Add files using upload-large-folder tool
Browse filesThis view is limited to 50 files because it contains too many changes. See raw diff
- DBCA_original_Estonian/Qwen3.5-2B-Base_original_features_structural_train_original_features_structural_test1/checkpoint-2568/trainer_state.json +610 -0
- DBCA_original_Estonian/Qwen3.5-2B-Base_original_features_structural_train_original_features_structural_test1/checkpoint-2996/README.md +209 -0
- DBCA_original_Estonian/Qwen3.5-2B-Base_original_features_structural_train_original_features_structural_test1/checkpoint-2996/adapter_config.json +48 -0
- DBCA_original_Estonian/Qwen3.5-2B-Base_original_features_structural_train_original_features_structural_test1/checkpoint-2996/chat_template.jinja +154 -0
- DBCA_original_Estonian/Qwen3.5-2B-Base_original_features_structural_train_original_features_structural_test1/checkpoint-2996/tokenizer_config.json +31 -0
- DBCA_original_Estonian/Qwen3.5-2B-Base_original_features_structural_train_original_features_structural_test1/checkpoint-2996/trainer_state.json +701 -0
- DBCA_original_Estonian/Qwen3.5-2B-Base_original_features_structural_train_original_features_structural_test1/checkpoint-3424/README.md +209 -0
- DBCA_original_Estonian/Qwen3.5-2B-Base_original_features_structural_train_original_features_structural_test1/checkpoint-3424/adapter_config.json +48 -0
- DBCA_original_Estonian/Qwen3.5-2B-Base_original_features_structural_train_original_features_structural_test1/checkpoint-3424/chat_template.jinja +154 -0
- DBCA_original_Estonian/Qwen3.5-2B-Base_original_features_structural_train_original_features_structural_test1/checkpoint-3424/tokenizer_config.json +31 -0
- DBCA_original_Estonian/Qwen3.5-2B-Base_original_features_structural_train_original_features_structural_test1/checkpoint-3424/trainer_state.json +802 -0
- DBCA_original_Estonian/Qwen3.5-2B-Base_original_features_structural_train_original_features_structural_test1/checkpoint-3852/README.md +209 -0
- DBCA_original_Estonian/Qwen3.5-2B-Base_original_features_structural_train_original_features_structural_test1/checkpoint-3852/adapter_config.json +48 -0
- DBCA_original_Estonian/Qwen3.5-2B-Base_original_features_structural_train_original_features_structural_test1/checkpoint-4280/tokenizer_config.json +31 -0
- substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/README.md +58 -0
- substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-1224/README.md +209 -0
- substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-1224/adapter_config.json +48 -0
- substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-1224/chat_template.jinja +154 -0
- substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-1224/tokenizer_config.json +31 -0
- substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-1224/trainer_state.json +340 -0
- substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-1632/README.md +209 -0
- substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-1632/adapter_config.json +48 -0
- substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/README.md +58 -0
- substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-1224/README.md +209 -0
- substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-1224/adapter_config.json +48 -0
- substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-1224/chat_template.jinja +154 -0
- substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-1224/tokenizer_config.json +31 -0
- substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-1224/trainer_state.json +340 -0
- substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-1632/adapter_config.json +48 -0
- substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-1632/chat_template.jinja +154 -0
- substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-1632/tokenizer_config.json +31 -0
- substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-1632/trainer_state.json +442 -0
- substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-2040/README.md +209 -0
- substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-2040/adapter_config.json +48 -0
- substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-2040/chat_template.jinja +154 -0
- substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-2040/tokenizer_config.json +31 -0
- substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-2040/trainer_state.json +544 -0
- substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-2448/README.md +209 -0
- substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-2448/adapter_config.json +48 -0
- substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-2448/chat_template.jinja +154 -0
- substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-2448/tokenizer_config.json +31 -0
- substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-2448/trainer_state.json +646 -0
- substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-2856/README.md +209 -0
- substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-2856/adapter_config.json +48 -0
- substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-2856/chat_template.jinja +154 -0
- substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-2856/tokenizer_config.json +31 -0
- substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-2856/trainer_state.json +758 -0
- substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-3264/README.md +209 -0
- substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-3264/adapter_config.json +48 -0
- substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-3264/chat_template.jinja +154 -0
DBCA_original_Estonian/Qwen3.5-2B-Base_original_features_structural_train_original_features_structural_test1/checkpoint-2568/trainer_state.json
ADDED
|
@@ -0,0 +1,610 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": null,
|
| 3 |
+
"best_metric": null,
|
| 4 |
+
"best_model_checkpoint": null,
|
| 5 |
+
"epoch": 6.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 2568,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"entropy": 1.8341716039180755,
|
| 14 |
+
"epoch": 0.11689070718877849,
|
| 15 |
+
"grad_norm": 1.1992322206497192,
|
| 16 |
+
"learning_rate": 5.366728168873013e-05,
|
| 17 |
+
"loss": 1.6918942260742187,
|
| 18 |
+
"mean_token_accuracy": 0.631798365265131,
|
| 19 |
+
"num_tokens": 153689.0,
|
| 20 |
+
"step": 50
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"entropy": 0.9513179230690002,
|
| 24 |
+
"epoch": 0.23378141437755698,
|
| 25 |
+
"grad_norm": 1.1290454864501953,
|
| 26 |
+
"learning_rate": 0.00010842981402416902,
|
| 27 |
+
"loss": 0.8867578887939453,
|
| 28 |
+
"mean_token_accuracy": 0.7587438315153122,
|
| 29 |
+
"num_tokens": 310548.0,
|
| 30 |
+
"step": 100
|
| 31 |
+
},
|
| 32 |
+
{
|
| 33 |
+
"entropy": 0.8510265004634857,
|
| 34 |
+
"epoch": 0.3506721215663355,
|
| 35 |
+
"grad_norm": 0.7732136845588684,
|
| 36 |
+
"learning_rate": 0.00016319234635960792,
|
| 37 |
+
"loss": 0.7856448364257812,
|
| 38 |
+
"mean_token_accuracy": 0.7769897204637527,
|
| 39 |
+
"num_tokens": 468560.0,
|
| 40 |
+
"step": 150
|
| 41 |
+
},
|
| 42 |
+
{
