Upload final LLM2Seq training artifacts
Browse files
checkpoints/phase1_warmup_qwen_wikilingua/train.log
ADDED
|
@@ -0,0 +1,907 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
2026-06-24 15:25:26,228 [INFO] Seed: 42
|
| 2 |
+
2026-06-24 15:25:26,229 [INFO] Using CUDA: NVIDIA L40
|
| 3 |
+
2026-06-24 15:25:26,230 [INFO] Enabled TF32 matmul/cudnn for CUDA
|
| 4 |
+
2026-06-24 15:27:29,974 [INFO] Seed: 42
|
| 5 |
+
2026-06-24 15:27:29,975 [INFO] Using CUDA: NVIDIA L40
|
| 6 |
+
2026-06-24 15:27:29,975 [INFO] Enabled TF32 matmul/cudnn for CUDA
|
| 7 |
+
2026-06-24 15:27:30,132 [INFO] HTTP Request: GET https://huggingface.co/api/agent-harnesses "HTTP/1.1 200 OK"
|
| 8 |
+
2026-06-24 15:27:30,264 [INFO] HTTP Request: HEAD https://huggingface.co/Qwen/Qwen3-Embedding-0.6B/resolve/main/config.json "HTTP/1.1 307 Temporary Redirect"
|
| 9 |
+
2026-06-24 15:27:30,295 [INFO] HTTP Request: HEAD https://huggingface.co/api/resolve-cache/models/Qwen/Qwen3-Embedding-0.6B/97b0c614be4d77ee51c0cef4e5f07c00f9eb65b3/config.json "HTTP/1.1 200 OK"
|
| 10 |
+
2026-06-24 15:27:30,440 [INFO] HTTP Request: HEAD https://huggingface.co/Qwen/Qwen3-Embedding-0.6B/resolve/main/tokenizer_config.json "HTTP/1.1 307 Temporary Redirect"
|
| 11 |
+
2026-06-24 15:27:30,470 [INFO] HTTP Request: HEAD https://huggingface.co/api/resolve-cache/models/Qwen/Qwen3-Embedding-0.6B/97b0c614be4d77ee51c0cef4e5f07c00f9eb65b3/tokenizer_config.json "HTTP/1.1 200 OK"
|
| 12 |
+
2026-06-24 15:27:30,609 [INFO] HTTP Request: GET https://huggingface.co/api/models/Qwen/Qwen3-Embedding-0.6B/tree/main/additional_chat_templates?recursive=false&expand=false "HTTP/1.1 404 Not Found"
|
| 13 |
+
2026-06-24 15:27:30,744 [INFO] HTTP Request: GET https://huggingface.co/api/models/Qwen/Qwen3-Embedding-0.6B/tree/main?recursive=true&expand=false "HTTP/1.1 200 OK"
|
| 14 |
+
2026-06-24 15:27:31,398 [INFO] HTTP Request: GET https://huggingface.co/api/models/Qwen/Qwen3-Embedding-0.6B "HTTP/1.1 200 OK"
|
| 15 |
+
2026-06-24 15:27:31,565 [INFO] HTTP Request: HEAD https://huggingface.co/Qwen/Qwen3-Embedding-0.6B/resolve/main/config.json "HTTP/1.1 307 Temporary Redirect"
|
| 16 |
+
2026-06-24 15:27:31,571 [INFO] HTTP Request: HEAD https://huggingface.co/api/resolve-cache/models/Qwen/Qwen3-Embedding-0.6B/97b0c614be4d77ee51c0cef4e5f07c00f9eb65b3/config.json "HTTP/1.1 200 OK"
|
| 17 |
+
2026-06-24 15:27:31,705 [INFO] HTTP Request: HEAD https://huggingface.co/Qwen/Qwen3-Embedding-0.6B/resolve/main/adapter_config.json "HTTP/1.1 404 Not Found"
|
| 18 |
+
2026-06-24 15:27:31,943 [INFO] HTTP Request: HEAD https://huggingface.co/Qwen/Qwen3-Embedding-0.6B/resolve/main/config.json "HTTP/1.1 307 Temporary Redirect"
|
| 19 |
+
2026-06-24 15:27:31,949 [INFO] HTTP Request: HEAD https://huggingface.co/api/resolve-cache/models/Qwen/Qwen3-Embedding-0.6B/97b0c614be4d77ee51c0cef4e5f07c00f9eb65b3/config.json "HTTP/1.1 200 OK"
|
| 20 |
+
2026-06-24 15:27:32,091 [INFO] HTTP Request: HEAD https://huggingface.co/Qwen/Qwen3-Embedding-0.6B/resolve/main/model.safetensors "HTTP/1.1 302 Found"
|
| 21 |
+
2026-06-24 15:27:38,317 [INFO] LLM2Seq Model Summary
|
| 22 |
+
==================================================
|
| 23 |
+
Encoder: Qwen/Qwen3-Embedding-0.6B
|
| 24 |
+
Trainable: False
|
| 25 |
+
Adaptor: gated_residual
|
| 26 |
+
Layer Fusion: True
|
| 27 |
+
Fusion Type: tokenwise
|
| 28 |
+
Salience Gate: True
|
| 29 |
+
EncStack: True
|
| 30 |
+
Global Tokens: 16
|
| 31 |
+
Decoder:
|
| 32 |
+
Layers: 8
|
| 33 |
+
Hidden Size: 1024
|
| 34 |
+
Heads: 16
|
| 35 |
+
FFN Size: 4096
|
| 36 |
+
==================================================
|
| 37 |
+
Total params: 980,047,361
|
| 38 |
+
Trainable params: 384,270,849
|
| 39 |
+
Frozen params: 595,776,512
|
| 40 |
+
2026-06-24 15:29:03,803 [INFO] Seed: 42
|
| 41 |
+
2026-06-24 15:29:03,804 [INFO] Using CUDA: NVIDIA L40
|
| 42 |
+
2026-06-24 15:29:03,805 [INFO] Enabled TF32 matmul/cudnn for CUDA
|
| 43 |
+
2026-06-24 15:29:03,988 [INFO] HTTP Request: HEAD https://huggingface.co/Qwen/Qwen3-Embedding-0.6B/resolve/main/config.json "HTTP/1.1 307 Temporary Redirect"
|
| 44 |
+
2026-06-24 15:29:03,994 [INFO] HTTP Request: HEAD https://huggingface.co/api/resolve-cache/models/Qwen/Qwen3-Embedding-0.6B/97b0c614be4d77ee51c0cef4e5f07c00f9eb65b3/config.json "HTTP/1.1 200 OK"
|
| 45 |
+
2026-06-24 15:29:04,133 [INFO] HTTP Request: HEAD https://huggingface.co/Qwen/Qwen3-Embedding-0.6B/resolve/main/tokenizer_config.json "HTTP/1.1 307 Temporary Redirect"
|
| 46 |
+
2026-06-24 15:29:04,140 [INFO] HTTP Request: HEAD https://huggingface.co/api/resolve-cache/models/Qwen/Qwen3-Embedding-0.6B/97b0c614be4d77ee51c0cef4e5f07c00f9eb65b3/tokenizer_config.json "HTTP/1.1 200 OK"
|
| 47 |
+
2026-06-24 15:29:04,282 [INFO] HTTP Request: GET https://huggingface.co/api/models/Qwen/Qwen3-Embedding-0.6B/tree/main/additional_chat_templates?recursive=false&expand=false "HTTP/1.1 404 Not Found"
|
| 48 |
+
2026-06-24 15:29:04,422 [INFO] HTTP Request: GET https://huggingface.co/api/models/Qwen/Qwen3-Embedding-0.6B/tree/main?recursive=true&expand=false "HTTP/1.1 200 OK"
|
| 49 |
+
2026-06-24 15:29:05,098 [INFO] HTTP Request: GET https://huggingface.co/api/models/Qwen/Qwen3-Embedding-0.6B "HTTP/1.1 200 OK"
|
| 50 |
+
2026-06-24 15:29:05,260 [INFO] HTTP Request: HEAD https://huggingface.co/Qwen/Qwen3-Embedding-0.6B/resolve/main/config.json "HTTP/1.1 307 Temporary Redirect"
|
| 51 |
+
2026-06-24 15:29:05,267 [INFO] HTTP Request: HEAD https://huggingface.co/api/resolve-cache/models/Qwen/Qwen3-Embedding-0.6B/97b0c614be4d77ee51c0cef4e5f07c00f9eb65b3/config.json "HTTP/1.1 200 OK"
|
| 52 |
+
2026-06-24 15:29:05,506 [INFO] HTTP Request: HEAD https://huggingface.co/Qwen/Qwen3-Embedding-0.6B/resolve/main/config.json "HTTP/1.1 307 Temporary Redirect"
|
| 53 |
+
2026-06-24 15:29:05,513 [INFO] HTTP Request: HEAD https://huggingface.co/api/resolve-cache/models/Qwen/Qwen3-Embedding-0.6B/97b0c614be4d77ee51c0cef4e5f07c00f9eb65b3/config.json "HTTP/1.1 200 OK"
|
| 54 |
+
2026-06-24 15:29:09,790 [INFO] LLM2Seq Model Summary
|
| 55 |
+
==================================================
|
| 56 |
+
Encoder: Qwen/Qwen3-Embedding-0.6B
|
| 57 |
+
Trainable: False
|
| 58 |
+
Adaptor: gated_residual
|
| 59 |
+
Layer Fusion: True
|
| 60 |
+
Fusion Type: tokenwise
|
| 61 |
+
Salience Gate: True
|
| 62 |
+
EncStack: True
|
| 63 |
+
Global Tokens: 16
|
| 64 |
+
Decoder:
|
| 65 |
+
Layers: 8
|
| 66 |
+
Hidden Size: 1024
|
| 67 |
+
Heads: 16
|
| 68 |
+
FFN Size: 4096
|
| 69 |
+
==================================================
|
| 70 |
+
Total params: 980,047,361
|
| 71 |
+
Trainable params: 384,270,849
|
| 72 |
+
Frozen params: 595,776,512
|
| 73 |
+
2026-06-24 15:29:11,735 [INFO] HTTP Request: POST https://huggingface.co/api/repos/create "HTTP/1.1 409 Conflict"
|
| 74 |
+
2026-06-24 15:29:11,736 [INFO] Hugging Face uploads enabled: repo=BienKieu/llm2seq-wikilingua path=checkpoints/phase1_warmup_qwen_wikilingua
|
| 75 |
+
2026-06-24 15:29:11,737 [INFO] Train file: llm2seq/data/processed/train.jsonl (13,999 examples)
|
| 76 |
+
2026-06-24 15:29:11,737 [INFO] Eval file: llm2seq/data/processed/validation.jsonl (1,680 examples)
|
| 77 |
+
2026-06-24 15:29:11,737 [INFO] Starting training for 8 epochs (2,336 optimizer steps)...
|
| 78 |
+
2026-06-24 15:29:11,737 [INFO] Batch size: 24
|
| 79 |
+
2026-06-24 15:29:11,737 [INFO] Gradient accumulation: 2
|
| 80 |
+
2026-06-24 15:29:11,737 [INFO] Effective batch size: 48
|
| 81 |
+
2026-06-24 15:29:11,737 [INFO] Steps per epoch: 292
|
| 82 |
+
2026-06-24 15:29:11,737 [INFO] Warmup steps: 70
|
| 83 |
+
2026-06-24 15:29:11,737 [INFO] Eval every: 292 optimizer steps
|
| 84 |
+
2026-06-24 15:29:11,737 [INFO] Save every: 292 optimizer steps
|
| 85 |
+
2026-06-24 15:29:11,737 [INFO] Mixed precision: fp16=False, bf16=True
|
| 86 |
+
2026-06-24 15:29:11,737 [INFO] Initial optimizer LRs: adaptor=0.00e+00, decoder=0.00e+00
|
| 87 |
+
2026-06-24 15:30:28,194 [INFO] Seed: 42
|
| 88 |
+
2026-06-24 15:30:28,195 [INFO] Using CUDA: NVIDIA L40
|
| 89 |
+
2026-06-24 15:30:28,196 [INFO] Enabled TF32 matmul/cudnn for CUDA
|
| 90 |
+
2026-06-24 15:30:28,361 [INFO] HTTP Request: HEAD https://huggingface.co/Qwen/Qwen3-Embedding-0.6B/resolve/main/config.json "HTTP/1.1 307 Temporary Redirect"
|
| 91 |
+
2026-06-24 15:30:28,367 [INFO] HTTP Request: HEAD https://huggingface.co/api/resolve-cache/models/Qwen/Qwen3-Embedding-0.6B/97b0c614be4d77ee51c0cef4e5f07c00f9eb65b3/config.json "HTTP/1.1 200 OK"
|
| 92 |
+
2026-06-24 15:30:28,506 [INFO] HTTP Request: HEAD https://huggingface.co/Qwen/Qwen3-Embedding-0.6B/resolve/main/tokenizer_config.json "HTTP/1.1 307 Temporary Redirect"
|
| 93 |
+
2026-06-24 15:30:28,511 [INFO] HTTP Request: HEAD https://huggingface.co/api/resolve-cache/models/Qwen/Qwen3-Embedding-0.6B/97b0c614be4d77ee51c0cef4e5f07c00f9eb65b3/tokenizer_config.json "HTTP/1.1 200 OK"
|
| 94 |
+
2026-06-24 15:30:28,656 [INFO] HTTP Request: GET https://huggingface.co/api/models/Qwen/Qwen3-Embedding-0.6B/tree/main/additional_chat_templates?recursive=false&expand=false "HTTP/1.1 404 Not Found"
|
| 95 |
+
2026-06-24 15:30:28,822 [INFO] HTTP Request: GET https://huggingface.co/api/models/Qwen/Qwen3-Embedding-0.6B/tree/main?recursive=true&expand=false "HTTP/1.1 200 OK"
|
| 96 |
+
2026-06-24 15:30:29,506 [INFO] HTTP Request: GET https://huggingface.co/api/models/Qwen/Qwen3-Embedding-0.6B "HTTP/1.1 200 OK"
|
| 97 |
+
2026-06-24 15:30:29,667 [INFO] HTTP Request: HEAD https://huggingface.co/Qwen/Qwen3-Embedding-0.6B/resolve/main/config.json "HTTP/1.1 307 Temporary Redirect"
|
| 98 |
+
2026-06-24 15:30:29,672 [INFO] HTTP Request: HEAD https://huggingface.co/api/resolve-cache/models/Qwen/Qwen3-Embedding-0.6B/97b0c614be4d77ee51c0cef4e5f07c00f9eb65b3/config.json "HTTP/1.1 200 OK"
|
| 99 |
+
2026-06-24 15:30:29,915 [INFO] HTTP Request: HEAD https://huggingface.co/Qwen/Qwen3-Embedding-0.6B/resolve/main/config.json "HTTP/1.1 307 Temporary Redirect"
|
| 100 |
+
2026-06-24 15:30:29,921 [INFO] HTTP Request: HEAD https://huggingface.co/api/resolve-cache/models/Qwen/Qwen3-Embedding-0.6B/97b0c614be4d77ee51c0cef4e5f07c00f9eb65b3/config.json "HTTP/1.1 200 OK"
|
| 101 |
+
2026-06-24 15:30:34,396 [INFO] LLM2Seq Model Summary
|
| 102 |
+
==================================================
|
| 103 |
+
Encoder: Qwen/Qwen3-Embedding-0.6B
|
| 104 |
+
Trainable: False
|
| 105 |
+
Adaptor: gated_residual
|
| 106 |
+
Layer Fusion: True
|
| 107 |
+
Fusion Type: tokenwise
|
| 108 |
+
Salience Gate: True
|
| 109 |
+
EncStack: True
|
| 110 |
+
Global Tokens: 16
|
| 111 |
+
Decoder:
|
| 112 |
+
Layers: 8
|
| 113 |
+
Hidden Size: 1024
|
| 114 |
+
Heads: 16
|
| 115 |
+
FFN Size: 4096
|
| 116 |
+
==================================================
|
| 117 |
+
Total params: 980,047,361
|
| 118 |
+
Trainable params: 384,270,849
|
| 119 |
+
Frozen params: 595,776,512
|
| 120 |
+
2026-06-24 15:30:36,669 [INFO] HTTP Request: POST https://huggingface.co/api/repos/create "HTTP/1.1 409 Conflict"
|
| 121 |
+
2026-06-24 15:30:36,670 [INFO] Hugging Face uploads enabled: repo=BienKieu/llm2seq-wikilingua path=checkpoints/phase1_warmup_qwen_wikilingua
|
| 122 |
+
2026-06-24 15:30:36,670 [INFO] Train file: llm2seq/data/processed/train.jsonl (13,999 examples)
|
| 123 |
+
2026-06-24 15:30:36,671 [INFO] Eval file: llm2seq/data/processed/validation.jsonl (1,680 examples)
|
| 124 |
+
2026-06-24 15:30:36,671 [INFO] Starting training for 8 epochs (3,496 optimizer steps)...
|
| 125 |
+
2026-06-24 15:30:36,671 [INFO] Batch size: 16
|
| 126 |
+
2026-06-24 15:30:36,671 [INFO] Gradient accumulation: 2
|
| 127 |
+
2026-06-24 15:30:36,671 [INFO] Effective batch size: 32
|
| 128 |
+
2026-06-24 15:30:36,671 [INFO] Steps per epoch: 437
|
| 129 |
+
2026-06-24 15:30:36,671 [INFO] Warmup steps: 104
|
| 130 |
+
2026-06-24 15:30:36,671 [INFO] Eval every: 437 optimizer steps
|
| 131 |
+
2026-06-24 15:30:36,671 [INFO] Save every: 437 optimizer steps
|
| 132 |
+
2026-06-24 15:30:36,671 [INFO] Mixed precision: fp16=False, bf16=True
|
| 133 |
+
2026-06-24 15:30:36,671 [INFO] Initial optimizer LRs: adaptor=0.00e+00, decoder=0.00e+00
|
| 134 |
+
2026-06-24 15:30:50,587 [INFO] Step 5/3496 | Loss: 11.9001 | CE: 11.9001 | LR: 2.40e-05 | Epoch: 1/8
|
| 135 |
+
2026-06-24 15:31:28,173 [INFO] Seed: 42
|
| 136 |
+
2026-06-24 15:31:28,174 [INFO] Using CUDA: NVIDIA L40
|
| 137 |
+
2026-06-24 15:31:28,174 [INFO] Enabled TF32 matmul/cudnn for CUDA
|
| 138 |
+
2026-06-24 15:31:28,340 [INFO] HTTP Request: HEAD https://huggingface.co/Qwen/Qwen3-Embedding-0.6B/resolve/main/config.json "HTTP/1.1 307 Temporary Redirect"
|
| 139 |
+
2026-06-24 15:31:28,346 [INFO] HTTP Request: HEAD https://huggingface.co/api/resolve-cache/models/Qwen/Qwen3-Embedding-0.6B/97b0c614be4d77ee51c0cef4e5f07c00f9eb65b3/config.json "HTTP/1.1 200 OK"
|
| 140 |
+
2026-06-24 15:31:28,487 [INFO] HTTP Request: HEAD https://huggingface.co/Qwen/Qwen3-Embedding-0.6B/resolve/main/tokenizer_config.json "HTTP/1.1 307 Temporary Redirect"
|
| 141 |
+
2026-06-24 15:31:28,492 [INFO] HTTP Request: HEAD https://huggingface.co/api/resolve-cache/models/Qwen/Qwen3-Embedding-0.6B/97b0c614be4d77ee51c0cef4e5f07c00f9eb65b3/tokenizer_config.json "HTTP/1.1 200 OK"
|
| 142 |
+
2026-06-24 15:31:28,630 [INFO] HTTP Request: GET https://huggingface.co/api/models/Qwen/Qwen3-Embedding-0.6B/tree/main/additional_chat_templates?recursive=false&expand=false "HTTP/1.1 404 Not Found"
|
| 143 |
+
2026-06-24 15:31:28,820 [INFO] HTTP Request: GET https://huggingface.co/api/models/Qwen/Qwen3-Embedding-0.6B/tree/main?recursive=true&expand=false "HTTP/1.1 200 OK"
|
| 144 |
+
2026-06-24 15:31:29,450 [INFO] HTTP Request: GET https://huggingface.co/api/models/Qwen/Qwen3-Embedding-0.6B "HTTP/1.1 200 OK"
|
| 145 |
+
2026-06-24 15:31:29,609 [INFO] HTTP Request: HEAD https://huggingface.co/Qwen/Qwen3-Embedding-0.6B/resolve/main/config.json "HTTP/1.1 307 Temporary Redirect"
|
| 146 |
+
2026-06-24 15:31:29,615 [INFO] HTTP Request: HEAD https://huggingface.co/api/resolve-cache/models/Qwen/Qwen3-Embedding-0.6B/97b0c614be4d77ee51c0cef4e5f07c00f9eb65b3/config.json "HTTP/1.1 200 OK"
|
| 147 |
+
2026-06-24 15:31:29,855 [INFO] HTTP Request: HEAD https://huggingface.co/Qwen/Qwen3-Embedding-0.6B/resolve/main/config.json "HTTP/1.1 307 Temporary Redirect"
|
| 148 |
+
2026-06-24 15:31:29,861 [INFO] HTTP Request: HEAD https://huggingface.co/api/resolve-cache/models/Qwen/Qwen3-Embedding-0.6B/97b0c614be4d77ee51c0cef4e5f07c00f9eb65b3/config.json "HTTP/1.1 200 OK"
|
| 149 |
+
2026-06-24 15:31:33,917 [INFO] LLM2Seq Model Summary
|
| 150 |
+
==================================================
|
| 151 |
+
Encoder: Qwen/Qwen3-Embedding-0.6B
|
| 152 |
+
Trainable: False
|
| 153 |
+
Adaptor: gated_residual
|
| 154 |
+
Layer Fusion: True
|
| 155 |
+
Fusion Type: tokenwise
|
| 156 |
+
Salience Gate: True
|
| 157 |
+
EncStack: True
|
| 158 |
+
Global Tokens: 16
|
| 159 |
+
Decoder:
|
| 160 |
+
Layers: 8
|
| 161 |
+
Hidden Size: 1024
|
| 162 |
+
Heads: 16
|
| 163 |
+
FFN Size: 4096
|
| 164 |
+
==================================================
|
| 165 |
+
Total params: 980,047,361
|
| 166 |
+
Trainable params: 384,270,849
|
| 167 |
+
Frozen params: 595,776,512
|
| 168 |
+
2026-06-24 15:31:35,848 [INFO] HTTP Request: POST https://huggingface.co/api/repos/create "HTTP/1.1 409 Conflict"
|
| 169 |
+
2026-06-24 15:31:35,849 [INFO] Hugging Face uploads enabled: repo=BienKieu/llm2seq-wikilingua path=checkpoints/phase1_warmup_qwen_wikilingua
|
| 170 |
+
2026-06-24 15:31:35,849 [INFO] Train file: llm2seq/data/processed/train.jsonl (13,999 examples)
|
| 171 |
+
2026-06-24 15:31:35,849 [INFO] Eval file: llm2seq/data/processed/validation.jsonl (1,680 examples)
|
| 172 |
+
2026-06-24 15:31:35,849 [INFO] Starting training for 8 epochs (3,504 optimizer steps)...
