
"learning_rate": 9.114495205865766e-06,
|
| 781 |
+
"loss": 0.2303,
|
| 782 |
+
"step": 2325
|
| 783 |
+
},
|
| 784 |
+
{
|
| 785 |
+
"epoch": 11.930114358322745,
|
| 786 |
+
"grad_norm": 1.097733497619629,
|
| 787 |
+
"learning_rate": 8.973491257755218e-06,
|
| 788 |
+
"loss": 0.1571,
|
| 789 |
+
"step": 2350
|
| 790 |
+
},
|
| 791 |
+
{
|
| 792 |
+
"epoch": 12.0,
|
| 793 |
+
"eval_accuracy": 0.8385650224215246,
|
| 794 |
+
"eval_loss": 1.061612606048584,
|
| 795 |
+
"eval_macro_f1": 0.838626505293172,
|
| 796 |
+
"eval_runtime": 0.7443,
|
| 797 |
+
"eval_samples_per_second": 898.843,
|
| 798 |
+
"eval_steps_per_second": 56.43,
|
| 799 |
+
"eval_weighted_f1": 0.8385499237517174,
|
| 800 |
+
"step": 2364
|
| 801 |
+
},
|
| 802 |
+
{
|
| 803 |
+
"epoch": 12.055908513341805,
|
| 804 |
+
"grad_norm": 0.0526498481631279,
|
| 805 |
+
"learning_rate": 8.832487309644671e-06,
|
| 806 |
+
"loss": 0.1029,
|
| 807 |
+
"step": 2375
|
| 808 |
+
},
|
| 809 |
+
{
|
| 810 |
+
"epoch": 12.18297331639136,
|
| 811 |
+
"grad_norm": 0.04083005711436272,
|
| 812 |
+
"learning_rate": 8.691483361534123e-06,
|
| 813 |
+
"loss": 0.1087,
|
| 814 |
+
"step": 2400
|
| 815 |
+
},
|
| 816 |
+
{
|
| 817 |
+
"epoch": 12.310038119440915,
|
| 818 |
+
"grad_norm": 375.5091247558594,
|
| 819 |
+
"learning_rate": 8.550479413423576e-06,
|
| 820 |
+
"loss": 0.2106,
|
| 821 |
+
"step": 2425
|
| 822 |
+
},
|
| 823 |
+
{
|
| 824 |
+
"epoch": 12.437102922490471,
|
| 825 |
+
"grad_norm": 36.84629821777344,
|
| 826 |
+
"learning_rate": 8.40947546531303e-06,
|
| 827 |
+
"loss": 0.1549,
|
| 828 |
+
"step": 2450
|
| 829 |
+
},
|
| 830 |
+
{
|
| 831 |
+
"epoch": 12.564167725540026,
|
| 832 |
+
"grad_norm": 0.5839002132415771,
|
| 833 |
+
"learning_rate": 8.268471517202483e-06,
|
| 834 |
+
"loss": 0.0738,
|
| 835 |
+
"step": 2475
|
| 836 |
+
},
|
| 837 |
+
{
|
| 838 |
+
"epoch": 12.69123252858958,
|
| 839 |
+
"grad_norm": 0.035607073456048965,
|
| 840 |
+
"learning_rate": 8.127467569091936e-06,
|
| 841 |
+
"loss": 0.0822,
|
| 842 |
+
"step": 2500
|
| 843 |
+
},
|
| 844 |
+
{
|
| 845 |
+
"epoch": 12.818297331639137,
|
| 846 |
+
"grad_norm": 0.04663573578000069,
|
| 847 |
+
"learning_rate": 7.986463620981388e-06,
|
| 848 |
+
"loss": 0.0845,
|
| 849 |
+
"step": 2525
|
| 850 |
+
},
|
| 851 |
+
{
|
| 852 |
+
"epoch": 12.945362134688692,
|
| 853 |
+
"grad_norm": 0.032848622649908066,
|
| 854 |
+
"learning_rate": 7.845459672870841e-06,
|
| 855 |
+
"loss": 0.1394,
|
| 856 |
+
"step": 2550
|
| 857 |
+
},
|
| 858 |
+
{
|
| 859 |
+
"epoch": 13.0,
|
| 860 |
+
"eval_accuracy": 0.8370702541106129,
|
| 861 |
+
"eval_loss": 1.119581699371338,
|
| 862 |
+
"eval_macro_f1": 0.8380561154301436,
|
| 863 |
+
"eval_runtime": 0.7211,
|
| 864 |
+
"eval_samples_per_second": 927.687,
|
| 865 |
+
"eval_steps_per_second": 58.24,
|
| 866 |
+
"eval_weighted_f1": 0.8379579390360057,
|
| 867 |
+
"step": 2561
|
| 868 |
+
},
|
| 869 |
+
{
|
| 870 |
+
"epoch": 13.071156289707751,
|
| 871 |
+
"grad_norm": 0.035511430352926254,
|
| 872 |