|
| 43 |
+
"entropy": 0.8142367601394653,
|
| 44 |
+
"epoch": 0.46756282875511396,
|
| 45 |
+
"grad_norm": 0.7580806016921997,
|
| 46 |
+
"learning_rate": 0.00021795487869504684,
|
| 47 |
+
"loss": 0.7461666870117187,
|
| 48 |
+
"mean_token_accuracy": 0.7888966089487076,
|
| 49 |
+
"num_tokens": 619552.0,
|
| 50 |
+
"step": 200
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"entropy": 0.772578022480011,
|
| 54 |
+
"epoch": 0.5844535359438925,
|
| 55 |
+
"grad_norm": 0.5330358147621155,
|
| 56 |
+
"learning_rate": 0.00027271741103048575,
|
| 57 |
+
"loss": 0.7142900848388671,
|
| 58 |
+
"mean_token_accuracy": 0.7959607627987861,
|
| 59 |
+
"num_tokens": 771972.0,
|
| 60 |
+
"step": 250
|
| 61 |
+
},
|
| 62 |
+
{
|
| 63 |
+
"entropy": 0.7481484657526016,
|
| 64 |
+
"epoch": 0.701344243132671,
|
| 65 |
+
"grad_norm": 0.8242517709732056,
|
| 66 |
+
"learning_rate": 0.00032747994336592464,
|
| 67 |
+
"loss": 0.6911422729492187,
|
| 68 |
+
"mean_token_accuracy": 0.8028205358982086,
|
| 69 |
+
"num_tokens": 928088.0,
|
| 70 |
+
"step": 300
|
| 71 |
+
},
|
| 72 |
+
{
|
| 73 |
+
"entropy": 0.741372903585434,
|
| 74 |
+
"epoch": 0.8182349503214494,
|
| 75 |
+
"grad_norm": 0.7722981572151184,
|
| 76 |
+
"learning_rate": 0.00038224247570136353,
|
| 77 |
+
"loss": 0.6908904266357422,
|
| 78 |
+
"mean_token_accuracy": 0.8018302822113037,
|
| 79 |
+
"num_tokens": 1086836.0,
|
| 80 |
+
"step": 350
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"entropy": 0.7257227802276611,
|
| 84 |
+
"epoch": 0.9351256575102279,
|
| 85 |
+
"grad_norm": 0.7969573140144348,
|
| 86 |
+
"learning_rate": 0.0004370050080368024,
|
| 87 |
+
"loss": 0.6714310455322265,
|
| 88 |
+
"mean_token_accuracy": 0.8057891410589219,
|
| 89 |
+
"num_tokens": 1241227.0,
|
| 90 |
+
"step": 400
|
| 91 |
+
},
|
| 92 |
+
{
|
| 93 |
+
"epoch": 1.0,
|
| 94 |
+
"eval_entropy": 0.8317563653766334,
|
| 95 |
+
"eval_loss": 0.8135058283805847,
|
| 96 |
+
"eval_mean_token_accuracy": 0.7755894215850087,
|
| 97 |
+
"eval_num_tokens": 1325643.0,
|
| 98 |
+
"eval_runtime": 37.4951,
|
| 99 |
+
"eval_samples_per_second": 32.831,
|
| 100 |
+
"eval_steps_per_second": 4.107,
|
| 101 |
+
"step": 428
|
| 102 |
+
},
|
| 103 |
+
{
|
| 104 |
+
"entropy": 0.7179872094087265,
|
| 105 |
+
"epoch": 1.0514319111630626,
|
| 106 |
+
"grad_norm": 0.6224768161773682,
|
| 107 |
+
"learning_rate": 0.00046873290101955413,
|
| 108 |
+
"loss": 0.6683222961425781,
|
| 109 |
+
"mean_token_accuracy": 0.8054787133207273,
|
| 110 |
+
"num_tokens": 1393922.0,
|
| 111 |
+
"step": 450
|
| 112 |
+
},
|
| 113 |
+
{
|
| 114 |
+
"entropy": 0.6991294291615486,
|
| 115 |
+
"epoch": 1.1683226183518411,
|
| 116 |
+
"grad_norm": 0.6813111901283264,
|
| 117 |
+
"learning_rate": 0.00046837443306086947,
|
| 118 |
+
"loss": 0.6493977355957031,
|
| 119 |
+
"mean_token_accuracy": 0.8089349576830864,
|
| 120 |
+
"num_tokens": 1551442.0,
|
| 121 |
+
"step": 500
|
| 122 |
+
},
|
| 123 |
+
{
|
| 124 |
+
"entropy": 0.6951853144168854,
|
| 125 |
+
"epoch": 1.2852133255406195,
|
| 126 |
+
"grad_norm": 0.617091953754425,
|
| 127 |
+
"learning_rate": 0.0004676269147738558,
|
| 128 |
+
"loss": 0.6484123229980469,
|
| 129 |
+
"mean_token_accuracy": 0.8088837671279907,
|
| 130 |
+
"num_tokens": 1706697.0,
|
| 131 |
+
"step": 550
|
| 132 |
+
},
|
| 133 |
+
{
|
| 134 |
+
"entropy": 0.6852494943141937,
|
| 135 |
+
"epoch": 1.4021040327293979,
|
| 136 |
+
"grad_norm": 0.4677598774433136,
|
| 137 |
+
"learning_rate": 0.0004664915890374708,
|
| 138 |
+
"loss": 0.6416233062744141,
|