|
| 173 |
+
2026-06-24 15:31:35,849 [INFO] Batch size: 8
|
| 174 |
+
2026-06-24 15:31:35,850 [INFO] Gradient accumulation: 4
|
| 175 |
+
2026-06-24 15:31:35,850 [INFO] Effective batch size: 32
|
| 176 |
+
2026-06-24 15:31:35,850 [INFO] Steps per epoch: 438
|
| 177 |
+
2026-06-24 15:31:35,850 [INFO] Warmup steps: 105
|
| 178 |
+
2026-06-24 15:31:35,850 [INFO] Eval every: 438 optimizer steps
|
| 179 |
+
2026-06-24 15:31:35,850 [INFO] Save every: 438 optimizer steps
|
| 180 |
+
2026-06-24 15:31:35,850 [INFO] Mixed precision: fp16=False, bf16=True
|
| 181 |
+
2026-06-24 15:31:35,850 [INFO] Initial optimizer LRs: adaptor=0.00e+00, decoder=0.00e+00
|
| 182 |
+
2026-06-24 15:31:47,199 [INFO] Step 5/3504 | Loss: 11.8994 | CE: 11.8994 | LR: 2.38e-05 | Epoch: 1/8
|
| 183 |
+
2026-06-24 15:31:55,957 [INFO] Step 10/3504 | Loss: 11.0158 | CE: 11.0158 | LR: 4.76e-05 | Epoch: 1/8
|
| 184 |
+
2026-06-24 15:32:05,306 [INFO] Step 15/3504 | Loss: 10.5336 | CE: 10.5336 | LR: 7.14e-05 | Epoch: 1/8
|
| 185 |
+
2026-06-24 15:32:13,234 [INFO] Step 20/3504 | Loss: 10.0759 | CE: 10.0759 | LR: 9.52e-05 | Epoch: 1/8
|
| 186 |
+
2026-06-24 15:32:21,523 [INFO] Step 25/3504 | Loss: 9.5130 | CE: 9.5130 | LR: 1.19e-04 | Epoch: 1/8
|
| 187 |
+
2026-06-24 15:32:29,807 [INFO] Step 30/3504 | Loss: 8.7934 | CE: 8.7934 | LR: 1.43e-04 | Epoch: 1/8
|
| 188 |
+
2026-06-24 15:32:40,131 [INFO] Step 35/3504 | Loss: 7.9035 | CE: 7.9035 | LR: 1.67e-04 | Epoch: 1/8
|
| 189 |
+
2026-06-24 15:32:48,824 [INFO] Step 40/3504 | Loss: 7.1752 | CE: 7.1752 | LR: 1.90e-04 | Epoch: 1/8
|
| 190 |
+
2026-06-24 15:32:56,822 [INFO] Step 45/3504 | Loss: 6.7221 | CE: 6.7221 | LR: 2.14e-04 | Epoch: 1/8
|
| 191 |
+
2026-06-24 15:33:04,152 [INFO] Step 50/3504 | Loss: 6.4773 | CE: 6.4773 | LR: 2.38e-04 | Epoch: 1/8
|
| 192 |
+
2026-06-24 15:33:12,698 [INFO] Step 55/3504 | Loss: 6.3695 | CE: 6.3695 | LR: 2.62e-04 | Epoch: 1/8
|
| 193 |
+
2026-06-24 15:33:22,047 [INFO] Step 60/3504 | Loss: 6.2721 | CE: 6.2721 | LR: 2.86e-04 | Epoch: 1/8
|
| 194 |
+
2026-06-24 15:33:30,494 [INFO] Step 65/3504 | Loss: 6.1980 | CE: 6.1980 | LR: 3.10e-04 | Epoch: 1/8
|
| 195 |
+
2026-06-24 15:33:40,655 [INFO] Step 70/3504 | Loss: 5.9787 | CE: 5.9787 | LR: 3.33e-04 | Epoch: 1/8
|
| 196 |
+
2026-06-24 15:33:50,222 [INFO] Step 75/3504 | Loss: 6.0126 | CE: 6.0126 | LR: 3.57e-04 | Epoch: 1/8
|
| 197 |
+
2026-06-24 15:33:58,220 [INFO] Step 80/3504 | Loss: 5.9178 | CE: 5.9178 | LR: 3.81e-04 | Epoch: 1/8
|
| 198 |
+
2026-06-24 15:34:08,946 [INFO] Step 85/3504 | Loss: 5.7340 | CE: 5.7340 | LR: 4.05e-04 | Epoch: 1/8
|
| 199 |
+
2026-06-24 15:34:16,988 [INFO] Step 90/3504 | Loss: 5.7256 | CE: 5.7256 | LR: 4.29e-04 | Epoch: 1/8
|
| 200 |
+
2026-06-24 15:34:26,742 [INFO] Step 95/3504 | Loss: 5.8148 | CE: 5.8148 | LR: 4.52e-04 | Epoch: 1/8
|
| 201 |
+
2026-06-24 15:34:37,216 [INFO] Step 100/3504 | Loss: 5.5746 | CE: 5.5746 | LR: 4.76e-04 | Epoch: 1/8
|
| 202 |
+
2026-06-24 15:34:46,609 [INFO] Step 105/3504 | Loss: 5.4389 | CE: 5.4389 | LR: 5.00e-04 | Epoch: 1/8
|
| 203 |
+
2026-06-24 15:34:55,774 [INFO] Step 110/3504 | Loss: 5.4149 | CE: 5.4149 | LR: 5.00e-04 | Epoch: 1/8
|
| 204 |
+
2026-06-24 15:35:05,496 [INFO] Step 115/3504 | Loss: 5.4051 | CE: 5.4051 | LR: 5.00e-04 | Epoch: 1/8
|
| 205 |
+
2026-06-24 15:35:13,049 [INFO] Step 120/3504 | Loss: 5.3995 | CE: 5.3995 | LR: 5.00e-04 | Epoch: 1/8
|
| 206 |
+
2026-06-24 15:35:21,806 [INFO] Step 125/3504 | Loss: 5.1879 | CE: 5.1879 | LR: 5.00e-04 | Epoch: 1/8
|
| 207 |
+
2026-06-24 15:35:30,248 [INFO] Step 130/3504 | Loss: 5.1576 | CE: 5.1576 | LR: 5.00e-04 | Epoch: 1/8
|
| 208 |
+
2026-06-24 15:35:38,761 [INFO] Step 135/3504 | Loss: 5.0880 | CE: 5.0880 | LR: 5.00e-04 | Epoch: 1/8
|
| 209 |
+
2026-06-24 15:35:47,704 [INFO] Step 140/3504 | Loss: 5.1887 | CE: 5.1887 | LR: 5.00e-04 | Epoch: 1/8
|
| 210 |
+
2026-06-24 15:35:56,598 [INFO] Step 145/3504 | Loss: 5.0271 | CE: 5.0271 | LR: 5.00e-04 | Epoch: 1/8
|
| 211 |
+
2026-06-24 15:36:06,252 [INFO] Step 150/3504 | Loss: 5.1243 | CE: 5.1243 | LR: 5.00e-04 | Epoch: 1/8
|
| 212 |
+
2026-06-24 15:36:15,062 [INFO] Step 155/3504 | Loss: 5.0605 | CE: 5.0605 | LR: 5.00e-04 | Epoch: 1/8
|
| 213 |
+
2026-06-24 15:36:23,329 [INFO] Step 160/3504 | Loss: 4.8040 | CE: 4.8040 | LR: 5.00e-04 | Epoch: 1/8
|
| 214 |
+
2026-06-24 15:36:30,871 [INFO] Step 165/3504 | Loss: 4.9998 | CE: 4.9998 | LR: 5.00e-04 | Epoch: 1/8
|
| 215 |
+
2026-06-24 15:36:39,747 [INFO] Step 170/3504 | Loss: 4.8400 | CE: 4.8400 | LR: 5.00e-04 | Epoch: 1/8
|
| 216 |
+
2026-06-24 15:36:48,606 [INFO] Step 175/3504 | Loss: 4.8995 | CE: 4.8995 | LR: 4.99e-04 | Epoch: 1/8
|
| 217 |
+
2026-06-24 15:36:56,562 [INFO] Step 180/3504 | Loss: 4.8623 | CE: 4.8623 | LR: 4.99e-04 | Epoch: 1/8
|
| 218 |
+
2026-06-24 15:37:05,010 [INFO] Step 185/3504 | Loss: 4.7630 | CE: 4.7630 | LR: 4.99e-04 | Epoch: 1/8
|
| 219 |
+
2026-06-24 15:37:15,446 [INFO] Step 190/3504 | Loss: 4.8097 | CE: 4.8097 | LR: 4.99e-04 | Epoch: 1/8
|
| 220 |
+
2026-06-24 15:37:24,808 [INFO] Step 195/3504 | Loss: 4.7014 | CE: 4.7014 | LR: 4.99e-04 | Epoch: 1/8
|
| 221 |
+
2026-06-24 15:37:32,106 [INFO] Step 200/3504 | Loss: 4.6825 | CE: 4.6825 | LR: 4.99e-04 | Epoch: 1/8
|
| 222 |
+
2026-06-24 15:37:41,291 [INFO] Step 205/3504 | Loss: 4.5497 | CE: 4.5497 | LR: 4.99e-04 | Epoch: 1/8
|
| 223 |
+
2026-06-24 15:37:49,372 [INFO] Step 210/3504 | Loss: 4.6699 | CE: 4.6699 | LR: 4.99e-04 | Epoch: 1/8
|
| 224 |
+
2026-06-24 15:37:58,576 [INFO] Step 215/3504 | Loss: 4.3792 | CE: 4.3792 | LR: 4.99e-04 | Epoch: 1/8
|
| 225 |
+
2026-06-24 15:38:06,586 [INFO] Step 220/3504 | Loss: 4.4395 | CE: 4.4395 | LR: 4.99e-04 | Epoch: 1/8
|
| 226 |
+
2026-06-24 15:38:15,075 [INFO] Step 225/3504 | Loss: 4.4622 | CE: 4.4622 | LR: 4.98e-04 | Epoch: 1/8
|
| 227 |
+
2026-06-24 15:38:24,209 [INFO] Step 230/3504 | Loss: 4.3598 | CE: 4.3598 | LR: 4.98e-04 | Epoch: 1/8
|
| 228 |
+
2026-06-24 15:38:33,251 [INFO] Step 235/3504 | Loss: 4.4105 | CE: 4.4105 | LR: 4.98e-04 | Epoch: 1/8
|
| 229 |
+
2026-06-24 15:38:42,370 [INFO] Step 240/3504 | Loss: 4.3439 | CE: 4.3439 | LR: 4.98e-04 | Epoch: 1/8
|
| 230 |
+
2026-06-24 15:38:51,333 [INFO] Step 245/3504 | Loss: 4.3100 | CE: 4.3100 | LR: 4.98e-04 | Epoch: 1/8
|
| 231 |
+
2026-06-24 15:38:59,392 [INFO] Step 250/3504 | Loss: 4.3957 | CE: 4.3957 | LR: 4.98e-04 | Epoch: 1/8
|
| 232 |
+
2026-06-24 15:39:07,956 [INFO] Step 255/3504 | Loss: 4.4233 | CE: 4.4233 | LR: 4.98e-04 | Epoch: 1/8
|
| 233 |
+
2026-06-24 15:39:16,990 [INFO] Step 260/3504 | Loss: 4.3814 | CE: 4.3814 | LR: 4.97e-04 | Epoch: 1/8
|
| 234 |
+
2026-06-24 15:39:24,982 [INFO] Step 265/3504 | Loss: 4.2711 | CE: 4.2711 | LR: 4.97e-04 | Epoch: 1/8
|
| 235 |
+
2026-06-24 15:39:32,608 [INFO] Step 270/3504 | Loss: 4.2424 | CE: 4.2424 | LR: 4.97e-04 | Epoch: 1/8
|
| 236 |
+
2026-06-24 15:39:41,555 [INFO] Step 275/3504 | Loss: 4.2557 | CE: 4.2557 | LR: 4.97e-04 | Epoch: 1/8
|
| 237 |
+
2026-06-24 15:39:51,952 [INFO] Step 280/3504 | Loss: 4.2074 | CE: 4.2074 | LR: 4.97e-04 | Epoch: 1/8
|
| 238 |
+
2026-06-24 15:40:00,146 [INFO] Step 285/3504 | Loss: 4.1360 | CE: 4.1360 | LR: 4.97e-04 | Epoch: 1/8
|
| 239 |
+
2026-06-24 15:40:09,409 [INFO] Step 290/3504 | Loss: 4.2171 | CE: 4.2171 | LR: 4.96e-04 | Epoch: 1/8
|
| 240 |
+
2026-06-24 15:40:18,770 [INFO] Step 295/3504 | Loss: 4.0780 | CE: 4.0780 | LR: 4.96e-04 | Epoch: 1/8
|
| 241 |
+
2026-06-24 15:40:27,995 [INFO] Step 300/3504 | Loss: 4.1449 | CE: 4.1449 | LR: 4.96e-04 | Epoch: 1/8
|
| 242 |
+
2026-06-24 15:40:36,786 [INFO] Step 305/3504 | Loss: 4.0702 | CE: 4.0702 | LR: 4.96e-04 | Epoch: 1/8
|
| 243 |
+
2026-06-24 15:40:45,158 [INFO] Step 310/3504 | Loss: 4.0008 | CE: 4.0008 | LR: 4.96e-04 | Epoch: 1/8
|
| 244 |
+
2026-06-24 15:40:53,152 [INFO] Step 315/3504 | Loss: 4.0473 | CE: 4.0473 | LR: 4.95e-04 | Epoch: 1/8
|
| 245 |
+
2026-06-24 15:41:00,626 [INFO] Step 320/3504 | Loss: 4.1152 | CE: 4.1152 | LR: 4.95e-04 | Epoch: 1/8
|
| 246 |
+
2026-06-24 15:41:09,458 [INFO] Step 325/3504 | Loss: 4.1337 | CE: 4.1337 | LR: 4.95e-04 | Epoch: 1/8
|
| 247 |
+
2026-06-24 15:41:18,543 [INFO] Step 330/3504 | Loss: 3.9870 | CE: 3.9870 | LR: 4.95e-04 | Epoch: 1/8
|
| 248 |
+
2026-06-24 15:41:26,576 [INFO] Step 335/3504 | Loss: 3.9279 | CE: 3.9279 | LR: 4.94e-04 | Epoch: 1/8
|
| 249 |
+
2026-06-24 15:41:36,137 [INFO] Step 340/3504 | Loss: 3.9647 | CE: 3.9647 | LR: 4.94e-04 | Epoch: 1/8
|
| 250 |
+
2026-06-24 15:41:44,715 [INFO] Step 345/3504 | Loss: 3.9517 | CE: 3.9517 | LR: 4.94e-04 | Epoch: 1/8
|
| 251 |
+
2026-06-24 15:41:53,341 [INFO] Step 350/3504 | Loss: 3.7424 | CE: 3.7424 | LR: 4.94e-04 | Epoch: 1/8
|
| 252 |
+
2026-06-24 15:42:01,095 [INFO] Step 355/3504 | Loss: 3.8924 | CE: 3.8924 | LR: 4.93e-04 | Epoch: 1/8
|
| 253 |
+
2026-06-24 15:42:09,195 [INFO] Step 360/3504 | Loss: 3.8113 | CE: 3.8113 | LR: 4.93e-04 | Epoch: 1/8
|
| 254 |
+
2026-06-24 15:42:16,598 [INFO] Step 365/3504 | Loss: 3.8940 | CE: 3.8940 | LR: 4.93e-04 | Epoch: 1/8
|
| 255 |
+
2026-06-24 15:42:24,467 [INFO] Step 370/3504 | Loss: 3.9236 | CE: 3.9236 | LR: 4.93e-04 | Epoch: 1/8
|
| 256 |
+
2026-06-24 15:42:33,121 [INFO] Step 375/3504 | Loss: 3.8203 | CE: 3.8203 | LR: 4.92e-04 | Epoch: 1/8
|
| 257 |
+
2026-06-24 15:42:42,556 [INFO] Step 380/3504 | Loss: 3.8788 | CE: 3.8788 | LR: 4.92e-04 | Epoch: 1/8
|
| 258 |
+
2026-06-24 15:42:51,295 [INFO] Step 385/3504 | Loss: 3.6876 | CE: 3.6876 | LR: 4.92e-04 | Epoch: 1/8
|
| 259 |
+
2026-06-24 15:42:59,654 [INFO] Step 390/3504 | Loss: 3.8804 | CE: 3.8804 | LR: 4.91e-04 | Epoch: 1/8
|
| 260 |
+
2026-06-24 15:43:08,628 [INFO] Step 395/3504 | Loss: 3.8794 | CE: 3.8794 | LR: 4.91e-04 | Epoch: 1/8
|
| 261 |
+
2026-06-24 15:43:17,072 [INFO] Step 400/3504 | Loss: 3.5935 | CE: 3.5935 | LR: 4.91e-04 | Epoch: 1/8
|
| 262 |
+
2026-06-24 15:43:25,878 [INFO] Step 405/3504 | Loss: 3.7812 | CE: 3.7812 | LR: 4.90e-04 | Epoch: 1/8
|
| 263 |
+
2026-06-24 15:43:34,449 [INFO] Step 410/3504 | Loss: 3.6309 | CE: 3.6309 | LR: 4.90e-04 | Epoch: 1/8
|
| 264 |
+
2026-06-24 15:43:42,129 [INFO] Step 415/3504 | Loss: 3.6847 | CE: 3.6847 | LR: 4.90e-04 | Epoch: 1/8
|
| 265 |
+
2026-06-24 15:43:50,565 [INFO] Step 420/3504 | Loss: 3.7767 | CE: 3.7767 | LR: 4.89e-04 | Epoch: 1/8
|
| 266 |
+
2026-06-24 15:43:58,515 [INFO] Step 425/3504 | Loss: 3.7170 | CE: 3.7170 | LR: 4.89e-04 | Epoch: 1/8
|
| 267 |
+
2026-06-24 15:44:06,691 [INFO] Step 430/3504 | Loss: 3.6850 | CE: 3.6850 | LR: 4.89e-04 | Epoch: 1/8
|
| 268 |
+
2026-06-24 15:44:15,604 [INFO] Step 435/3504 | Loss: 3.5626 | CE: 3.5626 | LR: 4.88e-04 | Epoch: 1/8
|
| 269 |
+
2026-06-24 15:45:18,994 [INFO] Step 438 | Eval Loss: 3.6298
|
| 270 |
+
2026-06-24 15:45:20,935 [INFO] Saved trainable-only checkpoint to runs/phase1_warmup_qwen_wikilingua/best.pt (183 model tensors, stores_base_encoder_weights=False)
|
| 271 |
+
2026-06-24 15:45:22,278 [INFO] New best eval loss: 3.6298
|
| 272 |
+
2026-06-24 15:45:25,767 [INFO] Step 440/3504 | Loss: 3.5792 | CE: 3.5792 | LR: 4.88e-04 | Epoch: 2/8
|
| 273 |
+
2026-06-24 15:45:34,269 [INFO] Step 445/3504 | Loss: 3.6109 | CE: 3.6109 | LR: 4.88e-04 | Epoch: 2/8
|
| 274 |
+
2026-06-24 15:45:40,979 [INFO] Step 450/3504 | Loss: 3.4676 | CE: 3.4676 | LR: 4.87e-04 | Epoch: 2/8
|
| 275 |
+
2026-06-24 15:45:49,011 [INFO] Step 455/3504 | Loss: 3.4070 | CE: 3.4070 | LR: 4.87e-04 | Epoch: 2/8
|
| 276 |
+
2026-06-24 15:45:58,773 [INFO] Step 460/3504 | Loss: 3.3765 | CE: 3.3765 | LR: 4.87e-04 | Epoch: 2/8
|
| 277 |
+
2026-06-24 15:46:06,993 [INFO] Step 465/3504 | Loss: 3.5223 | CE: 3.5223 | LR: 4.86e-04 | Epoch: 2/8
|
| 278 |
+
2026-06-24 15:46:16,610 [INFO] Step 470/3504 | Loss: 3.5444 | CE: 3.5444 | LR: 4.86e-04 | Epoch: 2/8
|
| 279 |
+
2026-06-24 15:46:23,973 [INFO] Step 475/3504 | Loss: 3.4474 | CE: 3.4474 | LR: 4.86e-04 | Epoch: 2/8
|
| 280 |
+
2026-06-24 15:46:31,671 [INFO] Step 480/3504 | Loss: 3.3227 | CE: 3.3227 | LR: 4.85e-04 | Epoch: 2/8
|
| 281 |
+
2026-06-24 15:46:39,912 [INFO] Step 485/3504 | Loss: 3.4797 | CE: 3.4797 | LR: 4.85e-04 | Epoch: 2/8