+
"learning_rate": 7.704455724760293e-06,
|
| 873 |
+
"loss": 0.1007,
|
| 874 |
+
"step": 2575
|
| 875 |
+
},
|
| 876 |
+
{
|
| 877 |
+
"epoch": 13.198221092757306,
|
| 878 |
+
"grad_norm": 0.08027854561805725,
|
| 879 |
+
"learning_rate": 7.563451776649747e-06,
|
| 880 |
+
"loss": 0.1057,
|
| 881 |
+
"step": 2600
|
| 882 |
+
},
|
| 883 |
+
{
|
| 884 |
+
"epoch": 13.32528589580686,
|
| 885 |
+
"grad_norm": 0.14604303240776062,
|
| 886 |
+
"learning_rate": 7.4224478285392e-06,
|
| 887 |
+
"loss": 0.0984,
|
| 888 |
+
"step": 2625
|
| 889 |
+
},
|
| 890 |
+
{
|
| 891 |
+
"epoch": 13.452350698856417,
|
| 892 |
+
"grad_norm": 0.03396645933389664,
|
| 893 |
+
"learning_rate": 7.281443880428652e-06,
|
| 894 |
+
"loss": 0.1031,
|
| 895 |
+
"step": 2650
|
| 896 |
+
},
|
| 897 |
+
{
|
| 898 |
+
"epoch": 13.579415501905972,
|
| 899 |
+
"grad_norm": 0.05524778366088867,
|
| 900 |
+
"learning_rate": 7.140439932318105e-06,
|
| 901 |
+
"loss": 0.056,
|
| 902 |
+
"step": 2675
|
| 903 |
+
},
|
| 904 |
+
{
|
| 905 |
+
"epoch": 13.706480304955527,
|
| 906 |
+
"grad_norm": 1.6461844444274902,
|
| 907 |
+
"learning_rate": 6.999435984207559e-06,
|
| 908 |
+
"loss": 0.1624,
|
| 909 |
+
"step": 2700
|
| 910 |
+
},
|
| 911 |
+
{
|
| 912 |
+
"epoch": 13.833545108005083,
|
| 913 |
+
"grad_norm": 0.1359899789094925,
|
| 914 |
+
"learning_rate": 6.8584320360970115e-06,
|
| 915 |
+
"loss": 0.0896,
|
| 916 |
+
"step": 2725
|
| 917 |
+
},
|
| 918 |
+
{
|
| 919 |
+
"epoch": 13.960609911054638,
|
| 920 |
+
"grad_norm": 0.3963850736618042,
|
| 921 |
+
"learning_rate": 6.717428087986464e-06,
|
| 922 |
+
"loss": 0.0855,
|
| 923 |
+
"step": 2750
|
| 924 |
+
},
|
| 925 |
+
{
|
| 926 |
+
"epoch": 14.0,
|
| 927 |
+
"eval_accuracy": 0.8415545590433483,
|
| 928 |
+
"eval_loss": 1.1025843620300293,
|
| 929 |
+
"eval_macro_f1": 0.8418671931234222,
|
| 930 |
+
"eval_runtime": 0.7688,
|
| 931 |
+
"eval_samples_per_second": 870.164,
|
| 932 |
+
"eval_steps_per_second": 54.629,
|
| 933 |
+
"eval_weighted_f1": 0.8417486116332702,
|
| 934 |
+
"step": 2758
|
| 935 |
+
}
|
| 936 |
+
],
|
| 937 |
+
"logging_steps": 25,
|
| 938 |
+
"max_steps": 3940,
|
| 939 |
+
"num_input_tokens_seen": 0,
|
| 940 |
+
"num_train_epochs": 20,
|
| 941 |
+
"save_steps": 500,
|
| 942 |
+
"stateful_callbacks": {
|
| 943 |
+
"EarlyStoppingCallback": {
|
| 944 |
+
"args": {
|
| 945 |
+
"early_stopping_patience": 3,
|
| 946 |
+
"early_stopping_threshold": 0.0
|
| 947 |
+
},
|
| 948 |
+
"attributes": {
|
| 949 |
+
"early_stopping_patience_counter": 3
|
| 950 |
+
}
|
| 951 |
+
},
|
| 952 |
+
"TrainerControl": {
|
| 953 |
+
"args": {
|
| 954 |
+
"should_epoch_stop": false,
|
| 955 |
+
"should_evaluate": false,
|
| 956 |
+
"should_log": false,
|
| 957 |
+
"should_save": true,
|
| 958 |
+
"should_training_stop": true
|
| 959 |
+
},
|
| 960 |
+
"attributes": {}
|
| 961 |
+
}
|
| 962 |
+
},
|
| 963 |
+
"total_flos": 0.0,
|
| 964 |
+
"train_batch_size": 8,
|
| 965 |
+
"trial_name": null,
|
| 966 |
+
"trial_params": null
|
| 967 |
+
}
|