| 139 |
+
"mean_token_accuracy": 0.8106312158703805,
|
| 140 |
+
"num_tokens": 1865246.0,
|
| 141 |
+
"step": 600
|
| 142 |
+
},
|
| 143 |
+
{
|
| 144 |
+
"entropy": 0.6703594943881035,
|
| 145 |
+
"epoch": 1.5189947399181765,
|
| 146 |
+
"grad_norm": 0.5708025693893433,
|
| 147 |
+
"learning_rate": 0.0004649703435278991,
|
| 148 |
+
"loss": 0.6273183441162109,
|
| 149 |
+
"mean_token_accuracy": 0.8155005398392677,
|
| 150 |
+
"num_tokens": 2022060.0,
|
| 151 |
+
"step": 650
|
| 152 |
+
},
|
| 153 |
+
{
|
| 154 |
+
"entropy": 0.6676286320388317,
|
| 155 |
+
"epoch": 1.635885447106955,
|
| 156 |
+
"grad_norm": 0.5555347800254822,
|
| 157 |
+
"learning_rate": 0.00046306570757996264,
|
| 158 |
+
"loss": 0.6276700973510743,
|
| 159 |
+
"mean_token_accuracy": 0.8154828292131424,
|
| 160 |
+
"num_tokens": 2178909.0,
|
| 161 |
+
"step": 700
|
| 162 |
+
},
|
| 163 |
+
{
|
| 164 |
+
"entropy": 0.6669242936372757,
|
| 165 |
+
"epoch": 1.7527761542957334,
|
| 166 |
+
"grad_norm": 0.5260112285614014,
|
| 167 |
+
"learning_rate": 0.0004607808479816624,
|
| 168 |
+
"loss": 0.6228141784667969,
|
| 169 |
+
"mean_token_accuracy": 0.8164840793609619,
|
| 170 |
+
"num_tokens": 2329314.0,
|
| 171 |
+
"step": 750
|
| 172 |
+
},
|
| 173 |
+
{
|
| 174 |
+
"entropy": 0.6598560312390327,
|
| 175 |
+
"epoch": 1.869666861484512,
|
| 176 |
+
"grad_norm": 0.5677736401557922,
|
| 177 |
+
"learning_rate": 0.0004581195637088436,
|
| 178 |
+
"loss": 0.6214292907714843,
|
| 179 |
+
"mean_token_accuracy": 0.8171795177459716,
|
| 180 |
+
"num_tokens": 2478559.0,
|
| 181 |
+
"step": 800
|
| 182 |
+
},
|
| 183 |
+
{
|
| 184 |
+
"entropy": 0.6687870016694069,
|
| 185 |
+
"epoch": 1.9865575686732906,
|
| 186 |
+
"grad_norm": 0.6449615955352783,
|
| 187 |
+
"learning_rate": 0.00045508627960873823,
|
| 188 |
+
"loss": 0.6243909454345703,
|
| 189 |
+
"mean_token_accuracy": 0.8147780740261078,
|
| 190 |
+
"num_tokens": 2631937.0,
|
| 191 |
+
"step": 850
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"epoch": 2.0,
|
| 195 |
+
"eval_entropy": 0.7786144471013701,
|
| 196 |
+
"eval_loss": 0.7599140405654907,
|
| 197 |
+
"eval_mean_token_accuracy": 0.7935443106409791,
|
| 198 |
+
"eval_num_tokens": 2651286.0,
|
| 199 |
+
"eval_runtime": 39.0048,
|
| 200 |
+
"eval_samples_per_second": 31.56,
|
| 201 |
+
"eval_steps_per_second": 3.948,
|
| 202 |
+
"step": 856
|
| 203 |
+
},
|
| 204 |
+
{
|
| 205 |
+
"entropy": 0.5821921329701966,
|
| 206 |
+
"epoch": 2.102863822326125,
|
| 207 |
+
"grad_norm": 0.6116960048675537,
|
| 208 |
+
"learning_rate": 0.00045168603904288863,
|
| 209 |
+
"loss": 0.5406004714965821,
|
| 210 |
+
"mean_token_accuracy": 0.8334921918921734,
|
| 211 |
+
"num_tokens": 2795200.0,
|
| 212 |
+
"step": 900
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"entropy": 0.5937331764400006,
|
| 216 |
+
"epoch": 2.2197545295149035,
|
| 217 |
+
"grad_norm": 0.5440122485160828,
|
| 218 |
+
"learning_rate": 0.00044792449550168286,
|
| 219 |
+
"loss": 0.5515737533569336,
|
| 220 |
+
"mean_token_accuracy": 0.8314691257476806,
|
| 221 |
+
"num_tokens": 2950534.0,
|
| 222 |
+
"step": 950
|
| 223 |
+
},
|
| 224 |
+
{
|
| 225 |
+
"entropy": 0.5877273553609847,
|
| 226 |
+
"epoch": 2.3366452367036823,
|
| 227 |
+
"grad_norm": 0.6505260467529297,
|
| 228 |
+
"learning_rate": 0.0004438079032044453,
|
| 229 |
+
"loss": 0.5507744979858399,
|
| 230 |
+
"mean_token_accuracy": 0.8314063146710395,
|
| 231 |
+
"num_tokens": 3101471.0,
|
| 232 |
+
"step": 1000
|
| 233 |
+
},
|
| 234 |
+
{
|
| 235 |
+
"entropy": 0.5897100016474723,