|
| 282 |
+
2026-06-24 15:46:49,352 [INFO] Step 490/3504 | Loss: 3.2878 | CE: 3.2878 | LR: 4.84e-04 | Epoch: 2/8
|
| 283 |
+
2026-06-24 15:46:57,458 [INFO] Step 495/3504 | Loss: 3.4894 | CE: 3.4894 | LR: 4.84e-04 | Epoch: 2/8
|
| 284 |
+
2026-06-24 15:47:05,556 [INFO] Step 500/3504 | Loss: 3.3222 | CE: 3.3222 | LR: 4.84e-04 | Epoch: 2/8
|
| 285 |
+
2026-06-24 15:47:12,834 [INFO] Step 505/3504 | Loss: 3.4277 | CE: 3.4277 | LR: 4.83e-04 | Epoch: 2/8
|
| 286 |
+
2026-06-24 15:47:22,429 [INFO] Step 510/3504 | Loss: 3.3461 | CE: 3.3461 | LR: 4.83e-04 | Epoch: 2/8
|
| 287 |
+
2026-06-24 15:47:30,596 [INFO] Step 515/3504 | Loss: 3.4142 | CE: 3.4142 | LR: 4.82e-04 | Epoch: 2/8
|
| 288 |
+
2026-06-24 15:47:39,359 [INFO] Step 520/3504 | Loss: 3.3020 | CE: 3.3020 | LR: 4.82e-04 | Epoch: 2/8
|
| 289 |
+
2026-06-24 15:47:49,251 [INFO] Step 525/3504 | Loss: 3.3385 | CE: 3.3385 | LR: 4.81e-04 | Epoch: 2/8
|
| 290 |
+
2026-06-24 15:47:56,891 [INFO] Step 530/3504 | Loss: 3.3764 | CE: 3.3764 | LR: 4.81e-04 | Epoch: 2/8
|
| 291 |
+
2026-06-24 15:48:06,377 [INFO] Step 535/3504 | Loss: 3.3064 | CE: 3.3064 | LR: 4.81e-04 | Epoch: 2/8
|
| 292 |
+
2026-06-24 15:48:13,969 [INFO] Step 540/3504 | Loss: 3.5091 | CE: 3.5091 | LR: 4.80e-04 | Epoch: 2/8
|
| 293 |
+
2026-06-24 15:48:23,044 [INFO] Step 545/3504 | Loss: 3.3540 | CE: 3.3540 | LR: 4.80e-04 | Epoch: 2/8
|
| 294 |
+
2026-06-24 15:48:30,715 [INFO] Step 550/3504 | Loss: 3.3091 | CE: 3.3091 | LR: 4.79e-04 | Epoch: 2/8
|
| 295 |
+
2026-06-24 15:48:40,289 [INFO] Step 555/3504 | Loss: 3.3155 | CE: 3.3155 | LR: 4.79e-04 | Epoch: 2/8
|
| 296 |
+
2026-06-24 15:48:48,757 [INFO] Step 560/3504 | Loss: 3.2676 | CE: 3.2676 | LR: 4.78e-04 | Epoch: 2/8
|
| 297 |
+
2026-06-24 15:48:57,052 [INFO] Step 565/3504 | Loss: 3.3811 | CE: 3.3811 | LR: 4.78e-04 | Epoch: 2/8
|
| 298 |
+
2026-06-24 15:49:06,906 [INFO] Step 570/3504 | Loss: 3.1720 | CE: 3.1720 | LR: 4.77e-04 | Epoch: 2/8
|
| 299 |
+
2026-06-24 15:49:15,033 [INFO] Step 575/3504 | Loss: 3.3162 | CE: 3.3162 | LR: 4.77e-04 | Epoch: 2/8
|
| 300 |
+
2026-06-24 15:49:23,097 [INFO] Step 580/3504 | Loss: 3.4106 | CE: 3.4106 | LR: 4.76e-04 | Epoch: 2/8
|
| 301 |
+
2026-06-24 15:49:32,560 [INFO] Step 585/3504 | Loss: 3.4298 | CE: 3.4298 | LR: 4.76e-04 | Epoch: 2/8
|
| 302 |
+
2026-06-24 15:49:41,204 [INFO] Step 590/3504 | Loss: 3.2985 | CE: 3.2985 | LR: 4.75e-04 | Epoch: 2/8
|
| 303 |
+
2026-06-24 15:49:49,348 [INFO] Step 595/3504 | Loss: 3.3157 | CE: 3.3157 | LR: 4.75e-04 | Epoch: 2/8
|
| 304 |
+
2026-06-24 15:49:57,631 [INFO] Step 600/3504 | Loss: 3.3765 | CE: 3.3765 | LR: 4.74e-04 | Epoch: 2/8
|
| 305 |
+
2026-06-24 15:50:08,032 [INFO] Step 605/3504 | Loss: 3.2169 | CE: 3.2169 | LR: 4.74e-04 | Epoch: 2/8
|
| 306 |
+
2026-06-24 15:50:16,046 [INFO] Step 610/3504 | Loss: 3.2954 | CE: 3.2954 | LR: 4.73e-04 | Epoch: 2/8
|
| 307 |
+
2026-06-24 15:50:27,112 [INFO] Step 615/3504 | Loss: 3.3336 | CE: 3.3336 | LR: 4.73e-04 | Epoch: 2/8
|
| 308 |
+
2026-06-24 15:50:35,979 [INFO] Step 620/3504 | Loss: 3.3302 | CE: 3.3302 | LR: 4.72e-04 | Epoch: 2/8
|
| 309 |
+
2026-06-24 15:50:44,399 [INFO] Step 625/3504 | Loss: 3.1371 | CE: 3.1371 | LR: 4.72e-04 | Epoch: 2/8
|
| 310 |
+
2026-06-24 15:50:54,632 [INFO] Step 630/3504 | Loss: 3.2272 | CE: 3.2272 | LR: 4.71e-04 | Epoch: 2/8
|
| 311 |
+
2026-06-24 15:51:03,736 [INFO] Step 635/3504 | Loss: 3.2678 | CE: 3.2678 | LR: 4.71e-04 | Epoch: 2/8
|
| 312 |
+
2026-06-24 15:51:13,033 [INFO] Step 640/3504 | Loss: 3.2136 | CE: 3.2136 | LR: 4.70e-04 | Epoch: 2/8
|
| 313 |
+
2026-06-24 15:51:21,949 [INFO] Step 645/3504 | Loss: 3.2889 | CE: 3.2889 | LR: 4.70e-04 | Epoch: 2/8
|
| 314 |
+
2026-06-24 15:51:31,244 [INFO] Step 650/3504 | Loss: 3.2790 | CE: 3.2790 | LR: 4.69e-04 | Epoch: 2/8
|
| 315 |
+
2026-06-24 15:51:39,273 [INFO] Step 655/3504 | Loss: 3.2365 | CE: 3.2365 | LR: 4.68e-04 | Epoch: 2/8
|
| 316 |
+
2026-06-24 15:51:47,093 [INFO] Step 660/3504 | Loss: 3.2323 | CE: 3.2323 | LR: 4.68e-04 | Epoch: 2/8
|
| 317 |
+
2026-06-24 15:51:55,711 [INFO] Step 665/3504 | Loss: 3.3179 | CE: 3.3179 | LR: 4.67e-04 | Epoch: 2/8
|
| 318 |
+
2026-06-24 15:52:03,561 [INFO] Step 670/3504 | Loss: 3.1795 | CE: 3.1795 | LR: 4.67e-04 | Epoch: 2/8
|
| 319 |
+
2026-06-24 15:52:13,655 [INFO] Step 675/3504 | Loss: 3.1378 | CE: 3.1378 | LR: 4.66e-04 | Epoch: 2/8
|
| 320 |
+
2026-06-24 15:52:22,719 [INFO] Step 680/3504 | Loss: 3.1014 | CE: 3.1014 | LR: 4.66e-04 | Epoch: 2/8
|
| 321 |
+
2026-06-24 15:52:31,552 [INFO] Step 685/3504 | Loss: 3.1685 | CE: 3.1685 | LR: 4.65e-04 | Epoch: 2/8
|
| 322 |
+
2026-06-24 15:52:40,529 [INFO] Step 690/3504 | Loss: 3.2212 | CE: 3.2212 | LR: 4.64e-04 | Epoch: 2/8
|
| 323 |
+
2026-06-24 15:52:48,443 [INFO] Step 695/3504 | Loss: 3.2689 | CE: 3.2689 | LR: 4.64e-04 | Epoch: 2/8
|
| 324 |
+
2026-06-24 15:52:55,201 [INFO] Step 700/3504 | Loss: 3.2432 | CE: 3.2432 | LR: 4.63e-04 | Epoch: 2/8
|
| 325 |
+
2026-06-24 15:53:04,178 [INFO] Step 705/3504 | Loss: 3.2091 | CE: 3.2091 | LR: 4.63e-04 | Epoch: 2/8
|
| 326 |
+
2026-06-24 15:53:13,122 [INFO] Step 710/3504 | Loss: 3.2886 | CE: 3.2886 | LR: 4.62e-04 | Epoch: 2/8
|
| 327 |
+
2026-06-24 15:53:21,204 [INFO] Step 715/3504 | Loss: 3.1453 | CE: 3.1453 | LR: 4.61e-04 | Epoch: 2/8
|
| 328 |
+
2026-06-24 15:53:29,937 [INFO] Step 720/3504 | Loss: 3.1662 | CE: 3.1662 | LR: 4.61e-04 | Epoch: 2/8
|
| 329 |
+
2026-06-24 15:53:38,020 [INFO] Step 725/3504 | Loss: 3.2001 | CE: 3.2001 | LR: 4.60e-04 | Epoch: 2/8
|
| 330 |
+
2026-06-24 15:53:48,049 [INFO] Step 730/3504 | Loss: 3.1201 | CE: 3.1201 | LR: 4.59e-04 | Epoch: 2/8
|
| 331 |
+
2026-06-24 15:53:56,398 [INFO] Step 735/3504 | Loss: 3.1014 | CE: 3.1014 | LR: 4.59e-04 | Epoch: 2/8
|
| 332 |
+
2026-06-24 15:54:04,454 [INFO] Step 740/3504 | Loss: 3.1933 | CE: 3.1933 | LR: 4.58e-04 | Epoch: 2/8
|
| 333 |
+
2026-06-24 15:54:14,432 [INFO] Step 745/3504 | Loss: 3.0940 | CE: 3.0940 | LR: 4.58e-04 | Epoch: 2/8
|
| 334 |
+
2026-06-24 15:54:22,402 [INFO] Step 750/3504 | Loss: 3.2106 | CE: 3.2106 | LR: 4.57e-04 | Epoch: 2/8
|
| 335 |
+
2026-06-24 15:54:30,887 [INFO] Step 755/3504 | Loss: 3.0634 | CE: 3.0634 | LR: 4.56e-04 | Epoch: 2/8
|
| 336 |
+
2026-06-24 15:54:39,257 [INFO] Step 760/3504 | Loss: 3.2546 | CE: 3.2546 | LR: 4.56e-04 | Epoch: 2/8
|
| 337 |
+
2026-06-24 15:54:48,912 [INFO] Step 765/3504 | Loss: 2.9755 | CE: 2.9755 | LR: 4.55e-04 | Epoch: 2/8
|
| 338 |
+
2026-06-24 15:54:56,711 [INFO] Step 770/3504 | Loss: 3.1427 | CE: 3.1427 | LR: 4.54e-04 | Epoch: 2/8
|
| 339 |
+
2026-06-24 15:55:06,678 [INFO] Step 775/3504 | Loss: 3.0667 | CE: 3.0667 | LR: 4.54e-04 | Epoch: 2/8
|
| 340 |
+
2026-06-24 15:55:15,806 [INFO] Step 780/3504 | Loss: 3.1215 | CE: 3.1215 | LR: 4.53e-04 | Epoch: 2/8
|
| 341 |
+
2026-06-24 15:55:25,013 [INFO] Step 785/3504 | Loss: 3.1915 | CE: 3.1915 | LR: 4.52e-04 | Epoch: 2/8
|
| 342 |
+
2026-06-24 15:55:34,327 [INFO] Step 790/3504 | Loss: 3.0573 | CE: 3.0573 | LR: 4.52e-04 | Epoch: 2/8
|
| 343 |
+
2026-06-24 15:55:43,640 [INFO] Step 795/3504 | Loss: 3.1904 | CE: 3.1904 | LR: 4.51e-04 | Epoch: 2/8
|
| 344 |
+
2026-06-24 15:55:52,767 [INFO] Step 800/3504 | Loss: 3.1951 | CE: 3.1951 | LR: 4.50e-04 | Epoch: 2/8
|
| 345 |
+
2026-06-24 15:56:02,785 [INFO] Step 805/3504 | Loss: 3.0239 | CE: 3.0239 | LR: 4.49e-04 | Epoch: 2/8
|
| 346 |
+
2026-06-24 15:56:10,806 [INFO] Step 810/3504 | Loss: 3.0280 | CE: 3.0280 | LR: 4.49e-04 | Epoch: 2/8
|
| 347 |
+
2026-06-24 15:56:21,206 [INFO] Step 815/3504 | Loss: 3.1670 | CE: 3.1670 | LR: 4.48e-04 | Epoch: 2/8
|
| 348 |
+
2026-06-24 15:56:29,252 [INFO] Step 820/3504 | Loss: 3.0084 | CE: 3.0084 | LR: 4.47e-04 | Epoch: 2/8
|
| 349 |
+
2026-06-24 15:56:38,450 [INFO] Step 825/3504 | Loss: 3.0182 | CE: 3.0182 | LR: 4.47e-04 | Epoch: 2/8
|
| 350 |
+
2026-06-24 15:56:45,895 [INFO] Step 830/3504 | Loss: 2.9384 | CE: 2.9384 | LR: 4.46e-04 | Epoch: 2/8
|
| 351 |
+
2026-06-24 15:56:54,402 [INFO] Step 835/3504 | Loss: 3.1581 | CE: 3.1581 | LR: 4.45e-04 | Epoch: 2/8
|
| 352 |
+
2026-06-24 15:57:03,602 [INFO] Step 840/3504 | Loss: 2.8852 | CE: 2.8852 | LR: 4.44e-04 | Epoch: 2/8
|
| 353 |
+
2026-06-24 15:57:12,113 [INFO] Step 845/3504 | Loss: 3.0107 | CE: 3.0107 | LR: 4.44e-04 | Epoch: 2/8
|
| 354 |
+
2026-06-24 15:57:21,205 [INFO] Step 850/3504 | Loss: 2.9701 | CE: 2.9701 | LR: 4.43e-04 | Epoch: 2/8
|
| 355 |
+
2026-06-24 15:57:30,065 [INFO] Step 855/3504 | Loss: 2.9731 | CE: 2.9731 | LR: 4.42e-04 | Epoch: 2/8
|
| 356 |
+
2026-06-24 15:57:38,813 [INFO] Step 860/3504 | Loss: 3.0085 | CE: 3.0085 | LR: 4.42e-04 | Epoch: 2/8
|
| 357 |
+
2026-06-24 15:57:47,588 [INFO] Step 865/3504 | Loss: 2.9527 | CE: 2.9527 | LR: 4.41e-04 | Epoch: 2/8
|
| 358 |
+
2026-06-24 15:57:56,133 [INFO] Step 870/3504 | Loss: 2.9545 | CE: 2.9545 | LR: 4.40e-04 | Epoch: 2/8
|
| 359 |
+
2026-06-24 15:58:05,483 [INFO] Step 875/3504 | Loss: 3.5718 | CE: 3.5718 | LR: 4.39e-04 | Epoch: 3/8
|
| 360 |
+
2026-06-24 15:59:05,387 [INFO] Step 876 | Eval Loss: 3.1011
|
| 361 |
+
2026-06-24 15:59:08,186 [INFO] Saved trainable-only checkpoint to runs/phase1_warmup_qwen_wikilingua/best.pt (183 model tensors, stores_base_encoder_weights=False)
|
| 362 |
+
2026-06-24 15:59:09,764 [INFO] New best eval loss: 3.1011
|
| 363 |
+
2026-06-24 15:59:16,212 [INFO] Step 880/3504 | Loss: 2.5805 | CE: 2.5805 | LR: 4.39e-04 | Epoch: 3/8
|
| 364 |
+
2026-06-24 15:59:25,836 [INFO] Step 885/3504 | Loss: 2.8468 | CE: 2.8468 | LR: 4.38e-04 | Epoch: 3/8
|
| 365 |
+
2026-06-24 15:59:33,215 [INFO] Step 890/3504 | Loss: 2.7945 | CE: 2.7945 | LR: 4.37e-04 | Epoch: 3/8
|
| 366 |
+
2026-06-24 15:59:42,569 [INFO] Step 895/3504 | Loss: 2.6429 | CE: 2.6429 | LR: 4.36e-04 | Epoch: 3/8
|
| 367 |
+
2026-06-24 15:59:51,033 [INFO] Step 900/3504 | Loss: 2.6209 | CE: 2.6209 | LR: 4.35e-04 | Epoch: 3/8
|
| 368 |
+
2026-06-24 15:59:59,535 [INFO] Step 905/3504 | Loss: 2.7512 | CE: 2.7512 | LR: 4.35e-04 | Epoch: 3/8
|
| 369 |
+
2026-06-24 16:00:06,664 [INFO] Step 910/3504 | Loss: 2.8121 | CE: 2.8121 | LR: 4.34e-04 | Epoch: 3/8
|
| 370 |
+
2026-06-24 16:00:16,615 [INFO] Step 915/3504 | Loss: 2.8310 | CE: 2.8310 | LR: 4.33e-04 | Epoch: 3/8
|
| 371 |
+
2026-06-24 16:00:26,583 [INFO] Step 920/3504 | Loss: 2.7239 | CE: 2.7239 | LR: 4.32e-04 | Epoch: 3/8
|
| 372 |
+
2026-06-24 16:00:36,345 [INFO] Step 925/3504 | Loss: 2.8202 | CE: 2.8202 | LR: 4.32e-04 | Epoch: 3/8
|
| 373 |
+
2026-06-24 16:00:43,723 [INFO] Step 930/3504 | Loss: 2.7002 | CE: 2.7002 | LR: 4.31e-04 | Epoch: 3/8
|
| 374 |
+
2026-06-24 16:00:54,146 [INFO] Step 935/3504 | Loss: 2.8012 | CE: 2.8012 | LR: 4.30e-04 | Epoch: 3/8
|
| 375 |
+
2026-06-24 16:01:03,982 [INFO] Step 940/3504 | Loss: 2.6946 | CE: 2.6946 | LR: 4.29e-04 | Epoch: 3/8
|
| 376 |
+
2026-06-24 16:01:10,801 [INFO] Step 945/3504 | Loss: 2.8363 | CE: 2.8363 | LR: 4.28e-04 | Epoch: 3/8
|
| 377 |
+
2026-06-24 16:01:19,107 [INFO] Step 950/3504 | Loss: 2.8266 | CE: 2.8266 | LR: 4.28e-04 | Epoch: 3/8
|
| 378 |
+
2026-06-24 16:01:28,303 [INFO] Step 955/3504 | Loss: 2.6443 | CE: 2.6443 | LR: 4.27e-04 | Epoch: 3/8
|
| 379 |
+
2026-06-24 16:01:37,540 [INFO] Step 960/3504 | Loss: 2.8581 | CE: 2.8581 | LR: 4.26e-04 | Epoch: 3/8
|
| 380 |
+
2026-06-24 16:01:45,956 [INFO] Step 965/3504 | Loss: 2.7262 | CE: 2.7262 | LR: 4.25e-04 | Epoch: 3/8
|
| 381 |
+
2026-06-24 16:01:56,002 [INFO] Step 970/3504 | Loss: 2.8103 | CE: 2.8103 | LR: 4.24e-04 | Epoch: 3/8
|
| 382 |
+
2026-06-24 16:02:05,649 [INFO] Step 975/3504 | Loss: 2.7571 | CE: 2.7571 | LR: 4.23e-04 | Epoch: 3/8
|
| 383 |
+
2026-06-24 16:02:13,766 [INFO] Step 980/3504 | Loss: 2.7804 | CE: 2.7804 | LR: 4.23e-04 | Epoch: 3/8
|
| 384 |
+
2026-06-24 16:02:22,356 [INFO] Step 985/3504 | Loss: 2.8489 | CE: 2.8489 | LR: 4.22e-04 | Epoch: 3/8
|
| 385 |
+
2026-06-24 16:02:30,398 [INFO] Step 990/3504 | Loss: 2.7532 | CE: 2.7532 | LR: 4.21e-04 | Epoch: 3/8
|
| 386 |
+
2026-06-24 16:02:38,819 [INFO] Step 995/3504 | Loss: 2.8773 | CE: 2.8773 | LR: 4.20e-04 | Epoch: 3/8
|
| 387 |
+
2026-06-24 16:02:47,824 [INFO] Step 1000/3504 | Loss: 2.6808 | CE: 2.6808 | LR: 4.19e-04 | Epoch: 3/8
|
| 388 |
+
2026-06-24 16:02:57,480 [INFO] Step 1005/3504 | Loss: 2.7219 | CE: 2.7219 | LR: 4.18e-04 | Epoch: 3/8
|
| 389 |
+
2026-06-24 16:03:06,413 [INFO] Step 1010/3504 | Loss: 2.8971 | CE: 2.8971 | LR: 4.18e-04 | Epoch: 3/8
|
| 390 |
+
2026-06-24 16:03:15,226 [INFO] Step 1015/3504 | Loss: 2.7598 | CE: 2.7598 | LR: 4.17e-04 | Epoch: 3/8
|
| 391 |
+
2026-06-24 16:03:22,709 [INFO] Step 1020/3504 | Loss: 2.6962 | CE: 2.6962 | LR: 4.16e-04 | Epoch: 3/8
|
| 392 |
+
2026-06-24 16:03:33,340 [INFO] Step 1025/3504 | Loss: 2.7058 | CE: 2.7058 | LR: 4.15e-04 | Epoch: 3/8
|
| 393 |
+
2026-06-24 16:03:40,523 [INFO] Step 1030/3504 | Loss: 2.7499 | CE: 2.7499 | LR: 4.14e-04 | Epoch: 3/8
|
| 394 |
+
2026-06-24 16:03:49,322 [INFO] Step 1035/3504 | Loss: 2.8416 | CE: 2.8416 | LR: 4.13e-04 | Epoch: 3/8
|
| 395 |
+
2026-06-24 16:03:57,585 [INFO] Step 1040/3504 | Loss: 2.8305 | CE: 2.8305 | LR: 4.12e-04 | Epoch: 3/8
|
| 396 |
+
2026-06-24 16:04:06,751 [INFO] Step 1045/3504 | Loss: 2.8956 | CE: 2.8956 | LR: 4.11e-04 | Epoch: 3/8
|
| 397 |
+
2026-06-24 16:04:14,301 [INFO] Step 1050/3504 | Loss: 2.7591 | CE: 2.7591 | LR: 4.11e-04 | Epoch: 3/8
|
| 398 |
+
2026-06-24 16:04:22,291 [INFO] Step 1055/3504 | Loss: 2.7619 | CE: 2.7619 | LR: 4.10e-04 | Epoch: 3/8
|
| 399 |
+
2026-06-24 16:04:31,412 [INFO] Step 1060/3504 | Loss: 2.6793 | CE: 2.6793 | LR: 4.09e-04 | Epoch: 3/8
|
| 400 |
+
2026-06-24 16:04:40,813 [INFO] Step 1065/3504 | Loss: 2.8954 | CE: 2.8954 | LR: 4.08e-04 | Epoch: 3/8
|
| 401 |
+
2026-06-24 16:04:48,763 [INFO] Step 1070/3504 | Loss: 2.7846 | CE: 2.7846 | LR: 4.07e-04 | Epoch: 3/8
|
| 402 |
+
2026-06-24 16:04:56,675 [INFO] Step 1075/3504 | Loss: 2.7975 | CE: 2.7975 | LR: 4.06e-04 | Epoch: 3/8
|
| 403 |
+
2026-06-24 16:05:06,103 [INFO] Step 1080/3504 | Loss: 2.6166 | CE: 2.6166 | LR: 4.05e-04 | Epoch: 3/8
|
| 404 |
+
2026-06-24 16:05:15,003 [INFO] Step 1085/3504 | Loss: 2.7101 | CE: 2.7101 | LR: 4.04e-04 | Epoch: 3/8
|
| 405 |
+
2026-06-24 16:05:23,250 [INFO] Step 1090/3504 | Loss: 2.7483 | CE: 2.7483 | LR: 4.03e-04 | Epoch: 3/8
|
| 406 |
+
2026-06-24 16:05:33,508 [INFO] Step 1095/3504 | Loss: 2.6977 | CE: 2.6977 | LR: 4.02e-04 | Epoch: 3/8
|
| 407 |
+
2026-06-24 16:05:41,764 [INFO] Step 1100/3504 | Loss: 2.6066 | CE: 2.6066 | LR: 4.02e-04 | Epoch: 3/8
|
| 408 |
+
2026-06-24 16:05:50,536 [INFO] Step 1105/3504 | Loss: 2.7113 | CE: 2.7113 | LR: 4.01e-04 | Epoch: 3/8
|
| 409 |
+
2026-06-24 16:06:00,019 [INFO] Step 1110/3504 | Loss: 2.6477 | CE: 2.6477 | LR: 4.00e-04 | Epoch: 3/8
|
| 410 |
+
2026-06-24 16:06:09,067 [INFO] Step 1115/3504 | Loss: 2.8057 | CE: 2.8057 | LR: 3.99e-04 | Epoch: 3/8
|
| 411 |
+
2026-06-24 16:06:16,914 [INFO] Step 1120/3504 | Loss: 2.8308 | CE: 2.8308 | LR: 3.98e-04 | Epoch: 3/8
|
| 412 |
+
2026-06-24 16:06:26,679 [INFO] Step 1125/3504 | Loss: 2.8634 | CE: 2.8634 | LR: 3.97e-04 | Epoch: 3/8
|
| 413 |
+
2026-06-24 16:06:33,947 [INFO] Step 1130/3504 | Loss: 2.7088 | CE: 2.7088 | LR: 3.96e-04 | Epoch: 3/8
|
| 414 |
+
2026-06-24 16:06:42,682 [INFO] Step 1135/3504 | Loss: 2.6773 | CE: 2.6773 | LR: 3.95e-04 | Epoch: 3/8
|
| 415 |
+
2026-06-24 16:06:51,430 [INFO] Step 1140/3504 | Loss: 2.7612 | CE: 2.7612 | LR: 3.94e-04 | Epoch: 3/8
|
| 416 |
+
2026-06-24 16:07:00,671 [INFO] Step 1145/3504 | Loss: 2.7048 | CE: 2.7048 | LR: 3.93e-04 | Epoch: 3/8
|
| 417 |
+
2026-06-24 16:07:09,898 [INFO] Step 1150/3504 | Loss: 2.7794 | CE: 2.7794 | LR: 3.92e-04 | Epoch: 3/8
|
| 418 |
+
2026-06-24 16:07:17,584 [INFO] Step 1155/3504 | Loss: 2.6423 | CE: 2.6423 | LR: 3.91e-04 | Epoch: 3/8
|
| 419 |
+
2026-06-24 16:07:24,905 [INFO] Step 1160/3504 | Loss: 2.8084 | CE: 2.8084 | LR: 3.90e-04 | Epoch: 3/8
|
| 420 |
+
2026-06-24 16:07:33,291 [INFO] Step 1165/3504 | Loss: 2.7718 | CE: 2.7718 | LR: 3.89e-04 | Epoch: 3/8
|
| 421 |
+
2026-06-24 16:07:43,151 [INFO] Step 1170/3504 | Loss: 2.8473 | CE: 2.8473 | LR: 3.88e-04 | Epoch: 3/8
|
| 422 |
+
2026-06-24 16:07:52,026 [INFO] Step 1175/3504 | Loss: 2.7020 | CE: 2.7020 | LR: 3.87e-04 | Epoch: 3/8
|
| 423 |
+
2026-06-24 16:08:00,930 [INFO] Step 1180/3504 | Loss: 2.6958 | CE: 2.6958 | LR: 3.86e-04 | Epoch: 3/8
|
| 424 |
+
2026-06-24 16:08:09,944 [INFO] Step 1185/3504 | Loss: 2.7750 | CE: 2.7750 | LR: 3.85e-04 | Epoch: 3/8
|
| 425 |
+
2026-06-24 16:08:19,459 [INFO] Step 1190/3504 | Loss: 2.7691 | CE: 2.7691 | LR: 3.84e-04 | Epoch: 3/8
|
| 426 |
+
2026-06-24 16:08:28,500 [INFO] Step 1195/3504 | Loss: 2.6779 | CE: 2.6779 | LR: 3.84e-04 | Epoch: 3/8
|
| 427 |
+
2026-06-24 16:08:38,625 [INFO] Step 1200/3504 | Loss: 2.6884 | CE: 2.6884 | LR: 3.83e-04 | Epoch: 3/8
|
| 428 |
+
2026-06-24 16:08:47,117 [INFO] Step 1205/3504 | Loss: 2.6584 | CE: 2.6584 | LR: 3.82e-04 | Epoch: 3/8
|
| 429 |
+
2026-06-24 16:08:55,562 [INFO] Step 1210/3504 | Loss: 2.5988 | CE: 2.5988 | LR: 3.81e-04 | Epoch: 3/8
|
| 430 |
+
2026-06-24 16:09:04,026 [INFO] Step 1215/3504 | Loss: 2.5267 | CE: 2.5267 | LR: 3.80e-04 | Epoch: 3/8
|
| 431 |
+
2026-06-24 16:09:12,900 [INFO] Step 1220/3504 | Loss: 2.6702 | CE: 2.6702 | LR: 3.79e-04 | Epoch: 3/8
|
| 432 |
+
2026-06-24 16:09:21,312 [INFO] Step 1225/3504 | Loss: 2.6465 | CE: 2.6465 | LR: 3.78e-04 | Epoch: 3/8
|
| 433 |
+
2026-06-24 16:09:29,118 [INFO] Step 1230/3504 | Loss: 2.5808 | CE: 2.5808 | LR: 3.77e-04 | Epoch: 3/8
|
| 434 |
+
2026-06-24 16:09:37,737 [INFO] Step 1235/3504 | Loss: 2.6339 | CE: 2.6339 | LR: 3.76e-04 | Epoch: 3/8
|
| 435 |
+
2026-06-24 16:09:47,962 [INFO] Step 1240/3504 | Loss: 2.6028 | CE: 2.6028 | LR: 3.75e-04 | Epoch: 3/8
|
| 436 |
+
2026-06-24 16:09:57,103 [INFO] Step 1245/3504 | Loss: 2.5635 | CE: 2.5635 | LR: 3.74e-04 | Epoch: 3/8
|
| 437 |
+
2026-06-24 16:10:07,450 [INFO] Step 1250/3504 | Loss: 2.6883 | CE: 2.6883 | LR: 3.73e-04 | Epoch: 3/8
|
| 438 |
+
2026-06-24 16:10:16,466 [INFO] Step 1255/3504 | Loss: 2.6264 | CE: 2.6264 | LR: 3.72e-04 | Epoch: 3/8
|
| 439 |
+
2026-06-24 16:10:24,245 [INFO] Step 1260/3504 | Loss: 2.6215 | CE: 2.6215 | LR: 3.71e-04 | Epoch: 3/8
|
| 440 |
+
2026-06-24 16:10:32,710 [INFO] Step 1265/3504 | Loss: 2.7072 | CE: 2.7072 | LR: 3.70e-04 | Epoch: 3/8
|
| 441 |
+
2026-06-24 16:10:40,159 [INFO] Step 1270/3504 | Loss: 2.6494 | CE: 2.6494 | LR: 3.69e-04 | Epoch: 3/8
|
| 442 |
+
2026-06-24 16:10:49,333 [INFO] Step 1275/3504 | Loss: 2.5466 | CE: 2.5466 | LR: 3.68e-04 | Epoch: 3/8
|
| 443 |
+
2026-06-24 16:10:59,096 [INFO] Step 1280/3504 | Loss: 2.5283 | CE: 2.5283 | LR: 3.67e-04 | Epoch: 3/8
|
| 444 |
+
2026-06-24 16:11:07,685 [INFO] Step 1285/3504 | Loss: 2.7235 | CE: 2.7235 | LR: 3.65e-04 | Epoch: 3/8
|
| 445 |
+
2026-06-24 16:11:14,636 [INFO] Step 1290/3504 | Loss: 2.6731 | CE: 2.6731 | LR: 3.64e-04 | Epoch: 3/8
|
| 446 |
+
2026-06-24 16:11:23,000 [INFO] Step 1295/3504 | Loss: 2.6776 | CE: 2.6776 | LR: 3.63e-04 | Epoch: 3/8
|
| 447 |
+
2026-06-24 16:11:31,405 [INFO] Step 1300/3504 | Loss: 2.7171 | CE: 2.7171 | LR: 3.62e-04 | Epoch: 3/8
|
| 448 |
+
2026-06-24 16:11:40,063 [INFO] Step 1305/3504 | Loss: 2.6887 | CE: 2.6887 | LR: 3.61e-04 | Epoch: 3/8
|
| 449 |
+
2026-06-24 16:11:50,419 [INFO] Step 1310/3504 | Loss: 2.6924 | CE: 2.6924 | LR: 3.60e-04 | Epoch: 3/8
|
| 450 |
+
2026-06-24 16:12:55,729 [INFO] Step 1314 | Eval Loss: 2.8291
|
| 451 |
+
2026-06-24 16:12:58,550 [INFO] Saved trainable-only checkpoint to runs/phase1_warmup_qwen_wikilingua/best.pt (183 model tensors, stores_base_encoder_weights=False)
|
| 452 |
+
2026-06-24 16:12:59,989 [INFO] New best eval loss: 2.8291
|
| 453 |
+
2026-06-24 16:13:02,016 [INFO] Step 1315/3504 | Loss: 2.4556 | CE: 2.4556 | LR: 3.59e-04 | Epoch: 4/8
|
| 454 |
+
2026-06-24 16:13:10,924 [INFO] Step 1320/3504 | Loss: 2.3209 | CE: 2.3209 | LR: 3.58e-04 | Epoch: 4/8
|
| 455 |
+
2026-06-24 16:13:20,158 [INFO] Step 1325/3504 | Loss: 2.4224 | CE: 2.4224 | LR: 3.57e-04 | Epoch: 4/8
|
| 456 |
+
2026-06-24 16:13:30,265 [INFO] Step 1330/3504 | Loss: 2.2311 | CE: 2.2311 | LR: 3.56e-04 | Epoch: 4/8
|
| 457 |
+
2026-06-24 16:13:38,574 [INFO] Step 1335/3504 | Loss: 2.4096 | CE: 2.4096 | LR: 3.55e-04 | Epoch: 4/8
|
| 458 |
+
2026-06-24 16:13:47,021 [INFO] Step 1340/3504 | Loss: 2.3838 | CE: 2.3838 | LR: 3.54e-04 | Epoch: 4/8
|
| 459 |
+
2026-06-24 16:13:54,638 [INFO] Step 1345/3504 | Loss: 2.5735 | CE: 2.5735 | LR: 3.53e-04 | Epoch: 4/8
|
| 460 |
+
2026-06-24 16:14:03,380 [INFO] Step 1350/3504 | Loss: 2.5266 | CE: 2.5266 | LR: 3.52e-04 | Epoch: 4/8
|
| 461 |
+
2026-06-24 16:14:11,477 [INFO] Step 1355/3504 | Loss: 2.7047 | CE: 2.7047 | LR: 3.51e-04 | Epoch: 4/8
|
| 462 |
+
2026-06-24 16:14:20,793 [INFO] Step 1360/3504 | Loss: 2.5692 | CE: 2.5692 | LR: 3.50e-04 | Epoch: 4/8
|
| 463 |
+
2026-06-24 16:14:30,565 [INFO] Step 1365/3504 | Loss: 2.4392 | CE: 2.4392 | LR: 3.49e-04 | Epoch: 4/8
|
| 464 |
+
2026-06-24 16:14:39,100 [INFO] Step 1370/3504 | Loss: 2.3793 | CE: 2.3793 | LR: 3.48e-04 | Epoch: 4/8
|
| 465 |
+
2026-06-24 16:14:48,495 [INFO] Step 1375/3504 | Loss: 2.3064 | CE: 2.3064 | LR: 3.47e-04 | Epoch: 4/8
|
| 466 |
+
2026-06-24 16:14:56,513 [INFO] Step 1380/3504 | Loss: 2.4392 | CE: 2.4392 | LR: 3.46e-04 | Epoch: 4/8
|
| 467 |
+
2026-06-24 16:15:04,566 [INFO] Step 1385/3504 | Loss: 2.3351 | CE: 2.3351 | LR: 3.45e-04 | Epoch: 4/8
|
| 468 |
+
2026-06-24 16:15:14,511 [INFO] Step 1390/3504 | Loss: 2.3596 | CE: 2.3596 | LR: 3.43e-04 | Epoch: 4/8
|
| 469 |
+
2026-06-24 16:15:23,120 [INFO] Step 1395/3504 | Loss: 2.3581 | CE: 2.3581 | LR: 3.42e-04 | Epoch: 4/8
|
| 470 |
+
2026-06-24 16:15:31,735 [INFO] Step 1400/3504 | Loss: 2.3758 | CE: 2.3758 | LR: 3.41e-04 | Epoch: 4/8
|
| 471 |
+
2026-06-24 16:15:40,149 [INFO] Step 1405/3504 | Loss: 2.4854 | CE: 2.4854 | LR: 3.40e-04 | Epoch: 4/8
|
| 472 |
+
2026-06-24 16:15:49,872 [INFO] Step 1410/3504 | Loss: 2.3632 | CE: 2.3632 | LR: 3.39e-04 | Epoch: 4/8
|
| 473 |
+
2026-06-24 16:15:58,161 [INFO] Step 1415/3504 | Loss: 2.4328 | CE: 2.4328 | LR: 3.38e-04 | Epoch: 4/8
|
| 474 |
+
2026-06-24 16:16:06,340 [INFO] Step 1420/3504 | Loss: 2.3565 | CE: 2.3565 | LR: 3.37e-04 | Epoch: 4/8
|
| 475 |
+
2026-06-24 16:16:13,700 [INFO] Step 1425/3504 | Loss: 2.4752 | CE: 2.4752 | LR: 3.36e-04 | Epoch: 4/8
|
| 476 |
+
2026-06-24 16:16:22,691 [INFO] Step 1430/3504 | Loss: 2.2841 | CE: 2.2841 | LR: 3.35e-04 | Epoch: 4/8
|
| 477 |
+
2026-06-24 16:16:30,281 [INFO] Step 1435/3504 | Loss: 2.3002 | CE: 2.3002 | LR: 3.34e-04 | Epoch: 4/8
|
| 478 |
+
2026-06-24 16:16:39,461 [INFO] Step 1440/3504 | Loss: 2.3401 | CE: 2.3401 | LR: 3.33e-04 | Epoch: 4/8
|
| 479 |
+
2026-06-24 16:16:47,133 [INFO] Step 1445/3504 | Loss: 2.3989 | CE: 2.3989 | LR: 3.32e-04 | Epoch: 4/8
|
| 480 |
+
2026-06-24 16:16:55,656 [INFO] Step 1450/3504 | Loss: 2.2997 | CE: 2.2997 | LR: 3.30e-04 | Epoch: 4/8
|
| 481 |
+
2026-06-24 16:17:02,826 [INFO] Step 1455/3504 | Loss: 2.4344 | CE: 2.4344 | LR: 3.29e-04 | Epoch: 4/8
|
| 482 |
+
2026-06-24 16:17:11,122 [INFO] Step 1460/3504 | Loss: 2.3521 | CE: 2.3521 | LR: 3.28e-04 | Epoch: 4/8
|
| 483 |
+
2026-06-24 16:17:20,652 [INFO] Step 1465/3504 | Loss: 2.2884 | CE: 2.2884 | LR: 3.27e-04 | Epoch: 4/8
|
| 484 |
+
2026-06-24 16:17:29,179 [INFO] Step 1470/3504 | Loss: 2.2405 | CE: 2.2405 | LR: 3.26e-04 | Epoch: 4/8
|
| 485 |
+
2026-06-24 16:17:37,171 [INFO] Step 1475/3504 | Loss: 2.3687 | CE: 2.3687 | LR: 3.25e-04 | Epoch: 4/8
|
| 486 |
+
2026-06-24 16:17:45,909 [INFO] Step 1480/3504 | Loss: 2.4791 | CE: 2.4791 | LR: 3.24e-04 | Epoch: 4/8
|
| 487 |
+
2026-06-24 16:17:54,235 [INFO] Step 1485/3504 | Loss: 2.2725 | CE: 2.2725 | LR: 3.23e-04 | Epoch: 4/8
|
| 488 |
+
2026-06-24 16:18:02,900 [INFO] Step 1490/3504 | Loss: 2.4637 | CE: 2.4637 | LR: 3.22e-04 | Epoch: 4/8
|
| 489 |
+
2026-06-24 16:18:11,177 [INFO] Step 1495/3504 | Loss: 2.2740 | CE: 2.2740 | LR: 3.21e-04 | Epoch: 4/8
|
| 490 |
+
2026-06-24 16:18:19,992 [INFO] Step 1500/3504 | Loss: 2.2725 | CE: 2.2725 | LR: 3.19e-04 | Epoch: 4/8
|
| 491 |
+
2026-06-24 16:18:30,648 [INFO] Step 1505/3504 | Loss: 2.3014 | CE: 2.3014 | LR: 3.18e-04 | Epoch: 4/8
|
| 492 |
+