|
| 236 |
+
"epoch": 2.4535359438924607,
|
| 237 |
+
"grad_norm": 0.54567551612854,
|
| 238 |
+
"learning_rate": 0.0004393431067007111,
|
| 239 |
+
"loss": 0.5528768157958984,
|
| 240 |
+
"mean_token_accuracy": 0.8315245220065117,
|
| 241 |
+
"num_tokens": 3258554.0,
|
| 242 |
+
"step": 1050
|
| 243 |
+
},
|
| 244 |
+
{
|
| 245 |
+
"entropy": 0.5831590622663498,
|
| 246 |
+
"epoch": 2.570426651081239,
|
| 247 |
+
"grad_norm": 0.44826796650886536,
|
| 248 |
+
"learning_rate": 0.00043453752948997376,
|
| 249 |
+
"loss": 0.5442767333984375,
|
| 250 |
+
"mean_token_accuracy": 0.8327079233527184,
|
| 251 |
+
"num_tokens": 3416088.0,
|
| 252 |
+
"step": 1100
|
| 253 |
+
},
|
| 254 |
+
{
|
| 255 |
+
"entropy": 0.5839906217157841,
|
| 256 |
+
"epoch": 2.6873173582700174,
|
| 257 |
+
"grad_norm": 0.5067696571350098,
|
| 258 |
+
"learning_rate": 0.0004293991616788285,
|
| 259 |
+
"loss": 0.5485663604736328,
|
| 260 |
+
"mean_token_accuracy": 0.8324281191825866,
|
| 261 |
+
"num_tokens": 3567251.0,
|
| 262 |
+
"step": 1150
|
| 263 |
+
},
|
| 264 |
+
{
|
| 265 |
+
"entropy": 0.5870274990797043,
|
| 266 |
+
"epoch": 2.8042080654587958,
|
| 267 |
+
"grad_norm": 0.44032806158065796,
|
| 268 |
+
"learning_rate": 0.00042393654669603217,
|
| 269 |
+
"loss": 0.5524474716186524,
|
| 270 |
+
"mean_token_accuracy": 0.8313613015413285,
|
| 271 |
+
"num_tokens": 3719464.0,
|
| 272 |
+
"step": 1200
|
| 273 |
+
},
|
| 274 |
+
{
|
| 275 |
+
"entropy": 0.5864466108381748,
|
| 276 |
+
"epoch": 2.9210987726475746,
|
| 277 |
+
"grad_norm": 0.5520864129066467,
|
| 278 |
+
"learning_rate": 0.00041815876708756964,
|
| 279 |
+
"loss": 0.5490786361694336,
|
| 280 |
+
"mean_token_accuracy": 0.8314764249324799,
|
| 281 |
+
"num_tokens": 3874423.0,
|
| 282 |
+
"step": 1250
|
| 283 |
+
},
|
| 284 |
+
{
|
| 285 |
+
"epoch": 3.0,
|
| 286 |
+
"eval_entropy": 0.6819627613990338,
|
| 287 |
+
"eval_loss": 0.7593190670013428,
|
| 288 |
+
"eval_mean_token_accuracy": 0.7857603395914102,
|
| 289 |
+
"eval_num_tokens": 3976929.0,
|
| 290 |
+
"eval_runtime": 39.3654,
|
| 291 |
+
"eval_samples_per_second": 31.271,
|
| 292 |
+
"eval_steps_per_second": 3.912,
|
| 293 |
+
"step": 1284
|
| 294 |
+
},
|
| 295 |
+
{
|
| 296 |
+
"entropy": 0.5587641523411525,
|
| 297 |
+
"epoch": 3.037405026300409,
|
| 298 |
+
"grad_norm": 0.622401773929596,
|
| 299 |
+
"learning_rate": 0.0004120754294153441,
|
| 300 |
+
"loss": 0.518932762145996,
|
| 301 |
+
"mean_token_accuracy": 0.8399260457436643,
|
| 302 |
+
"num_tokens": 4027609.0,
|
| 303 |
+
"step": 1300
|
| 304 |
+
},
|
| 305 |
+
{
|
| 306 |
+
"entropy": 0.5028628017008304,
|
| 307 |
+
"epoch": 3.1542957334891875,
|
| 308 |
+
"grad_norm": 0.577396810054779,
|
| 309 |
+
"learning_rate": 0.00040569664828459917,
|
| 310 |
+
"loss": 0.4589382171630859,
|
| 311 |
+
"mean_token_accuracy": 0.8538418188691139,
|
| 312 |
+
"num_tokens": 4180560.0,
|
| 313 |
+
"step": 1350
|
| 314 |
+
},
|
| 315 |
+
{
|
| 316 |
+
"entropy": 0.5227407096326351,
|
| 317 |
+
"epoch": 3.2711864406779663,
|
| 318 |
+
"grad_norm": 0.5113071203231812,
|
| 319 |
+
"learning_rate": 0.00039903302952663176,
|
| 320 |
+
"loss": 0.47863777160644533,
|
| 321 |
+
"mean_token_accuracy": 0.847364938557148,
|
| 322 |
+
"num_tokens": 4329954.0,
|
| 323 |
+
"step": 1400
|
| 324 |
+
},
|
| 325 |
+
{
|
| 326 |
+
"entropy": 0.5050077450275421,
|
| 327 |
+
"epoch": 3.3880771478667446,
|
| 328 |
+
"grad_norm": 0.48977184295654297,
|
| 329 |
+
"learning_rate": 0.0003920956525647558,
|
| 330 |
+
"loss": 0.46841480255126955,