2026-06-24 16:18:40,399 [INFO] Step 1510/3504 | Loss: 2.2779 | CE: 2.2779 | LR: 3.17e-04 | Epoch: 4/8
|
| 493 |
+
2026-06-24 16:18:49,520 [INFO] Step 1515/3504 | Loss: 2.3889 | CE: 2.3889 | LR: 3.16e-04 | Epoch: 4/8
|
| 494 |
+
2026-06-24 16:18:59,459 [INFO] Step 1520/3504 | Loss: 2.3256 | CE: 2.3256 | LR: 3.15e-04 | Epoch: 4/8
|
| 495 |
+
2026-06-24 16:19:07,897 [INFO] Step 1525/3504 | Loss: 2.2546 | CE: 2.2546 | LR: 3.14e-04 | Epoch: 4/8
|
| 496 |
+
2026-06-24 16:19:15,501 [INFO] Step 1530/3504 | Loss: 2.2042 | CE: 2.2042 | LR: 3.13e-04 | Epoch: 4/8
|
| 497 |
+
2026-06-24 16:19:23,062 [INFO] Step 1535/3504 | Loss: 2.4474 | CE: 2.4474 | LR: 3.12e-04 | Epoch: 4/8
|
| 498 |
+
2026-06-24 16:19:32,380 [INFO] Step 1540/3504 | Loss: 2.3269 | CE: 2.3269 | LR: 3.11e-04 | Epoch: 4/8
|
| 499 |
+
2026-06-24 16:19:41,460 [INFO] Step 1545/3504 | Loss: 2.3137 | CE: 2.3137 | LR: 3.09e-04 | Epoch: 4/8
|
| 500 |
+
2026-06-24 16:19:49,253 [INFO] Step 1550/3504 | Loss: 2.3398 | CE: 2.3398 | LR: 3.08e-04 | Epoch: 4/8
|
| 501 |
+
2026-06-24 16:19:57,021 [INFO] Step 1555/3504 | Loss: 2.4491 | CE: 2.4491 | LR: 3.07e-04 | Epoch: 4/8
|
| 502 |
+
2026-06-24 16:20:06,123 [INFO] Step 1560/3504 | Loss: 2.4111 | CE: 2.4111 | LR: 3.06e-04 | Epoch: 4/8
|
| 503 |
+
2026-06-24 16:20:14,914 [INFO] Step 1565/3504 | Loss: 2.4614 | CE: 2.4614 | LR: 3.05e-04 | Epoch: 4/8
|
| 504 |
+
2026-06-24 16:20:25,096 [INFO] Step 1570/3504 | Loss: 2.4284 | CE: 2.4284 | LR: 3.04e-04 | Epoch: 4/8
|
| 505 |
+
2026-06-24 16:20:35,220 [INFO] Step 1575/3504 | Loss: 2.3711 | CE: 2.3711 | LR: 3.03e-04 | Epoch: 4/8
|
| 506 |
+
2026-06-24 16:20:43,592 [INFO] Step 1580/3504 | Loss: 2.2723 | CE: 2.2723 | LR: 3.02e-04 | Epoch: 4/8
|
| 507 |
+
2026-06-24 16:20:53,205 [INFO] Step 1585/3504 | Loss: 2.2914 | CE: 2.2914 | LR: 3.00e-04 | Epoch: 4/8
|
| 508 |
+
2026-06-24 16:21:01,064 [INFO] Step 1590/3504 | Loss: 2.3045 | CE: 2.3045 | LR: 2.99e-04 | Epoch: 4/8
|
| 509 |
+
2026-06-24 16:21:08,828 [INFO] Step 1595/3504 | Loss: 2.3093 | CE: 2.3093 | LR: 2.98e-04 | Epoch: 4/8
|
| 510 |
+
2026-06-24 16:21:18,041 [INFO] Step 1600/3504 | Loss: 2.2612 | CE: 2.2612 | LR: 2.97e-04 | Epoch: 4/8
|
| 511 |
+
2026-06-24 16:21:25,660 [INFO] Step 1605/3504 | Loss: 2.2925 | CE: 2.2925 | LR: 2.96e-04 | Epoch: 4/8
|
| 512 |
+
2026-06-24 16:21:33,391 [INFO] Step 1610/3504 | Loss: 2.2530 | CE: 2.2530 | LR: 2.95e-04 | Epoch: 4/8
|
| 513 |
+
2026-06-24 16:21:43,315 [INFO] Step 1615/3504 | Loss: 2.3369 | CE: 2.3369 | LR: 2.94e-04 | Epoch: 4/8
|
| 514 |
+
2026-06-24 16:21:52,919 [INFO] Step 1620/3504 | Loss: 2.3243 | CE: 2.3243 | LR: 2.92e-04 | Epoch: 4/8
|
| 515 |
+
2026-06-24 16:22:01,087 [INFO] Step 1625/3504 | Loss: 2.3176 | CE: 2.3176 | LR: 2.91e-04 | Epoch: 4/8
|
| 516 |
+
2026-06-24 16:22:09,855 [INFO] Step 1630/3504 | Loss: 2.3623 | CE: 2.3623 | LR: 2.90e-04 | Epoch: 4/8
|
| 517 |
+
2026-06-24 16:22:18,467 [INFO] Step 1635/3504 | Loss: 2.3673 | CE: 2.3673 | LR: 2.89e-04 | Epoch: 4/8
|
| 518 |
+
2026-06-24 16:22:27,683 [INFO] Step 1640/3504 | Loss: 2.3330 | CE: 2.3330 | LR: 2.88e-04 | Epoch: 4/8
|
| 519 |
+
2026-06-24 16:22:37,039 [INFO] Step 1645/3504 | Loss: 2.3236 | CE: 2.3236 | LR: 2.87e-04 | Epoch: 4/8
|
| 520 |
+
2026-06-24 16:22:46,147 [INFO] Step 1650/3504 | Loss: 2.1773 | CE: 2.1773 | LR: 2.86e-04 | Epoch: 4/8
|
| 521 |
+
2026-06-24 16:22:53,530 [INFO] Step 1655/3504 | Loss: 2.2164 | CE: 2.2164 | LR: 2.84e-04 | Epoch: 4/8
|
| 522 |
+
2026-06-24 16:23:01,990 [INFO] Step 1660/3504 | Loss: 2.1839 | CE: 2.1839 | LR: 2.83e-04 | Epoch: 4/8
|
| 523 |
+
2026-06-24 16:23:11,166 [INFO] Step 1665/3504 | Loss: 2.1883 | CE: 2.1883 | LR: 2.82e-04 | Epoch: 4/8
|
| 524 |
+
2026-06-24 16:23:20,311 [INFO] Step 1670/3504 | Loss: 2.4545 | CE: 2.4545 | LR: 2.81e-04 | Epoch: 4/8
|
| 525 |
+
2026-06-24 16:23:28,876 [INFO] Step 1675/3504 | Loss: 2.2929 | CE: 2.2929 | LR: 2.80e-04 | Epoch: 4/8
|
| 526 |
+
2026-06-24 16:23:37,939 [INFO] Step 1680/3504 | Loss: 2.3493 | CE: 2.3493 | LR: 2.79e-04 | Epoch: 4/8
|
| 527 |
+
2026-06-24 16:23:47,665 [INFO] Step 1685/3504 | Loss: 2.3224 | CE: 2.3224 | LR: 2.78e-04 | Epoch: 4/8
|
| 528 |
+
2026-06-24 16:23:56,809 [INFO] Step 1690/3504 | Loss: 2.4095 | CE: 2.4095 | LR: 2.76e-04 | Epoch: 4/8
|
| 529 |
+
2026-06-24 16:24:05,480 [INFO] Step 1695/3504 | Loss: 2.2843 | CE: 2.2843 | LR: 2.75e-04 | Epoch: 4/8
|
| 530 |
+
2026-06-24 16:24:13,648 [INFO] Step 1700/3504 | Loss: 2.2560 | CE: 2.2560 | LR: 2.74e-04 | Epoch: 4/8
|
| 531 |
+
2026-06-24 16:24:22,460 [INFO] Step 1705/3504 | Loss: 2.3526 | CE: 2.3526 | LR: 2.73e-04 | Epoch: 4/8
|
| 532 |
+
2026-06-24 16:24:30,612 [INFO] Step 1710/3504 | Loss: 2.3053 | CE: 2.3053 | LR: 2.72e-04 | Epoch: 4/8
|
| 533 |
+
2026-06-24 16:24:40,246 [INFO] Step 1715/3504 | Loss: 2.1890 | CE: 2.1890 | LR: 2.71e-04 | Epoch: 4/8
|
| 534 |
+
2026-06-24 16:24:50,436 [INFO] Step 1720/3504 | Loss: 2.3749 | CE: 2.3749 | LR: 2.70e-04 | Epoch: 4/8
|
| 535 |
+
2026-06-24 16:24:59,021 [INFO] Step 1725/3504 | Loss: 2.3852 | CE: 2.3852 | LR: 2.68e-04 | Epoch: 4/8
|
| 536 |
+
2026-06-24 16:25:07,822 [INFO] Step 1730/3504 | Loss: 2.3107 | CE: 2.3107 | LR: 2.67e-04 | Epoch: 4/8
|
| 537 |
+
2026-06-24 16:25:17,437 [INFO] Step 1735/3504 | Loss: 2.4042 | CE: 2.4042 | LR: 2.66e-04 | Epoch: 4/8
|
| 538 |
+
2026-06-24 16:25:25,807 [INFO] Step 1740/3504 | Loss: 2.3426 | CE: 2.3426 | LR: 2.65e-04 | Epoch: 4/8
|
| 539 |
+
2026-06-24 16:25:34,413 [INFO] Step 1745/3504 | Loss: 2.2735 | CE: 2.2735 | LR: 2.64e-04 | Epoch: 4/8
|
| 540 |
+
2026-06-24 16:25:42,525 [INFO] Step 1750/3504 | Loss: 2.2943 | CE: 2.2943 | LR: 2.63e-04 | Epoch: 5/8
|
| 541 |
+
2026-06-24 16:26:44,489 [INFO] Step 1752 | Eval Loss: 2.6967
|
| 542 |
+
2026-06-24 16:26:47,436 [INFO] Saved trainable-only checkpoint to runs/phase1_warmup_qwen_wikilingua/best.pt (183 model tensors, stores_base_encoder_weights=False)
|
| 543 |
+
2026-06-24 16:26:48,925 [INFO] New best eval loss: 2.6967
|
| 544 |
+
2026-06-24 16:26:53,680 [INFO] Step 1755/3504 | Loss: 2.0162 | CE: 2.0162 | LR: 2.61e-04 | Epoch: 5/8
|
| 545 |
+
2026-06-24 16:27:00,781 [INFO] Step 1760/3504 | Loss: 1.9818 | CE: 1.9818 | LR: 2.60e-04 | Epoch: 5/8
|
| 546 |
+
2026-06-24 16:27:09,463 [INFO] Step 1765/3504 | Loss: 2.0042 | CE: 2.0042 | LR: 2.59e-04 | Epoch: 5/8
|
| 547 |
+
2026-06-24 16:27:18,972 [INFO] Step 1770/3504 | Loss: 1.9985 | CE: 1.9985 | LR: 2.58e-04 | Epoch: 5/8
|
| 548 |
+
2026-06-24 16:27:28,972 [INFO] Step 1775/3504 | Loss: 1.9047 | CE: 1.9047 | LR: 2.57e-04 | Epoch: 5/8
|
| 549 |
+
2026-06-24 16:27:37,459 [INFO] Step 1780/3504 | Loss: 1.9441 | CE: 1.9441 | LR: 2.56e-04 | Epoch: 5/8
|
| 550 |
+
2026-06-24 16:27:47,261 [INFO] Step 1785/3504 | Loss: 1.8457 | CE: 1.8457 | LR: 2.55e-04 | Epoch: 5/8
|
| 551 |
+
2026-06-24 16:27:57,438 [INFO] Step 1790/3504 | Loss: 2.0199 | CE: 2.0199 | LR: 2.53e-04 | Epoch: 5/8
|
| 552 |
+
2026-06-24 16:28:06,755 [INFO] Step 1795/3504 | Loss: 1.8412 | CE: 1.8412 | LR: 2.52e-04 | Epoch: 5/8
|
| 553 |
+
2026-06-24 16:28:15,674 [INFO] Step 1800/3504 | Loss: 1.9640 | CE: 1.9640 | LR: 2.51e-04 | Epoch: 5/8
|
| 554 |
+
2026-06-24 16:28:25,657 [INFO] Step 1805/3504 | Loss: 2.0019 | CE: 2.0019 | LR: 2.50e-04 | Epoch: 5/8
|
| 555 |
+
2026-06-24 16:28:35,069 [INFO] Step 1810/3504 | Loss: 1.9387 | CE: 1.9387 | LR: 2.49e-04 | Epoch: 5/8
|
| 556 |
+
2026-06-24 16:28:43,451 [INFO] Step 1815/3504 | Loss: 1.9462 | CE: 1.9462 | LR: 2.48e-04 | Epoch: 5/8
|
| 557 |
+
2026-06-24 16:28:53,514 [INFO] Step 1820/3504 | Loss: 1.9522 | CE: 1.9522 | LR: 2.46e-04 | Epoch: 5/8
|
| 558 |
+
2026-06-24 16:29:01,511 [INFO] Step 1825/3504 | Loss: 1.9048 | CE: 1.9048 | LR: 2.45e-04 | Epoch: 5/8
|
| 559 |
+
2026-06-24 16:29:10,064 [INFO] Step 1830/3504 | Loss: 1.9671 | CE: 1.9671 | LR: 2.44e-04 | Epoch: 5/8
|
| 560 |
+
2026-06-24 16:29:17,567 [INFO] Step 1835/3504 | Loss: 1.9374 | CE: 1.9374 | LR: 2.43e-04 | Epoch: 5/8
|
| 561 |
+
2026-06-24 16:29:27,074 [INFO] Step 1840/3504 | Loss: 2.0357 | CE: 2.0357 | LR: 2.42e-04 | Epoch: 5/8
|
| 562 |
+
2026-06-24 16:29:35,585 [INFO] Step 1845/3504 | Loss: 2.0110 | CE: 2.0110 | LR: 2.41e-04 | Epoch: 5/8
|
| 563 |
+
2026-06-24 16:29:44,451 [INFO] Step 1850/3504 | Loss: 1.8878 | CE: 1.8878 | LR: 2.39e-04 | Epoch: 5/8
|
| 564 |
+
2026-06-24 16:29:53,411 [INFO] Step 1855/3504 | Loss: 1.9137 | CE: 1.9137 | LR: 2.38e-04 | Epoch: 5/8
|
| 565 |
+
2026-06-24 16:30:02,685 [INFO] Step 1860/3504 | Loss: 1.9007 | CE: 1.9007 | LR: 2.37e-04 | Epoch: 5/8
|
| 566 |
+
2026-06-24 16:30:10,701 [INFO] Step 1865/3504 | Loss: 1.9384 | CE: 1.9384 | LR: 2.36e-04 | Epoch: 5/8
|
| 567 |
+
2026-06-24 16:30:19,190 [INFO] Step 1870/3504 | Loss: 2.0280 | CE: 2.0280 | LR: 2.35e-04 | Epoch: 5/8
|
| 568 |
+
2026-06-24 16:30:26,649 [INFO] Step 1875/3504 | Loss: 1.9541 | CE: 1.9541 | LR: 2.34e-04 | Epoch: 5/8
|
| 569 |
+
2026-06-24 16:30:35,508 [INFO] Step 1880/3504 | Loss: 1.9725 | CE: 1.9725 | LR: 2.33e-04 | Epoch: 5/8
|
| 570 |
+
2026-06-24 16:30:43,057 [INFO] Step 1885/3504 | Loss: 1.9080 | CE: 1.9080 | LR: 2.31e-04 | Epoch: 5/8
|
| 571 |
+
2026-06-24 16:30:51,540 [INFO] Step 1890/3504 | Loss: 2.0389 | CE: 2.0389 | LR: 2.30e-04 | Epoch: 5/8
|
| 572 |
+
2026-06-24 16:31:00,415 [INFO] Step 1895/3504 | Loss: 1.9080 | CE: 1.9080 | LR: 2.29e-04 | Epoch: 5/8
|
| 573 |
+
2026-06-24 16:31:10,987 [INFO] Step 1900/3504 | Loss: 1.9454 | CE: 1.9454 | LR: 2.28e-04 | Epoch: 5/8
|
| 574 |
+
2026-06-24 16:31:20,507 [INFO] Step 1905/3504 | Loss: 1.9328 | CE: 1.9328 | LR: 2.27e-04 | Epoch: 5/8
|
| 575 |
+
2026-06-24 16:31:28,639 [INFO] Step 1910/3504 | Loss: 2.0451 | CE: 2.0451 | LR: 2.26e-04 | Epoch: 5/8
|
| 576 |
+
2026-06-24 16:31:38,743 [INFO] Step 1915/3504 | Loss: 1.9842 | CE: 1.9842 | LR: 2.25e-04 | Epoch: 5/8
|
| 577 |
+
2026-06-24 16:31:46,949 [INFO] Step 1920/3504 | Loss: 2.0288 | CE: 2.0288 | LR: 2.23e-04 | Epoch: 5/8
|
| 578 |
+
2026-06-24 16:31:56,337 [INFO] Step 1925/3504 | Loss: 2.0099 | CE: 2.0099 | LR: 2.22e-04 | Epoch: 5/8
|
| 579 |
+
2026-06-24 16:32:04,718 [INFO] Step 1930/3504 | Loss: 1.9584 | CE: 1.9584 | LR: 2.21e-04 | Epoch: 5/8
|
| 580 |
+
2026-06-24 16:32:12,052 [INFO] Step 1935/3504 | Loss: 1.9246 | CE: 1.9246 | LR: 2.20e-04 | Epoch: 5/8
|
| 581 |
+
2026-06-24 16:32:20,360 [INFO] Step 1940/3504 | Loss: 1.9589 | CE: 1.9589 | LR: 2.19e-04 | Epoch: 5/8
|
| 582 |
+
2026-06-24 16:32:27,972 [INFO] Step 1945/3504 | Loss: 1.9887 | CE: 1.9887 | LR: 2.18e-04 | Epoch: 5/8
|
| 583 |
+
2026-06-24 16:32:36,013 [INFO] Step 1950/3504 | Loss: 1.8941 | CE: 1.8941 | LR: 2.16e-04 | Epoch: 5/8
|
| 584 |
+
2026-06-24 16:32:44,607 [INFO] Step 1955/3504 | Loss: 1.9999 | CE: 1.9999 | LR: 2.15e-04 | Epoch: 5/8
|
| 585 |
+
2026-06-24 16:32:54,582 [INFO] Step 1960/3504 | Loss: 2.0079 | CE: 2.0079 | LR: 2.14e-04 | Epoch: 5/8
|
| 586 |
+
2026-06-24 16:33:03,234 [INFO] Step 1965/3504 | Loss: 1.9926 | CE: 1.9926 | LR: 2.13e-04 | Epoch: 5/8
|
| 587 |
+
2026-06-24 16:33:13,077 [INFO] Step 1970/3504 | Loss: 2.0141 | CE: 2.0141 | LR: 2.12e-04 | Epoch: 5/8
|
| 588 |
+
2026-06-24 16:33:22,444 [INFO] Step 1975/3504 | Loss: 1.9745 | CE: 1.9745 | LR: 2.11e-04 | Epoch: 5/8
|
| 589 |
+
2026-06-24 16:33:30,899 [INFO] Step 1980/3504 | Loss: 2.0030 | CE: 2.0030 | LR: 2.10e-04 | Epoch: 5/8
|
| 590 |
+
2026-06-24 16:33:38,308 [INFO] Step 1985/3504 | Loss: 1.8668 | CE: 1.8668 | LR: 2.08e-04 | Epoch: 5/8
|
| 591 |
+
2026-06-24 16:33:46,846 [INFO] Step 1990/3504 | Loss: 1.9300 | CE: 1.9300 | LR: 2.07e-04 | Epoch: 5/8
|
| 592 |
+
2026-06-24 16:33:56,545 [INFO] Step 1995/3504 | Loss: 2.0195 | CE: 2.0195 | LR: 2.06e-04 | Epoch: 5/8
|
| 593 |
+
2026-06-24 16:34:05,558 [INFO] Step 2000/3504 | Loss: 1.7939 | CE: 1.7939 | LR: 2.05e-04 | Epoch: 5/8
|
| 594 |
+
2026-06-24 16:34:14,641 [INFO] Step 2005/3504 | Loss: 1.9375 | CE: 1.9375 | LR: 2.04e-04 | Epoch: 5/8
|
| 595 |
+
2026-06-24 16:34:25,164 [INFO] Step 2010/3504 | Loss: 2.1099 | CE: 2.1099 | LR: 2.03e-04 | Epoch: 5/8
|
| 596 |
+
2026-06-24 16:34:33,330 [INFO] Step 2015/3504 | Loss: 2.0513 | CE: 2.0513 | LR: 2.02e-04 | Epoch: 5/8