|
| 331 |
+
"mean_token_accuracy": 0.8519425508379936,
|
| 332 |
+
"num_tokens": 4484354.0,
|
| 333 |
+
"step": 1450
|
| 334 |
+
},
|
| 335 |
+
{
|
| 336 |
+
"entropy": 0.5163291451334954,
|
| 337 |
+
"epoch": 3.504967855055523,
|
| 338 |
+
"grad_norm": 0.509864330291748,
|
| 339 |
+
"learning_rate": 0.000384896051992837,
|
| 340 |
+
"loss": 0.47587432861328127,
|
| 341 |
+
"mean_token_accuracy": 0.8486990982294083,
|
| 342 |
+
"num_tokens": 4638919.0,
|
| 343 |
+
"step": 1500
|
| 344 |
+
},
|
| 345 |
+
{
|
| 346 |
+
"entropy": 0.5134807989001274,
|
| 347 |
+
"epoch": 3.6218585622443014,
|
| 348 |
+
"grad_norm": 0.46544620394706726,
|
| 349 |
+
"learning_rate": 0.00037744619839702735,
|
| 350 |
+
"loss": 0.47692710876464844,
|
| 351 |
+
"mean_token_accuracy": 0.8489899519085884,
|
| 352 |
+
"num_tokens": 4795540.0,
|
| 353 |
+
"step": 1550
|
| 354 |
+
},
|
| 355 |
+
{
|
| 356 |
+
"entropy": 0.5216251534223556,
|
| 357 |
+
"epoch": 3.73874926943308,
|
| 358 |
+
"grad_norm": 0.5167004466056824,
|
| 359 |
+
"learning_rate": 0.0003697584784525874,
|
| 360 |
+
"loss": 0.4837848663330078,
|
| 361 |
+
"mean_token_accuracy": 0.847066233754158,
|
| 362 |
+
"num_tokens": 4951177.0,
|
| 363 |
+
"step": 1600
|
| 364 |
+
},
|
| 365 |
+
{
|
| 366 |
+
"entropy": 0.5254101701080799,
|
| 367 |
+
"epoch": 3.8556399766218585,
|
| 368 |
+
"grad_norm": 0.5504006743431091,
|
| 369 |
+
"learning_rate": 0.00036184567432888745,
|
| 370 |
+
"loss": 0.48759506225585936,
|
| 371 |
+
"mean_token_accuracy": 0.8457375919818878,
|
| 372 |
+
"num_tokens": 5104814.0,
|
| 373 |
+
"step": 1650
|
| 374 |
+
},
|
| 375 |
+
{
|
| 376 |
+
"entropy": 0.5107753933966159,
|
| 377 |
+
"epoch": 3.972530683810637,
|
| 378 |
+
"grad_norm": 0.421975314617157,
|
| 379 |
+
"learning_rate": 0.0003537209424368311,
|
| 380 |
+
"loss": 0.4743759536743164,
|
| 381 |
+
"mean_token_accuracy": 0.8505110186338425,
|
| 382 |
+
"num_tokens": 5265185.0,
|
| 383 |
+
"step": 1700
|
| 384 |
+
},
|
| 385 |
+
{
|
| 386 |
+
"epoch": 4.0,
|
| 387 |
+
"eval_entropy": 0.665543915002377,
|
| 388 |
+
"eval_loss": 0.7487082481384277,
|
| 389 |
+
"eval_mean_token_accuracy": 0.7999678904359991,
|
| 390 |
+
"eval_num_tokens": 5302572.0,
|
| 391 |
+
"eval_runtime": 38.5439,
|
| 392 |
+
"eval_samples_per_second": 31.938,
|
| 393 |
+
"eval_steps_per_second": 3.995,
|
| 394 |
+
"step": 1712
|
| 395 |
+
},
|
| 396 |
+
{
|
| 397 |
+
"entropy": 0.4414465431891494,
|
| 398 |
+
"epoch": 4.088836937463472,
|
| 399 |
+
"grad_norm": 0.5308493971824646,
|
| 400 |
+
"learning_rate": 0.0003453977915540383,
|
| 401 |
+
"loss": 0.39650299072265627,
|
| 402 |
+
"mean_token_accuracy": 0.8703725772287378,
|
| 403 |
+
"num_tokens": 5419002.0,
|
| 404 |
+
"step": 1750
|
| 405 |
+
},
|
| 406 |
+
{
|
| 407 |
+
"entropy": 0.42591195791959763,
|
| 408 |
+
"epoch": 4.20572764465225,
|
| 409 |
+
"grad_norm": 0.6111563444137573,
|
| 410 |
+
"learning_rate": 0.00033689006036415585,
|
| 411 |
+
"loss": 0.37838283538818357,
|
| 412 |
+
"mean_token_accuracy": 0.874235480427742,
|
| 413 |
+
"num_tokens": 5572969.0,
|
| 414 |
+
"step": 1800
|
| 415 |
+
},
|
| 416 |
+
{
|
| 417 |
+
"entropy": 0.44086351931095125,
|
| 418 |
+
"epoch": 4.322618351841029,
|
| 419 |
+
"grad_norm": 0.5524880290031433,
|
| 420 |
+
"learning_rate": 0.0003282118944476435,
|
| 421 |
+
"loss": 0.3920352554321289,
|
| 422 |
+
"mean_token_accuracy": 0.8694414687156677,
|
| 423 |
+
"num_tokens": 5733805.0,
|
| 424 |
+
"step": 1850
|
| 425 |
+
},
|
| 426 |
+
{
|
| 427 |
+
"entropy": 0.4380896310508251,