|
| 597 |
+
2026-06-24 16:34:41,508 [INFO] Step 2020/3504 | Loss: 1.9949 | CE: 1.9949 | LR: 2.01e-04 | Epoch: 5/8
|
| 598 |
+
2026-06-24 16:34:50,157 [INFO] Step 2025/3504 | Loss: 1.9823 | CE: 1.9823 | LR: 1.99e-04 | Epoch: 5/8
|
| 599 |
+
2026-06-24 16:34:58,751 [INFO] Step 2030/3504 | Loss: 1.9489 | CE: 1.9489 | LR: 1.98e-04 | Epoch: 5/8
|
| 600 |
+
2026-06-24 16:35:06,312 [INFO] Step 2035/3504 | Loss: 1.8564 | CE: 1.8564 | LR: 1.97e-04 | Epoch: 5/8
|
| 601 |
+
2026-06-24 16:35:15,982 [INFO] Step 2040/3504 | Loss: 1.9117 | CE: 1.9117 | LR: 1.96e-04 | Epoch: 5/8
|
| 602 |
+
2026-06-24 16:35:24,753 [INFO] Step 2045/3504 | Loss: 1.9064 | CE: 1.9064 | LR: 1.95e-04 | Epoch: 5/8
|
| 603 |
+
2026-06-24 16:35:33,275 [INFO] Step 2050/3504 | Loss: 2.0113 | CE: 2.0113 | LR: 1.94e-04 | Epoch: 5/8
|
| 604 |
+
2026-06-24 16:35:42,106 [INFO] Step 2055/3504 | Loss: 1.9755 | CE: 1.9755 | LR: 1.93e-04 | Epoch: 5/8
|
| 605 |
+
2026-06-24 16:35:50,439 [INFO] Step 2060/3504 | Loss: 1.9973 | CE: 1.9973 | LR: 1.92e-04 | Epoch: 5/8
|
| 606 |
+
2026-06-24 16:35:58,769 [INFO] Step 2065/3504 | Loss: 2.0658 | CE: 2.0658 | LR: 1.90e-04 | Epoch: 5/8
|
| 607 |
+
2026-06-24 16:36:08,196 [INFO] Step 2070/3504 | Loss: 1.8421 | CE: 1.8421 | LR: 1.89e-04 | Epoch: 5/8
|
| 608 |
+
2026-06-24 16:36:16,709 [INFO] Step 2075/3504 | Loss: 1.9169 | CE: 1.9169 | LR: 1.88e-04 | Epoch: 5/8
|
| 609 |
+
2026-06-24 16:36:24,696 [INFO] Step 2080/3504 | Loss: 1.9423 | CE: 1.9423 | LR: 1.87e-04 | Epoch: 5/8
|
| 610 |
+
2026-06-24 16:36:33,502 [INFO] Step 2085/3504 | Loss: 1.8850 | CE: 1.8850 | LR: 1.86e-04 | Epoch: 5/8
|
| 611 |
+
2026-06-24 16:36:42,042 [INFO] Step 2090/3504 | Loss: 2.0096 | CE: 2.0096 | LR: 1.85e-04 | Epoch: 5/8
|
| 612 |
+
2026-06-24 16:36:50,720 [INFO] Step 2095/3504 | Loss: 1.8298 | CE: 1.8298 | LR: 1.84e-04 | Epoch: 5/8
|
| 613 |
+
2026-06-24 16:36:58,796 [INFO] Step 2100/3504 | Loss: 1.9371 | CE: 1.9371 | LR: 1.83e-04 | Epoch: 5/8
|
| 614 |
+
2026-06-24 16:37:07,175 [INFO] Step 2105/3504 | Loss: 1.8742 | CE: 1.8742 | LR: 1.81e-04 | Epoch: 5/8
|
| 615 |
+
2026-06-24 16:37:15,335 [INFO] Step 2110/3504 | Loss: 2.0428 | CE: 2.0428 | LR: 1.80e-04 | Epoch: 5/8
|
| 616 |
+
2026-06-24 16:37:26,044 [INFO] Step 2115/3504 | Loss: 1.9163 | CE: 1.9163 | LR: 1.79e-04 | Epoch: 5/8
|
| 617 |
+
2026-06-24 16:37:34,359 [INFO] Step 2120/3504 | Loss: 1.9930 | CE: 1.9930 | LR: 1.78e-04 | Epoch: 5/8
|
| 618 |
+
2026-06-24 16:37:42,475 [INFO] Step 2125/3504 | Loss: 1.8808 | CE: 1.8808 | LR: 1.77e-04 | Epoch: 5/8
|
| 619 |
+
2026-06-24 16:37:50,881 [INFO] Step 2130/3504 | Loss: 1.8596 | CE: 1.8596 | LR: 1.76e-04 | Epoch: 5/8
|
| 620 |
+
2026-06-24 16:37:59,401 [INFO] Step 2135/3504 | Loss: 1.8416 | CE: 1.8416 | LR: 1.75e-04 | Epoch: 5/8
|
| 621 |
+
2026-06-24 16:38:07,369 [INFO] Step 2140/3504 | Loss: 1.8295 | CE: 1.8295 | LR: 1.74e-04 | Epoch: 5/8
|
| 622 |
+
2026-06-24 16:38:16,580 [INFO] Step 2145/3504 | Loss: 1.8588 | CE: 1.8588 | LR: 1.73e-04 | Epoch: 5/8
|
| 623 |
+
2026-06-24 16:38:28,168 [INFO] Step 2150/3504 | Loss: 1.9232 | CE: 1.9232 | LR: 1.72e-04 | Epoch: 5/8
|
| 624 |
+
2026-06-24 16:38:37,263 [INFO] Step 2155/3504 | Loss: 1.9600 | CE: 1.9600 | LR: 1.70e-04 | Epoch: 5/8
|
| 625 |
+
2026-06-24 16:38:45,221 [INFO] Step 2160/3504 | Loss: 1.8925 | CE: 1.8925 | LR: 1.69e-04 | Epoch: 5/8
|
| 626 |
+
2026-06-24 16:38:54,726 [INFO] Step 2165/3504 | Loss: 1.9112 | CE: 1.9112 | LR: 1.68e-04 | Epoch: 5/8
|
| 627 |
+
2026-06-24 16:39:04,515 [INFO] Step 2170/3504 | Loss: 1.8545 | CE: 1.8545 | LR: 1.67e-04 | Epoch: 5/8
|
| 628 |
+
2026-06-24 16:39:12,793 [INFO] Step 2175/3504 | Loss: 2.0029 | CE: 2.0029 | LR: 1.66e-04 | Epoch: 5/8
|
| 629 |
+
2026-06-24 16:39:21,580 [INFO] Step 2180/3504 | Loss: 1.9173 | CE: 1.9173 | LR: 1.65e-04 | Epoch: 5/8
|
| 630 |
+
2026-06-24 16:39:30,017 [INFO] Step 2185/3504 | Loss: 1.9543 | CE: 1.9543 | LR: 1.64e-04 | Epoch: 5/8
|
| 631 |
+
2026-06-24 16:39:39,038 [INFO] Step 2190/3504 | Loss: 1.7082 | CE: 1.7082 | LR: 1.63e-04 | Epoch: 6/8
|
| 632 |
+
2026-06-24 16:40:37,150 [INFO] Step 2190 | Eval Loss: 2.6334
|
| 633 |
+
2026-06-24 16:40:39,790 [INFO] Saved trainable-only checkpoint to runs/phase1_warmup_qwen_wikilingua/best.pt (183 model tensors, stores_base_encoder_weights=False)
|
| 634 |
+
2026-06-24 16:40:41,219 [INFO] New best eval loss: 2.6334
|
| 635 |
+
2026-06-24 16:40:48,280 [INFO] Step 2195/3504 | Loss: 1.6227 | CE: 1.6227 | LR: 1.62e-04 | Epoch: 6/8
|
| 636 |
+
2026-06-24 16:40:57,559 [INFO] Step 2200/3504 | Loss: 1.5654 | CE: 1.5654 | LR: 1.61e-04 | Epoch: 6/8
|
| 637 |
+
2026-06-24 16:41:07,173 [INFO] Step 2205/3504 | Loss: 1.5361 | CE: 1.5361 | LR: 1.60e-04 | Epoch: 6/8
|
| 638 |
+
2026-06-24 16:41:17,007 [INFO] Step 2210/3504 | Loss: 1.5356 | CE: 1.5356 | LR: 1.58e-04 | Epoch: 6/8
|
| 639 |
+
2026-06-24 16:41:25,385 [INFO] Step 2215/3504 | Loss: 1.6220 | CE: 1.6220 | LR: 1.57e-04 | Epoch: 6/8
|
| 640 |
+
2026-06-24 16:41:35,525 [INFO] Step 2220/3504 | Loss: 1.5380 | CE: 1.5380 | LR: 1.56e-04 | Epoch: 6/8
|
| 641 |
+
2026-06-24 16:41:43,748 [INFO] Step 2225/3504 | Loss: 1.5593 | CE: 1.5593 | LR: 1.55e-04 | Epoch: 6/8
|
| 642 |
+
2026-06-24 16:41:50,892 [INFO] Step 2230/3504 | Loss: 1.5095 | CE: 1.5095 | LR: 1.54e-04 | Epoch: 6/8
|
| 643 |
+
2026-06-24 16:42:00,270 [INFO] Step 2235/3504 | Loss: 1.5636 | CE: 1.5636 | LR: 1.53e-04 | Epoch: 6/8
|
| 644 |
+
2026-06-24 16:42:08,323 [INFO] Step 2240/3504 | Loss: 1.5988 | CE: 1.5988 | LR: 1.52e-04 | Epoch: 6/8
|
| 645 |
+
2026-06-24 16:42:16,372 [INFO] Step 2245/3504 | Loss: 1.5341 | CE: 1.5341 | LR: 1.51e-04 | Epoch: 6/8
|
| 646 |
+
2026-06-24 16:42:25,677 [INFO] Step 2250/3504 | Loss: 1.5863 | CE: 1.5863 | LR: 1.50e-04 | Epoch: 6/8
|
| 647 |
+
2026-06-24 16:42:34,494 [INFO] Step 2255/3504 | Loss: 1.5113 | CE: 1.5113 | LR: 1.49e-04 | Epoch: 6/8
|
| 648 |
+
2026-06-24 16:42:43,543 [INFO] Step 2260/3504 | Loss: 1.5472 | CE: 1.5472 | LR: 1.48e-04 | Epoch: 6/8
|
| 649 |
+
2026-06-24 16:42:52,521 [INFO] Step 2265/3504 | Loss: 1.6036 | CE: 1.6036 | LR: 1.47e-04 | Epoch: 6/8
|
| 650 |
+
2026-06-24 16:43:01,439 [INFO] Step 2270/3504 | Loss: 1.5479 | CE: 1.5479 | LR: 1.46e-04 | Epoch: 6/8
|
| 651 |
+
2026-06-24 16:43:10,832 [INFO] Step 2275/3504 | Loss: 1.5883 | CE: 1.5883 | LR: 1.45e-04 | Epoch: 6/8
|
| 652 |
+
2026-06-24 16:43:19,798 [INFO] Step 2280/3504 | Loss: 1.4744 | CE: 1.4744 | LR: 1.44e-04 | Epoch: 6/8
|
| 653 |
+
2026-06-24 16:43:27,092 [INFO] Step 2285/3504 | Loss: 1.5694 | CE: 1.5694 | LR: 1.43e-04 | Epoch: 6/8
|
| 654 |
+
2026-06-24 16:43:35,172 [INFO] Step 2290/3504 | Loss: 1.4821 | CE: 1.4821 | LR: 1.42e-04 | Epoch: 6/8
|
| 655 |
+
2026-06-24 16:43:43,491 [INFO] Step 2295/3504 | Loss: 1.5149 | CE: 1.5149 | LR: 1.41e-04 | Epoch: 6/8
|
| 656 |
+
2026-06-24 16:43:52,582 [INFO] Step 2300/3504 | Loss: 1.5584 | CE: 1.5584 | LR: 1.39e-04 | Epoch: 6/8
|
| 657 |
+
2026-06-24 16:44:01,159 [INFO] Step 2305/3504 | Loss: 1.6116 | CE: 1.6116 | LR: 1.38e-04 | Epoch: 6/8
|
| 658 |
+
2026-06-24 16:44:08,096 [INFO] Step 2310/3504 | Loss: 1.5768 | CE: 1.5768 | LR: 1.37e-04 | Epoch: 6/8
|
| 659 |
+
2026-06-24 16:44:17,799 [INFO] Step 2315/3504 | Loss: 1.6267 | CE: 1.6267 | LR: 1.36e-04 | Epoch: 6/8
|
| 660 |
+
2026-06-24 16:44:26,702 [INFO] Step 2320/3504 | Loss: 1.5567 | CE: 1.5567 | LR: 1.35e-04 | Epoch: 6/8
|
| 661 |
+
2026-06-24 16:44:36,873 [INFO] Step 2325/3504 | Loss: 1.5794 | CE: 1.5794 | LR: 1.34e-04 | Epoch: 6/8
|
| 662 |
+
2026-06-24 16:44:44,428 [INFO] Step 2330/3504 | Loss: 1.5889 | CE: 1.5889 | LR: 1.33e-04 | Epoch: 6/8
|
| 663 |
+
2026-06-24 16:44:53,483 [INFO] Step 2335/3504 | Loss: 1.5919 | CE: 1.5919 | LR: 1.32e-04 | Epoch: 6/8
|
| 664 |
+
2026-06-24 16:45:02,801 [INFO] Step 2340/3504 | Loss: 1.6062 | CE: 1.6062 | LR: 1.31e-04 | Epoch: 6/8
|
| 665 |
+
2026-06-24 16:45:12,582 [INFO] Step 2345/3504 | Loss: 1.6136 | CE: 1.6136 | LR: 1.30e-04 | Epoch: 6/8
|
| 666 |
+
2026-06-24 16:45:21,688 [INFO] Step 2350/3504 | Loss: 1.5732 | CE: 1.5732 | LR: 1.29e-04 | Epoch: 6/8
|
| 667 |
+
2026-06-24 16:45:30,957 [INFO] Step 2355/3504 | Loss: 1.4686 | CE: 1.4686 | LR: 1.28e-04 | Epoch: 6/8
|
| 668 |
+
2026-06-24 16:45:40,172 [INFO] Step 2360/3504 | Loss: 1.4952 | CE: 1.4952 | LR: 1.27e-04 | Epoch: 6/8
|
| 669 |
+
2026-06-24 16:45:48,457 [INFO] Step 2365/3504 | Loss: 1.5508 | CE: 1.5508 | LR: 1.26e-04 | Epoch: 6/8
|
| 670 |
+
2026-06-24 16:45:56,447 [INFO] Step 2370/3504 | Loss: 1.5160 | CE: 1.5160 | LR: 1.25e-04 | Epoch: 6/8
|
| 671 |
+
2026-06-24 16:46:06,477 [INFO] Step 2375/3504 | Loss: 1.4839 | CE: 1.4839 | LR: 1.24e-04 | Epoch: 6/8
|
| 672 |
+
2026-06-24 16:46:14,316 [INFO] Step 2380/3504 | Loss: 1.5473 | CE: 1.5473 | LR: 1.23e-04 | Epoch: 6/8
|
| 673 |
+
2026-06-24 16:46:23,437 [INFO] Step 2385/3504 | Loss: 1.5522 | CE: 1.5522 | LR: 1.22e-04 | Epoch: 6/8
|
| 674 |
+
2026-06-24 16:46:31,120 [INFO] Step 2390/3504 | Loss: 1.5829 | CE: 1.5829 | LR: 1.21e-04 | Epoch: 6/8
|
| 675 |
+
2026-06-24 16:46:39,042 [INFO] Step 2395/3504 | Loss: 1.5617 | CE: 1.5617 | LR: 1.20e-04 | Epoch: 6/8
|
| 676 |
+
2026-06-24 16:46:47,653 [INFO] Step 2400/3504 | Loss: 1.5624 | CE: 1.5624 | LR: 1.19e-04 | Epoch: 6/8
|
| 677 |
+
2026-06-24 16:46:59,463 [INFO] Step 2405/3504 | Loss: 1.6444 | CE: 1.6444 | LR: 1.18e-04 | Epoch: 6/8
|
| 678 |
+
2026-06-24 16:47:07,854 [INFO] Step 2410/3504 | Loss: 1.5968 | CE: 1.5968 | LR: 1.17e-04 | Epoch: 6/8
|
| 679 |
+
2026-06-24 16:47:16,962 [INFO] Step 2415/3504 | Loss: 1.5017 | CE: 1.5017 | LR: 1.16e-04 | Epoch: 6/8
|
| 680 |
+
2026-06-24 16:47:26,402 [INFO] Step 2420/3504 | Loss: 1.5379 | CE: 1.5379 | LR: 1.15e-04 | Epoch: 6/8
|
| 681 |
+
2026-06-24 16:47:34,461 [INFO] Step 2425/3504 | Loss: 1.5296 | CE: 1.5296 | LR: 1.14e-04 | Epoch: 6/8
|
| 682 |
+
2026-06-24 16:47:42,663 [INFO] Step 2430/3504 | Loss: 1.5367 | CE: 1.5367 | LR: 1.13e-04 | Epoch: 6/8
|
| 683 |
+
2026-06-24 16:47:50,986 [INFO] Step 2435/3504 | Loss: 1.5207 | CE: 1.5207 | LR: 1.12e-04 | Epoch: 6/8
|
| 684 |
+
2026-06-24 16:48:00,171 [INFO] Step 2440/3504 | Loss: 1.5464 | CE: 1.5464 | LR: 1.11e-04 | Epoch: 6/8
|
| 685 |
+
2026-06-24 16:48:09,583 [INFO] Step 2445/3504 | Loss: 1.5678 | CE: 1.5678 | LR: 1.10e-04 | Epoch: 6/8
|
| 686 |
+
2026-06-24 16:48:18,290 [INFO] Step 2450/3504 | Loss: 1.6943 | CE: 1.6943 | LR: 1.10e-04 | Epoch: 6/8
|
| 687 |
+
2026-06-24 16:48:27,747 [INFO] Step 2455/3504 | Loss: 1.5399 | CE: 1.5399 | LR: 1.09e-04 | Epoch: 6/8
|
| 688 |
+
2026-06-24 16:48:36,759 [INFO] Step 2460/3504 | Loss: 1.6787 | CE: 1.6787 | LR: 1.08e-04 | Epoch: 6/8
|
| 689 |
+
2026-06-24 16:48:44,372 [INFO] Step 2465/3504 | Loss: 1.5448 | CE: 1.5448 | LR: 1.07e-04 | Epoch: 6/8
|
| 690 |
+
2026-06-24 16:48:54,275 [INFO] Step 2470/3504 | Loss: 1.4917 | CE: 1.4917 | LR: 1.06e-04 | Epoch: 6/8
|
| 691 |
+
2026-06-24 16:49:04,079 [INFO] Step 2475/3504 | Loss: 1.5718 | CE: 1.5718 | LR: 1.05e-04 | Epoch: 6/8
|
| 692 |
+
2026-06-24 16:49:12,935 [INFO] Step 2480/3504 | Loss: 1.4931 | CE: 1.4931 | LR: 1.04e-04 | Epoch: 6/8
|
| 693 |
+
2026-06-24 16:49:22,550 [INFO] Step 2485/3504 | Loss: 1.4946 | CE: 1.4946 | LR: 1.03e-04 | Epoch: 6/8
|
| 694 |
+
2026-06-24 16:49:30,175 [INFO] Step 2490/3504 | Loss: 1.4375 | CE: 1.4375 | LR: 1.02e-04 | Epoch: 6/8
|
| 695 |
+
2026-06-24 16:49:37,460 [INFO] Step 2495/3504 | Loss: 1.5828 | CE: 1.5828 | LR: 1.01e-04 | Epoch: 6/8
|
| 696 |
+
2026-06-24 16:49:45,918 [INFO] Step 2500/3504 | Loss: 1.6845 | CE: 1.6845 | LR: 1.00e-04 | Epoch: 6/8
|
| 697 |
+
2026-06-24 16:49:54,612 [INFO] Step 2505/3504 | Loss: 1.5668 | CE: 1.5668 | LR: 9.92e-05 | Epoch: 6/8
|
| 698 |
+
2026-06-24 16:50:03,287 [INFO] Step 2510/3504 | Loss: 1.5377 | CE: 1.5377 | LR: 9.83e-05 | Epoch: 6/8
|
| 699 |
+
2026-06-24 16:50:13,317 [INFO] Step 2515/3504 | Loss: 1.5874 | CE: 1.5874 | LR: 9.74e-05 | Epoch: 6/8
|
| 700 |
+
2026-06-24 16:50:21,976 [INFO] Step 2520/3504 | Loss: 1.5764 | CE: 1.5764 | LR: 9.65e-05 | Epoch: 6/8
|
| 701 |
+
2026-06-24 16:50:30,267 [INFO] Step 2525/3504 | Loss: 1.5471 | CE: 1.5471 | LR: 9.56e-05 | Epoch: 6/8
|
| 702 |
+
2026-06-24 16:50:40,779 [INFO] Step 2530/3504 | Loss: 1.5912 | CE: 1.5912 | LR: 9.46e-05 | Epoch: 6/8
|
| 703 |
+
2026-06-24 16:50:49,079 [INFO] Step 2535/3504 | Loss: 1.4822 | CE: 1.4822 | LR: 9.37e-05 | Epoch: 6/8
|
| 704 |
+
2026-06-24 16:50:57,878 [INFO] Step 2540/3504 | Loss: 1.5815 | CE: 1.5815 | LR: 9.28e-05 | Epoch: 6/8
|
| 705 |
+
2026-06-24 16:51:06,434 [INFO] Step 2545/3504 | Loss: 1.4661 | CE: 1.4661 | LR: 9.19e-05 | Epoch: 6/8
|
| 706 |
+
2026-06-24 16:51:14,183 [INFO] Step 2550/3504 | Loss: 1.5225 | CE: 1.5225 | LR: 9.11e-05 | Epoch: 6/8
|
| 707 |
+
2026-06-24 16:51:23,156 [INFO] Step 2555/3504 | Loss: 1.5090 | CE: 1.5090 | LR: 9.02e-05 | Epoch: 6/8
|
| 708 |
+
2026-06-24 16:51:30,766 [INFO] Step 2560/3504 | Loss: 1.5782 | CE: 1.5782 | LR: 8.93e-05 | Epoch: 6/8
|
| 709 |
+
2026-06-24 16:51:38,978 [INFO] Step 2565/3504 | Loss: 1.5495 | CE: 1.5495 | LR: 8.84e-05 | Epoch: 6/8
|
| 710 |
+
2026-06-24 16:51:47,348 [INFO] Step 2570/3504 | Loss: 1.5260 | CE: 1.5260 | LR: 8.75e-05 | Epoch: 6/8
|
| 711 |
+
2026-06-24 16:51:56,116 [INFO] Step 2575/3504 | Loss: 1.5838 | CE: 1.5838 | LR: 8.66e-05 | Epoch: 6/8
|
| 712 |
+
2026-06-24 16:52:05,390 [INFO] Step 2580/3504 | Loss: 1.4240 | CE: 1.4240 | LR: 8.58e-05 | Epoch: 6/8
|
| 713 |
+
2026-06-24 16:52:13,518 [INFO] Step 2585/3504 | Loss: 1.5162 | CE: 1.5162 | LR: 8.49e-05 | Epoch: 6/8
|
| 714 |
+
2026-06-24 16:52:21,387 [INFO] Step 2590/3504 | Loss: 1.5277 | CE: 1.5277 | LR: 8.40e-05 | Epoch: 6/8
|
| 715 |
+
2026-06-24 16:52:30,486 [INFO] Step 2595/3504 | Loss: 1.5363 | CE: 1.5363 | LR: 8.32e-05 | Epoch: 6/8
|
| 716 |
+
2026-06-24 16:52:38,886 [INFO] Step 2600/3504 | Loss: 1.5102 | CE: 1.5102 | LR: 8.23e-05 | Epoch: 6/8
|
| 717 |
+
2026-06-24 16:52:48,809 [INFO] Step 2605/3504 | Loss: 1.5586 | CE: 1.5586 | LR: 8.15e-05 | Epoch: 6/8
|
| 718 |
+
2026-06-24 16:52:56,421 [INFO] Step 2610/3504 | Loss: 1.5937 | CE: 1.5937 | LR: 8.06e-05 | Epoch: 6/8
|
| 719 |
+
2026-06-24 16:53:04,673 [INFO] Step 2615/3504 | Loss: 1.5580 | CE: 1.5580 | LR: 7.98e-05 | Epoch: 6/8
|
| 720 |
+
2026-06-24 16:53:13,549 [INFO] Step 2620/3504 | Loss: 1.5618 | CE: 1.5618 | LR: 7.89e-05 | Epoch: 6/8
|
| 721 |
+
2026-06-24 16:53:23,710 [INFO] Step 2625/3504 | Loss: 1.5362 | CE: 1.5362 | LR: 7.81e-05 | Epoch: 7/8
|
| 722 |
+
2026-06-24 16:54:27,375 [INFO] Step 2628 | Eval Loss: 2.6351
|
| 723 |
+
2026-06-24 16:54:31,145 [INFO] Step 2630/3504 | Loss: 1.2871 | CE: 1.2871 | LR: 7.72e-05 | Epoch: 7/8
|
| 724 |
+
2026-06-24 16:54:40,755 [INFO] Step 2635/3504 | Loss: 1.2199 | CE: 1.2199 | LR: 7.64e-05 | Epoch: 7/8
|
| 725 |
+
2026-06-24 16:54:49,303 [INFO] Step 2640/3504 | Loss: 1.2574 | CE: 1.2574 | LR: 7.56e-05 | Epoch: 7/8
|
| 726 |
+
2026-06-24 16:54:59,005 [INFO] Step 2645/3504 | Loss: 1.2248 | CE: 1.2248 | LR: 7.47e-05 | Epoch: 7/8
|
| 727 |
+
2026-06-24 16:55:07,810 [INFO] Step 2650/3504 | Loss: 1.2801 | CE: 1.2801 | LR: 7.39e-05 | Epoch: 7/8
|
| 728 |
+
2026-06-24 16:55:16,255 [INFO] Step 2655/3504 | Loss: 1.2826 | CE: 1.2826 | LR: 7.31e-05 | Epoch: 7/8
|
| 729 |
+
2026-06-24 16:55:26,599 [INFO] Step 2660/3504 | Loss: 1.2413 | CE: 1.2413 | LR: 7.23e-05 | Epoch: 7/8
|
| 730 |
+
2026-06-24 16:55:34,475 [INFO] Step 2665/3504 | Loss: 1.1808 | CE: 1.1808 | LR: 7.15e-05 | Epoch: 7/8
|
| 731 |
+
2026-06-24 16:55:43,461 [INFO] Step 2670/3504 | Loss: 1.2845 | CE: 1.2845 | LR: 7.07e-05 | Epoch: 7/8
|
| 732 |
+
2026-06-24 16:55:51,367 [INFO] Step 2675/3504 | Loss: 1.2569 | CE: 1.2569 | LR: 6.99e-05 | Epoch: 7/8
|
| 733 |
+
2026-06-24 16:55:59,338 [INFO] Step 2680/3504 | Loss: 1.2504 | CE: 1.2504 | LR: 6.91e-05 | Epoch: 7/8
|
| 734 |
+
2026-06-24 16:56:07,737 [INFO] Step 2685/3504 | Loss: 1.3041 | CE: 1.3041 | LR: 6.83e-05 | Epoch: 7/8
|
| 735 |
+
2026-06-24 16:56:16,295 [INFO] Step 2690/3504 | Loss: 1.2622 | CE: 1.2622 | LR: 6.75e-05 | Epoch: 7/8
|
| 736 |
+
2026-06-24 16:56:24,548 [INFO] Step 2695/3504 | Loss: 1.3073 | CE: 1.3073 | LR: 6.67e-05 | Epoch: 7/8
|
| 737 |
+
2026-06-24 16:56:35,177 [INFO] Step 2700/3504 | Loss: 1.2991 | CE: 1.2991 | LR: 6.59e-05 | Epoch: 7/8
|
| 738 |
+
2026-06-24 16:56:43,633 [INFO] Step 2705/3504 | Loss: 1.1730 | CE: 1.1730 | LR: 6.51e-05 | Epoch: 7/8
|
| 739 |
+
2026-06-24 16:56:52,767 [INFO] Step 2710/3504 | Loss: 1.2647 | CE: 1.2647 | LR: 6.44e-05 | Epoch: 7/8
|
| 740 |
+
2026-06-24 16:57:03,372 [INFO] Step 2715/3504 | Loss: 1.2940 | CE: 1.2940 | LR: 6.36e-05 | Epoch: 7/8
|
| 741 |
+
2026-06-24 16:57:11,305 [INFO] Step 2720/3504 | Loss: 1.2371 | CE: 1.2371 | LR: 6.28e-05 | Epoch: 7/8
|
| 742 |
+
2026-06-24 16:57:19,417 [INFO] Step 2725/3504 | Loss: 1.2010 | CE: 1.2010 | LR: 6.20e-05 | Epoch: 7/8
|
| 743 |
+
2026-06-24 16:57:27,956 [INFO] Step 2730/3504 | Loss: 1.2657 | CE: 1.2657 | LR: 6.13e-05 | Epoch: 7/8
|
| 744 |
+
2026-06-24 16:57:35,631 [INFO] Step 2735/3504 | Loss: 1.1820 | CE: 1.1820 | LR: 6.05e-05 | Epoch: 7/8
|
| 745 |
+
2026-06-24 16:57:43,546 [INFO] Step 2740/3504 | Loss: 1.2539 | CE: 1.2539 | LR: 5.98e-05 | Epoch: 7/8
|
| 746 |
+
2026-06-24 16:57:53,196 [INFO] Step 2745/3504 | Loss: 1.2176 | CE: 1.2176 | LR: 5.90e-05 | Epoch: 7/8
|
| 747 |
+
2026-06-24 16:58:01,269 [INFO] Step 2750/3504 | Loss: 1.2481 | CE: 1.2481 | LR: 5.83e-05 | Epoch: 7/8
|
| 748 |
+
2026-06-24 16:58:09,331 [INFO] Step 2755/3504 | Loss: 1.2423 | CE: 1.2423 | LR: 5.76e-05 | Epoch: 7/8
|
| 749 |
+
2026-06-24 16:58:18,164 [INFO] Step 2760/3504 | Loss: 1.2690 | CE: 1.2690 | LR: 5.68e-05 | Epoch: 7/8
|
| 750 |
+
2026-06-24 16:58:26,721 [INFO] Step 2765/3504 | Loss: 1.2827 | CE: 1.2827 | LR: 5.61e-05 | Epoch: 7/8
|
| 751 |
+
2026-06-24 16:58:35,034 [INFO] Step 2770/3504 | Loss: 1.2318 | CE: 1.2318 | LR: 5.54e-05 | Epoch: 7/8
|
| 752 |
+
2026-06-24 16:58:44,978 [INFO] Step 2775/3504 | Loss: 1.2237 | CE: 1.2237 | LR: 5.46e-05 | Epoch: 7/8
|
| 753 |
+
2026-06-24 16:58:54,006 [INFO] Step 2780/3504 | Loss: 1.2116 | CE: 1.2116 | LR: 5.39e-05 | Epoch: 7/8
|
| 754 |
+
2026-06-24 16:59:03,172 [INFO] Step 2785/3504 | Loss: 1.2172 | CE: 1.2172 | LR: 5.32e-05 | Epoch: 7/8
|
| 755 |
+
2026-06-24 16:59:11,101 [INFO] Step 2790/3504 | Loss: 1.2553 | CE: 1.2553 | LR: 5.25e-05 | Epoch: 7/8
|
| 756 |
+
2026-06-24 16:59:19,605 [INFO] Step 2795/3504 | Loss: 1.3206 | CE: 1.3206 | LR: 5.18e-05 | Epoch: 7/8
|
| 757 |
+
2026-06-24 16:59:27,978 [INFO] Step 2800/3504 | Loss: 1.2884 | CE: 1.2884 | LR: 5.11e-05 | Epoch: 7/8
|
| 758 |
+
2026-06-24 16:59:36,716 [INFO] Step 2805/3504 | Loss: 1.2218 | CE: 1.2218 | LR: 5.04e-05 | Epoch: 7/8
|
| 759 |
+
2026-06-24 16:59:45,636 [INFO] Step 2810/3504 | Loss: 1.2733 | CE: 1.2733 | LR: 5.00e-05 | Epoch: 7/8
|
| 760 |
+
2026-06-24 16:59:54,400 [INFO] Step 2815/3504 | Loss: 1.2501 | CE: 1.2501 | LR: 5.00e-05 | Epoch: 7/8
|
| 761 |
+
2026-06-24 17:00:02,013 [INFO] Step 2820/3504 | Loss: 1.2717 | CE: 1.2717 | LR: 5.00e-05 | Epoch: 7/8
|
| 762 |
+
2026-06-24 17:00:10,800 [INFO] Step 2825/3504 | Loss: 1.1686 | CE: 1.1686 | LR: 5.00e-05 | Epoch: 7/8
|
| 763 |
+
2026-06-24 17:00:19,586 [INFO] Step 2830/3504 | Loss: 1.2446 | CE: 1.2446 | LR: 5.00e-05 | Epoch: 7/8
|
| 764 |
+
2026-06-24 17:00:27,097 [INFO] Step 2835/3504 | Loss: 1.2885 | CE: 1.2885 | LR: 5.00e-05 | Epoch: 7/8
|
| 765 |
+
2026-06-24 17:00:35,430 [INFO] Step 2840/3504 | Loss: 1.3170 | CE: 1.3170 | LR: 5.00e-05 | Epoch: 7/8
|
| 766 |
+
2026-06-24 17:00:44,086 [INFO] Step 2845/3504 | Loss: 1.2198 | CE: 1.2198 | LR: 5.00e-05 | Epoch: 7/8
|
| 767 |
+
2026-06-24 17:00:51,729 [INFO] Step 2850/3504 | Loss: 1.2409 | CE: 1.2409 | LR: 5.00e-05 | Epoch: 7/8
|
| 768 |
+
2026-06-24 17:01:00,808 [INFO] Step 2855/3504 | Loss: 1.2978 | CE: 1.2978 | LR: 5.00e-05 | Epoch: 7/8
|
| 769 |
+
2026-06-24 17:01:10,139 [INFO] Step 2860/3504 | Loss: 1.1634 | CE: 1.1634 | LR: 5.00e-05 | Epoch: 7/8
|
| 770 |
+
2026-06-24 17:01:19,589 [INFO] Step 2865/3504 | Loss: 1.2111 | CE: 1.2111 | LR: 5.00e-05 | Epoch: 7/8
|
| 771 |
+
2026-06-24 17:01:27,642 [INFO] Step 2870/3504 | Loss: 1.3122 | CE: 1.3122 | LR: 5.00e-05 | Epoch: 7/8
|
| 772 |
+
2026-06-24 17:01:37,904 [INFO] Step 2875/3504 | Loss: 1.2221 | CE: 1.2221 | LR: 5.00e-05 | Epoch: 7/8
|
| 773 |
+
2026-06-24 17:01:46,220 [INFO] Step 2880/3504 | Loss: 1.3376 | CE: 1.3376 | LR: 5.00e-05 | Epoch: 7/8
|
| 774 |
+
2026-06-24 17:01:55,669 [INFO] Step 2885/3504 | Loss: 1.2144 | CE: 1.2144 | LR: 5.00e-05 | Epoch: 7/8
|
| 775 |
+
2026-06-24 17:02:05,035 [INFO] Step 2890/3504 | Loss: 1.2292 | CE: 1.2292 | LR: 5.00e-05 | Epoch: 7/8
|
| 776 |
+
2026-06-24 17:02:13,008 [INFO] Step 2895/3504 | Loss: 1.2050 | CE: 1.2050 | LR: 5.00e-05 | Epoch: 7/8
|
| 777 |
+
2026-06-24 17:02:21,962 [INFO] Step 2900/3504 | Loss: 1.1742 | CE: 1.1742 | LR: 5.00e-05 | Epoch: 7/8
|
| 778 |
+
2026-06-24 17:02:29,753 [INFO] Step 2905/3504 | Loss: 1.2187 | CE: 1.2187 | LR: 5.00e-05 | Epoch: 7/8
|
| 779 |
+
2026-06-24 17:02:37,517 [INFO] Step 2910/3504 | Loss: 1.2367 | CE: 1.2367 | LR: 5.00e-05 | Epoch: 7/8
|
| 780 |
+
2026-06-24 17:02:47,448 [INFO] Step 2915/3504 | Loss: 1.2187 | CE: 1.2187 | LR: 5.00e-05 | Epoch: 7/8
|
| 781 |
+
2026-06-24 17:02:56,286 [INFO] Step 2920/3504 | Loss: 1.2771 | CE: 1.2771 | LR: 5.00e-05 | Epoch: 7/8
|
| 782 |
+
2026-06-24 17:03:04,707 [INFO] Step 2925/3504 | Loss: 1.2711 | CE: 1.2711 | LR: 5.00e-05 | Epoch: 7/8
|
| 783 |
+
2026-06-24 17:03:12,717 [INFO] Step 2930/3504 | Loss: 1.2559 | CE: 1.2559 | LR: 5.00e-05 | Epoch: 7/8
|
| 784 |
+
2026-06-24 17:03:22,406 [INFO] Step 2935/3504 | Loss: 1.2374 | CE: 1.2374 | LR: 5.00e-05 | Epoch: 7/8
|
| 785 |
+
2026-06-24 17:03:30,306 [INFO] Step 2940/3504 | Loss: 1.2137 | CE: 1.2137 | LR: 5.00e-05 | Epoch: 7/8
|
| 786 |
+
2026-06-24 17:03:39,572 [INFO] Step 2945/3504 | Loss: 1.1891 | CE: 1.1891 | LR: 5.00e-05 | Epoch: 7/8
|
| 787 |
+
2026-06-24 17:03:49,409 [INFO] Step 2950/3504 | Loss: 1.2384 | CE: 1.2384 | LR: 5.00e-05 | Epoch: 7/8
|
| 788 |
+
2026-06-24 17:03:57,456 [INFO] Step 2955/3504 | Loss: 1.2571 | CE: 1.2571 | LR: 5.00e-05 | Epoch: 7/8
|
| 789 |
+
2026-06-24 17:04:04,617 [INFO] Step 2960/3504 | Loss: 1.2147 | CE: 1.2147 | LR: 5.00e-05 | Epoch: 7/8
|
| 790 |
+
2026-06-24 17:04:12,895 [INFO] Step 2965/3504 | Loss: 1.2045 | CE: 1.2045 | LR: 5.00e-05 | Epoch: 7/8
|
| 791 |
+
2026-06-24 17:04:22,786 [INFO] Step 2970/3504 | Loss: 1.2671 | CE: 1.2671 | LR: 5.00e-05 | Epoch: 7/8
|
| 792 |
+
2026-06-24 17:04:31,320 [INFO] Step 2975/3504 | Loss: 1.2700 | CE: 1.2700 | LR: 5.00e-05 | Epoch: 7/8
|
| 793 |
+
2026-06-24 17:04:39,140 [INFO] Step 2980/3504 | Loss: 1.1828 | CE: 1.1828 | LR: 5.00e-05 | Epoch: 7/8
|
| 794 |
+
2026-06-24 17:04:48,205 [INFO] Step 2985/3504 | Loss: 1.2159 | CE: 1.2159 | LR: 5.00e-05 | Epoch: 7/8
|
| 795 |
+
2026-06-24 17:04:57,354 [INFO] Step 2990/3504 | Loss: 1.2225 | CE: 1.2225 | LR: 5.00e-05 | Epoch: 7/8
|
| 796 |
+
2026-06-24 17:05:05,657 [INFO] Step 2995/3504 | Loss: 1.1358 | CE: 1.1358 | LR: 5.00e-05 | Epoch: 7/8
|
| 797 |
+
2026-06-24 17:05:15,043 [INFO] Step 3000/3504 | Loss: 1.2189 | CE: 1.2189 | LR: 5.00e-05 | Epoch: 7/8
|
| 798 |
+
2026-06-24 17:05:24,957 [INFO] Step 3005/3504 | Loss: 1.2489 | CE: 1.2489 | LR: 5.00e-05 | Epoch: 7/8
|
| 799 |
+
2026-06-24 17:05:33,680 [INFO] Step 3010/3504 | Loss: 1.2563 | CE: 1.2563 | LR: 5.00e-05 | Epoch: 7/8
|
| 800 |
+
2026-06-24 17:05:42,187 [INFO] Step 3015/3504 | Loss: 1.1639 | CE: 1.1639 | LR: 5.00e-05 | Epoch: 7/8
|
| 801 |
+
2026-06-24 17:05:51,501 [INFO] Step 3020/3504 | Loss: 1.2041 | CE: 1.2041 | LR: 5.00e-05 | Epoch: 7/8
|
| 802 |
+
2026-06-24 17:06:00,160 [INFO] Step 3025/3504 | Loss: 1.2429 | CE: 1.2429 | LR: 5.00e-05 | Epoch: 7/8
|
| 803 |
+
2026-06-24 17:06:09,100 [INFO] Step 3030/3504 | Loss: 1.2188 | CE: 1.2188 | LR: 5.00e-05 | Epoch: 7/8
|
| 804 |
+
2026-06-24 17:06:17,826 [INFO] Step 3035/3504 | Loss: 1.2677 | CE: 1.2677 | LR: 5.00e-05 | Epoch: 7/8
|
| 805 |
+
2026-06-24 17:06:26,970 [INFO] Step 3040/3504 | Loss: 1.3035 | CE: 1.3035 | LR: 5.00e-05 | Epoch: 7/8
|
| 806 |
+
2026-06-24 17:06:34,707 [INFO] Step 3045/3504 | Loss: 1.1814 | CE: 1.1814 | LR: 5.00e-05 | Epoch: 7/8