|
| 428 |
+
"epoch": 4.439509059029807,
|
| 429 |
+
"grad_norm": 0.5023863315582275,
|
| 430 |
+
"learning_rate": 0.0003193777227622898,
|
| 431 |
+
"loss": 0.3916081237792969,
|
| 432 |
+
"mean_token_accuracy": 0.8697139009833336,
|
| 433 |
+
"num_tokens": 5888694.0,
|
| 434 |
+
"step": 1900
|
| 435 |
+
},
|
| 436 |
+
{
|
| 437 |
+
"entropy": 0.44005885019898416,
|
| 438 |
+
"epoch": 4.556399766218585,
|
| 439 |
+
"grad_norm": 0.455997109413147,
|
| 440 |
+
"learning_rate": 0.000310402233652564,
|
| 441 |
+
"loss": 0.3955466842651367,
|
| 442 |
+
"mean_token_accuracy": 0.8693471103906631,
|
| 443 |
+
"num_tokens": 6042655.0,
|
| 444 |
+
"step": 1950
|
| 445 |
+
},
|
| 446 |
+
{
|
| 447 |
+
"entropy": 0.444269048422575,
|
| 448 |
+
"epoch": 4.673290473407365,
|
| 449 |
+
"grad_norm": 0.4542011320590973,
|
| 450 |
+
"learning_rate": 0.00030130035042769316,
|
| 451 |
+
"loss": 0.3988466262817383,
|
| 452 |
+
"mean_token_accuracy": 0.8680117425322532,
|
| 453 |
+
"num_tokens": 6203980.0,
|
| 454 |
+
"step": 2000
|
| 455 |
+
},
|
| 456 |
+
{
|
| 457 |
+
"entropy": 0.43520899042487143,
|
| 458 |
+
"epoch": 4.790181180596143,
|
| 459 |
+
"grad_norm": 0.49051445722579956,
|
| 460 |
+
"learning_rate": 0.0002920872065490688,
|
| 461 |
+
"loss": 0.39819797515869143,
|
| 462 |
+
"mean_token_accuracy": 0.8678940117359162,
|
| 463 |
+
"num_tokens": 6354604.0,
|
| 464 |
+
"step": 2050
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"entropy": 0.429456724524498,
|
| 468 |
+
"epoch": 4.907071887784921,
|
| 469 |
+
"grad_norm": 0.5267980098724365,
|
| 470 |
+
"learning_rate": 0.0002827781204682396,
|
| 471 |
+
"loss": 0.39413917541503907,
|
| 472 |
+
"mean_token_accuracy": 0.870051506459713,
|
| 473 |
+
"num_tokens": 6507484.0,
|
| 474 |
+
"step": 2100
|
| 475 |
+
},
|
| 476 |
+
{
|
| 477 |
+
"epoch": 5.0,
|
| 478 |
+
"eval_entropy": 0.5826305385146823,
|
| 479 |
+
"eval_loss": 0.7694042921066284,
|
| 480 |
+
"eval_mean_token_accuracy": 0.7950637824349589,
|
| 481 |
+
"eval_num_tokens": 6628215.0,
|
| 482 |
+
"eval_runtime": 38.5641,
|
| 483 |
+
"eval_samples_per_second": 31.921,
|
| 484 |
+
"eval_steps_per_second": 3.993,
|
| 485 |
+
"step": 2140
|
| 486 |
+
},
|
| 487 |
+
{
|
| 488 |
+
"entropy": 0.41423329688496324,
|
| 489 |
+
"epoch": 5.023378141437756,
|
| 490 |
+
"grad_norm": 0.6393507719039917,
|
| 491 |
+
"learning_rate": 0.0002733885701573256,
|
| 492 |
+
"loss": 0.3733938598632813,
|
| 493 |
+
"mean_token_accuracy": 0.8762160557598325,
|
| 494 |
+
"num_tokens": 6662015.0,
|
| 495 |
+
"step": 2150
|
| 496 |
+
},
|
| 497 |
+
{
|
| 498 |
+
"entropy": 0.32116023637354374,
|
| 499 |
+
"epoch": 5.140268848626534,
|
| 500 |
+
"grad_norm": 0.5128306150436401,
|
| 501 |
+
"learning_rate": 0.00026393416737420213,
|
| 502 |
+
"loss": 0.2751211166381836,
|
| 503 |
+
"mean_token_accuracy": 0.9052674892544746,
|
| 504 |
+
"num_tokens": 6817989.0,
|
| 505 |
+
"step": 2200
|
| 506 |
+
},
|
| 507 |
+
{
|
| 508 |
+
"entropy": 0.33186347484588624,
|
| 509 |
+
"epoch": 5.257159555815313,
|
| 510 |
+
"grad_norm": 0.6007080078125,
|
| 511 |
+
"learning_rate": 0.0002544306317052408,
|
| 512 |
+
"loss": 0.2850212097167969,
|
| 513 |
+
"mean_token_accuracy": 0.9016379952430725,
|
| 514 |
+
"num_tokens": 6977021.0,
|
| 515 |
+
"step": 2250
|
| 516 |
+
},
|
| 517 |
+
{
|
| 518 |
+
"entropy": 0.3474555689096451,
|
| 519 |
+
"epoch": 5.374050263004091,
|
| 520 |
+
"grad_norm": 0.6231185793876648,
|
| 521 |
+
"learning_rate": 0.0002448937644287679,
|
| 522 |
+
"loss": 0.29555959701538087,
|