|
| 807 |
+
2026-06-24 17:06:43,000 [INFO] Step 3050/3504 | Loss: 1.1590 | CE: 1.1590 | LR: 5.00e-05 | Epoch: 7/8
|
| 808 |
+
2026-06-24 17:06:51,331 [INFO] Step 3055/3504 | Loss: 1.2151 | CE: 1.2151 | LR: 5.00e-05 | Epoch: 7/8
|
| 809 |
+
2026-06-24 17:07:01,263 [INFO] Step 3060/3504 | Loss: 1.1839 | CE: 1.1839 | LR: 5.00e-05 | Epoch: 7/8
|
| 810 |
+
2026-06-24 17:07:09,659 [INFO] Step 3065/3504 | Loss: 1.1418 | CE: 1.1418 | LR: 5.00e-05 | Epoch: 8/8
|
| 811 |
+
2026-06-24 17:08:09,257 [INFO] Step 3066 | Eval Loss: 2.6812
|
| 812 |
+
2026-06-24 17:08:15,140 [INFO] Step 3070/3504 | Loss: 1.0684 | CE: 1.0684 | LR: 5.00e-05 | Epoch: 8/8
|
| 813 |
+
2026-06-24 17:08:23,324 [INFO] Step 3075/3504 | Loss: 1.0601 | CE: 1.0601 | LR: 5.00e-05 | Epoch: 8/8
|
| 814 |
+
2026-06-24 17:08:32,471 [INFO] Step 3080/3504 | Loss: 1.0977 | CE: 1.0977 | LR: 5.00e-05 | Epoch: 8/8
|
| 815 |
+
2026-06-24 17:08:41,061 [INFO] Step 3085/3504 | Loss: 1.0842 | CE: 1.0842 | LR: 5.00e-05 | Epoch: 8/8
|
| 816 |
+
2026-06-24 17:08:50,287 [INFO] Step 3090/3504 | Loss: 1.0327 | CE: 1.0327 | LR: 5.00e-05 | Epoch: 8/8
|
| 817 |
+
2026-06-24 17:08:59,107 [INFO] Step 3095/3504 | Loss: 1.0401 | CE: 1.0401 | LR: 5.00e-05 | Epoch: 8/8
|
| 818 |
+
2026-06-24 17:09:07,063 [INFO] Step 3100/3504 | Loss: 1.0621 | CE: 1.0621 | LR: 5.00e-05 | Epoch: 8/8
|
| 819 |
+
2026-06-24 17:09:15,621 [INFO] Step 3105/3504 | Loss: 1.1417 | CE: 1.1417 | LR: 5.00e-05 | Epoch: 8/8
|
| 820 |
+
2026-06-24 17:09:24,323 [INFO] Step 3110/3504 | Loss: 1.0687 | CE: 1.0687 | LR: 5.00e-05 | Epoch: 8/8
|
| 821 |
+
2026-06-24 17:09:32,535 [INFO] Step 3115/3504 | Loss: 1.0806 | CE: 1.0806 | LR: 5.00e-05 | Epoch: 8/8
|
| 822 |
+
2026-06-24 17:09:40,717 [INFO] Step 3120/3504 | Loss: 1.0944 | CE: 1.0944 | LR: 5.00e-05 | Epoch: 8/8
|
| 823 |
+
2026-06-24 17:09:49,383 [INFO] Step 3125/3504 | Loss: 1.0994 | CE: 1.0994 | LR: 5.00e-05 | Epoch: 8/8
|
| 824 |
+
2026-06-24 17:09:58,886 [INFO] Step 3130/3504 | Loss: 1.0407 | CE: 1.0407 | LR: 5.00e-05 | Epoch: 8/8
|
| 825 |
+
2026-06-24 17:10:08,583 [INFO] Step 3135/3504 | Loss: 1.0843 | CE: 1.0843 | LR: 5.00e-05 | Epoch: 8/8
|
| 826 |
+
2026-06-24 17:10:17,895 [INFO] Step 3140/3504 | Loss: 1.0776 | CE: 1.0776 | LR: 5.00e-05 | Epoch: 8/8
|
| 827 |
+
2026-06-24 17:10:27,560 [INFO] Step 3145/3504 | Loss: 1.0524 | CE: 1.0524 | LR: 5.00e-05 | Epoch: 8/8
|
| 828 |
+
2026-06-24 17:10:36,551 [INFO] Step 3150/3504 | Loss: 1.0814 | CE: 1.0814 | LR: 5.00e-05 | Epoch: 8/8
|
| 829 |
+
2026-06-24 17:10:45,648 [INFO] Step 3155/3504 | Loss: 1.0884 | CE: 1.0884 | LR: 5.00e-05 | Epoch: 8/8
|
| 830 |
+
2026-06-24 17:10:53,829 [INFO] Step 3160/3504 | Loss: 1.0479 | CE: 1.0479 | LR: 5.00e-05 | Epoch: 8/8
|
| 831 |
+
2026-06-24 17:11:02,986 [INFO] Step 3165/3504 | Loss: 1.0725 | CE: 1.0725 | LR: 5.00e-05 | Epoch: 8/8
|
| 832 |
+
2026-06-24 17:11:11,931 [INFO] Step 3170/3504 | Loss: 1.0596 | CE: 1.0596 | LR: 5.00e-05 | Epoch: 8/8
|
| 833 |
+
2026-06-24 17:11:19,309 [INFO] Step 3175/3504 | Loss: 1.0867 | CE: 1.0867 | LR: 5.00e-05 | Epoch: 8/8
|
| 834 |
+
2026-06-24 17:11:27,292 [INFO] Step 3180/3504 | Loss: 1.1391 | CE: 1.1391 | LR: 5.00e-05 | Epoch: 8/8
|
| 835 |
+
2026-06-24 17:11:36,401 [INFO] Step 3185/3504 | Loss: 1.0168 | CE: 1.0168 | LR: 5.00e-05 | Epoch: 8/8
|
| 836 |
+
2026-06-24 17:11:44,622 [INFO] Step 3190/3504 | Loss: 1.1125 | CE: 1.1125 | LR: 5.00e-05 | Epoch: 8/8
|
| 837 |
+
2026-06-24 17:11:53,902 [INFO] Step 3195/3504 | Loss: 1.0146 | CE: 1.0146 | LR: 5.00e-05 | Epoch: 8/8
|
| 838 |
+
2026-06-24 17:12:03,041 [INFO] Step 3200/3504 | Loss: 1.1037 | CE: 1.1037 | LR: 5.00e-05 | Epoch: 8/8
|
| 839 |
+
2026-06-24 17:12:11,516 [INFO] Step 3205/3504 | Loss: 1.0358 | CE: 1.0358 | LR: 5.00e-05 | Epoch: 8/8
|
| 840 |
+
2026-06-24 17:12:20,214 [INFO] Step 3210/3504 | Loss: 1.1129 | CE: 1.1129 | LR: 5.00e-05 | Epoch: 8/8
|
| 841 |
+
2026-06-24 17:12:28,398 [INFO] Step 3215/3504 | Loss: 1.0455 | CE: 1.0455 | LR: 5.00e-05 | Epoch: 8/8
|
| 842 |
+
2026-06-24 17:12:37,496 [INFO] Step 3220/3504 | Loss: 1.0481 | CE: 1.0481 | LR: 5.00e-05 | Epoch: 8/8
|
| 843 |
+
2026-06-24 17:12:45,820 [INFO] Step 3225/3504 | Loss: 1.0862 | CE: 1.0862 | LR: 5.00e-05 | Epoch: 8/8
|
| 844 |
+
2026-06-24 17:12:53,999 [INFO] Step 3230/3504 | Loss: 1.0128 | CE: 1.0128 | LR: 5.00e-05 | Epoch: 8/8
|
| 845 |
+
2026-06-24 17:13:03,542 [INFO] Step 3235/3504 | Loss: 1.1036 | CE: 1.1036 | LR: 5.00e-05 | Epoch: 8/8
|
| 846 |
+
2026-06-24 17:13:11,709 [INFO] Step 3240/3504 | Loss: 1.0775 | CE: 1.0775 | LR: 5.00e-05 | Epoch: 8/8
|
| 847 |
+
2026-06-24 17:13:21,018 [INFO] Step 3245/3504 | Loss: 1.1680 | CE: 1.1680 | LR: 5.00e-05 | Epoch: 8/8
|
| 848 |
+
2026-06-24 17:13:29,114 [INFO] Step 3250/3504 | Loss: 1.1458 | CE: 1.1458 | LR: 5.00e-05 | Epoch: 8/8
|
| 849 |
+
2026-06-24 17:13:36,663 [INFO] Step 3255/3504 | Loss: 1.0704 | CE: 1.0704 | LR: 5.00e-05 | Epoch: 8/8
|
| 850 |
+
2026-06-24 17:13:44,989 [INFO] Step 3260/3504 | Loss: 1.1079 | CE: 1.1079 | LR: 5.00e-05 | Epoch: 8/8
|
| 851 |
+
2026-06-24 17:13:55,797 [INFO] Step 3265/3504 | Loss: 1.1026 | CE: 1.1026 | LR: 5.00e-05 | Epoch: 8/8
|
| 852 |
+
2026-06-24 17:14:03,519 [INFO] Step 3270/3504 | Loss: 1.0504 | CE: 1.0504 | LR: 5.00e-05 | Epoch: 8/8
|
| 853 |
+
2026-06-24 17:14:12,674 [INFO] Step 3275/3504 | Loss: 1.0668 | CE: 1.0668 | LR: 5.00e-05 | Epoch: 8/8
|
| 854 |
+
2026-06-24 17:14:21,184 [INFO] Step 3280/3504 | Loss: 1.0384 | CE: 1.0384 | LR: 5.00e-05 | Epoch: 8/8
|
| 855 |
+
2026-06-24 17:14:30,205 [INFO] Step 3285/3504 | Loss: 1.1260 | CE: 1.1260 | LR: 5.00e-05 | Epoch: 8/8
|
| 856 |
+
2026-06-24 17:14:40,357 [INFO] Step 3290/3504 | Loss: 1.0677 | CE: 1.0677 | LR: 5.00e-05 | Epoch: 8/8
|
| 857 |
+
2026-06-24 17:14:49,490 [INFO] Step 3295/3504 | Loss: 1.0834 | CE: 1.0834 | LR: 5.00e-05 | Epoch: 8/8
|
| 858 |
+
2026-06-24 17:14:57,876 [INFO] Step 3300/3504 | Loss: 1.0580 | CE: 1.0580 | LR: 5.00e-05 | Epoch: 8/8
|
| 859 |
+
2026-06-24 17:15:06,289 [INFO] Step 3305/3504 | Loss: 1.1043 | CE: 1.1043 | LR: 5.00e-05 | Epoch: 8/8
|
| 860 |
+
2026-06-24 17:15:14,992 [INFO] Step 3310/3504 | Loss: 1.0681 | CE: 1.0681 | LR: 5.00e-05 | Epoch: 8/8
|
| 861 |
+
2026-06-24 17:15:23,431 [INFO] Step 3315/3504 | Loss: 1.1497 | CE: 1.1497 | LR: 5.00e-05 | Epoch: 8/8
|
| 862 |
+
2026-06-24 17:15:33,903 [INFO] Step 3320/3504 | Loss: 1.1511 | CE: 1.1511 | LR: 5.00e-05 | Epoch: 8/8
|
| 863 |
+
2026-06-24 17:15:42,776 [INFO] Step 3325/3504 | Loss: 1.0734 | CE: 1.0734 | LR: 5.00e-05 | Epoch: 8/8
|
| 864 |
+
2026-06-24 17:15:50,874 [INFO] Step 3330/3504 | Loss: 1.0898 | CE: 1.0898 | LR: 5.00e-05 | Epoch: 8/8
|
| 865 |
+
2026-06-24 17:16:00,683 [INFO] Step 3335/3504 | Loss: 1.1085 | CE: 1.1085 | LR: 5.00e-05 | Epoch: 8/8
|
| 866 |
+
2026-06-24 17:16:08,684 [INFO] Step 3340/3504 | Loss: 1.0743 | CE: 1.0743 | LR: 5.00e-05 | Epoch: 8/8
|
| 867 |
+
2026-06-24 17:16:16,990 [INFO] Step 3345/3504 | Loss: 1.0448 | CE: 1.0448 | LR: 5.00e-05 | Epoch: 8/8
|
| 868 |
+
2026-06-24 17:16:25,844 [INFO] Step 3350/3504 | Loss: 1.1234 | CE: 1.1234 | LR: 5.00e-05 | Epoch: 8/8
|
| 869 |
+
2026-06-24 17:16:34,674 [INFO] Step 3355/3504 | Loss: 1.1063 | CE: 1.1063 | LR: 5.00e-05 | Epoch: 8/8
|
| 870 |
+
2026-06-24 17:16:44,345 [INFO] Step 3360/3504 | Loss: 1.1318 | CE: 1.1318 | LR: 5.00e-05 | Epoch: 8/8
|
| 871 |
+
2026-06-24 17:16:54,014 [INFO] Step 3365/3504 | Loss: 1.1150 | CE: 1.1150 | LR: 5.00e-05 | Epoch: 8/8
|
| 872 |
+
2026-06-24 17:17:02,069 [INFO] Step 3370/3504 | Loss: 1.0629 | CE: 1.0629 | LR: 5.00e-05 | Epoch: 8/8
|
| 873 |
+
2026-06-24 17:17:10,953 [INFO] Step 3375/3504 | Loss: 1.0602 | CE: 1.0602 | LR: 5.00e-05 | Epoch: 8/8
|
| 874 |
+
2026-06-24 17:17:20,202 [INFO] Step 3380/3504 | Loss: 1.1153 | CE: 1.1153 | LR: 5.00e-05 | Epoch: 8/8
|
| 875 |
+
2026-06-24 17:17:29,510 [INFO] Step 3385/3504 | Loss: 1.0641 | CE: 1.0641 | LR: 5.00e-05 | Epoch: 8/8
|
| 876 |
+
2026-06-24 17:17:38,649 [INFO] Step 3390/3504 | Loss: 1.0015 | CE: 1.0015 | LR: 5.00e-05 | Epoch: 8/8
|
| 877 |
+
2026-06-24 17:17:46,463 [INFO] Step 3395/3504 | Loss: 1.0814 | CE: 1.0814 | LR: 5.00e-05 | Epoch: 8/8
|
| 878 |
+
2026-06-24 17:17:53,819 [INFO] Step 3400/3504 | Loss: 1.0898 | CE: 1.0898 | LR: 5.00e-05 | Epoch: 8/8
|
| 879 |
+
2026-06-24 17:18:01,767 [INFO] Step 3405/3504 | Loss: 1.0217 | CE: 1.0217 | LR: 5.00e-05 | Epoch: 8/8
|
| 880 |
+
2026-06-24 17:18:10,418 [INFO] Step 3410/3504 | Loss: 1.0445 | CE: 1.0445 | LR: 5.00e-05 | Epoch: 8/8
|
| 881 |
+
2026-06-24 17:18:19,515 [INFO] Step 3415/3504 | Loss: 1.1072 | CE: 1.1072 | LR: 5.00e-05 | Epoch: 8/8
|
| 882 |
+
2026-06-24 17:18:27,510 [INFO] Step 3420/3504 | Loss: 1.0693 | CE: 1.0693 | LR: 5.00e-05 | Epoch: 8/8
|
| 883 |
+
2026-06-24 17:18:35,979 [INFO] Step 3425/3504 | Loss: 1.0640 | CE: 1.0640 | LR: 5.00e-05 | Epoch: 8/8
|
| 884 |
+
2026-06-24 17:18:44,506 [INFO] Step 3430/3504 | Loss: 1.1398 | CE: 1.1398 | LR: 5.00e-05 | Epoch: 8/8
|
| 885 |
+
2026-06-24 17:18:52,501 [INFO] Step 3435/3504 | Loss: 1.0829 | CE: 1.0829 | LR: 5.00e-05 | Epoch: 8/8
|
| 886 |
+
2026-06-24 17:19:01,788 [INFO] Step 3440/3504 | Loss: 1.1106 | CE: 1.1106 | LR: 5.00e-05 | Epoch: 8/8
|
| 887 |
+
2026-06-24 17:19:10,797 [INFO] Step 3445/3504 | Loss: 1.1526 | CE: 1.1526 | LR: 5.00e-05 | Epoch: 8/8
|
| 888 |
+
2026-06-24 17:19:19,857 [INFO] Step 3450/3504 | Loss: 1.1004 | CE: 1.1004 | LR: 5.00e-05 | Epoch: 8/8
|
| 889 |
+
2026-06-24 17:19:27,216 [INFO] Step 3455/3504 | Loss: 1.1104 | CE: 1.1104 | LR: 5.00e-05 | Epoch: 8/8
|
| 890 |
+
2026-06-24 17:19:34,482 [INFO] Step 3460/3504 | Loss: 1.0896 | CE: 1.0896 | LR: 5.00e-05 | Epoch: 8/8
|
| 891 |
+
2026-06-24 17:19:43,928 [INFO] Step 3465/3504 | Loss: 1.0503 | CE: 1.0503 | LR: 5.00e-05 | Epoch: 8/8
|
| 892 |
+
2026-06-24 17:19:52,504 [INFO] Step 3470/3504 | Loss: 1.1018 | CE: 1.1018 | LR: 5.00e-05 | Epoch: 8/8
|
| 893 |
+
2026-06-24 17:20:01,001 [INFO] Step 3475/3504 | Loss: 1.0500 | CE: 1.0500 | LR: 5.00e-05 | Epoch: 8/8
|
| 894 |
+
2026-06-24 17:20:09,889 [INFO] Step 3480/3504 | Loss: 1.1511 | CE: 1.1511 | LR: 5.00e-05 | Epoch: 8/8
|
| 895 |
+
2026-06-24 17:20:18,646 [INFO] Step 3485/3504 | Loss: 1.0059 | CE: 1.0059 | LR: 5.00e-05 | Epoch: 8/8
|
| 896 |
+
2026-06-24 17:20:28,646 [INFO] Step 3490/3504 | Loss: 1.0887 | CE: 1.0887 | LR: 5.00e-05 | Epoch: 8/8
|
| 897 |
+
2026-06-24 17:20:36,398 [INFO] Step 3495/3504 | Loss: 1.0453 | CE: 1.0453 | LR: 5.00e-05 | Epoch: 8/8
|
| 898 |
+
2026-06-24 17:20:46,195 [INFO] Step 3500/3504 | Loss: 1.1233 | CE: 1.1233 | LR: 5.00e-05 | Epoch: 9/8
|
| 899 |
+
2026-06-24 17:21:51,993 [INFO] Step 3504 | Eval Loss: 2.7424
|
| 900 |
+
2026-06-24 17:21:53,591 [INFO] Uploading to HF: runs/phase1_warmup_qwen_wikilingua/best.pt -> BienKieu/llm2seq-wikilingua/checkpoints/phase1_warmup_qwen_wikilingua/best.pt
|
| 901 |
+
2026-06-24 17:21:53,762 [INFO] HTTP Request: POST https://huggingface.co/api/models/BienKieu/llm2seq-wikilingua/preupload/main "HTTP/1.1 200 OK"
|
| 902 |
+
2026-06-24 17:26:38,807 [INFO] HTTP Request: POST https://huggingface.co/api/models/BienKieu/llm2seq-wikilingua/commit/main "HTTP/1.1 200 OK"
|
| 903 |
+
2026-06-24 17:26:38,808 [INFO] Uploading to HF: runs/phase1_warmup_qwen_wikilingua/config.yaml -> BienKieu/llm2seq-wikilingua/checkpoints/phase1_warmup_qwen_wikilingua/config.yaml
|
| 904 |
+
2026-06-24 17:26:38,966 [INFO] HTTP Request: POST https://huggingface.co/api/models/BienKieu/llm2seq-wikilingua/preupload/main "HTTP/1.1 200 OK"
|
| 905 |
+
2026-06-24 17:26:39,432 [INFO] HTTP Request: POST https://huggingface.co/api/models/BienKieu/llm2seq-wikilingua/commit/main "HTTP/1.1 200 OK"
|
| 906 |
+
2026-06-24 17:26:39,432 [INFO] Uploading to HF: runs/phase1_warmup_qwen_wikilingua/train.log -> BienKieu/llm2seq-wikilingua/checkpoints/phase1_warmup_qwen_wikilingua/train.log
|
| 907 |
+
2026-06-24 17:26:39,615 [INFO] HTTP Request: POST https://huggingface.co/api/models/BienKieu/llm2seq-wikilingua/preupload/main "HTTP/1.1 200 OK"
|