| 523 |
+
"mean_token_accuracy": 0.8966628012061119,
|
| 524 |
+
"num_tokens": 7133360.0,
|
| 525 |
+
"step": 2300
|
| 526 |
+
},
|
| 527 |
+
{
|
| 528 |
+
"entropy": 0.346554354429245,
|
| 529 |
+
"epoch": 5.490940970192869,
|
| 530 |
+
"grad_norm": 0.5146846175193787,
|
| 531 |
+
"learning_rate": 0.0002353394222426952,
|
| 532 |
+
"loss": 0.29524572372436525,
|
| 533 |
+
"mean_token_accuracy": 0.8978548383712769,
|
| 534 |
+
"num_tokens": 7282372.0,
|
| 535 |
+
"step": 2350
|
| 536 |
+
},
|
| 537 |
+
{
|
| 538 |
+
"entropy": 0.3467242659628391,
|
| 539 |
+
"epoch": 5.607831677381649,
|
| 540 |
+
"grad_norm": 0.6118131875991821,
|
| 541 |
+
"learning_rate": 0.00022578349090000624,
|
| 542 |
+
"loss": 0.2970552635192871,
|
| 543 |
+
"mean_token_accuracy": 0.8985717830061912,
|
| 544 |
+
"num_tokens": 7435766.0,
|
| 545 |
+
"step": 2400
|
| 546 |
+
},
|
| 547 |
+
{
|
| 548 |
+
"entropy": 0.34968974225223065,
|
| 549 |
+
"epoch": 5.724722384570427,
|
| 550 |
+
"grad_norm": 0.6346496939659119,
|
| 551 |
+
"learning_rate": 0.0002162418587959333,
|
| 552 |
+
"loss": 0.29940626144409177,
|
| 553 |
+
"mean_token_accuracy": 0.8972864350676537,
|
| 554 |
+
"num_tokens": 7578408.0,
|
| 555 |
+
"step": 2450
|
| 556 |
+
},
|
| 557 |
+
{
|
| 558 |
+
"entropy": 0.35084108062088487,
|
| 559 |
+
"epoch": 5.841613091759205,
|
| 560 |
+
"grad_norm": 0.6566210389137268,
|
| 561 |
+
"learning_rate": 0.00020673039055074148,
|
| 562 |
+
"loss": 0.302426700592041,
|
| 563 |
+
"mean_token_accuracy": 0.8953999072313309,
|
| 564 |
+
"num_tokens": 7734599.0,
|
| 565 |
+
"step": 2500
|
| 566 |
+
},
|
| 567 |
+
{
|
| 568 |
+
"entropy": 0.33895430400967597,
|
| 569 |
+
"epoch": 5.958503798947984,
|
| 570 |
+
"grad_norm": 0.517574667930603,
|
| 571 |
+
"learning_rate": 0.00019726490063204247,
|
| 572 |
+
"loss": 0.292838020324707,
|
| 573 |
+
"mean_token_accuracy": 0.8984868070483207,
|
| 574 |
+
"num_tokens": 7898028.0,
|
| 575 |
+
"step": 2550
|
| 576 |
+
},
|
| 577 |
+
{
|
| 578 |
+
"epoch": 6.0,
|
| 579 |
+
"eval_entropy": 0.5211211553254684,
|
| 580 |
+
"eval_loss": 0.8269560933113098,
|
| 581 |
+
"eval_mean_token_accuracy": 0.7899065927251593,
|
| 582 |
+
"eval_num_tokens": 7953858.0,
|
| 583 |
+
"eval_runtime": 39.369,
|
| 584 |
+
"eval_samples_per_second": 31.268,
|
| 585 |
+
"eval_steps_per_second": 3.912,
|
| 586 |
+
"step": 2568
|
| 587 |
+
}
|
| 588 |
+
],
|
| 589 |
+
"logging_steps": 50,
|
| 590 |
+
"max_steps": 4280,
|
| 591 |
+
"num_input_tokens_seen": 0,
|
| 592 |
+
"num_train_epochs": 10,
|
| 593 |
+
"save_steps": 500,
|
| 594 |
+
"stateful_callbacks": {
|
| 595 |
+
"TrainerControl": {
|
| 596 |
+
"args": {
|
| 597 |
+
"should_epoch_stop": false,
|
| 598 |
+
"should_evaluate": false,
|
| 599 |
+
"should_log": false,
|
| 600 |
+
"should_save": true,
|
| 601 |
+
"should_training_stop": false
|
| 602 |
+
},
|
| 603 |
+
"attributes": {}
|
| 604 |
+
}
|
| 605 |
+
},
|
| 606 |
+
"total_flos": 1.1913679730121984e+17,
|
| 607 |
+
"train_batch_size": 4,
|
| 608 |
+
"trial_name": null,
|
| 609 |
+
"trial_params": null
|
| 610 |
+
}
|
DBCA_original_Estonian/Qwen3.5-2B-Base_original_features_structural_train_original_features_structural_test1/checkpoint-2996/README.md
ADDED
|
@@ -0,0 +1,209 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
base_model: Qwen/Qwen3.5-2B-Base
|
| 3 |
+
library_name: peft
|
| 4 |
+
pipeline_tag: text-generation
|
| 5 |
+
tags:
|
| 6 |
+
- base_model:adapter:Qwen/Qwen3.5-2B-Base
|
| 7 |
+
- lora
|
| 8 |
+
- sft
|
| 9 |
+
- transformers
|
| 10 |
+
- trl
|
| 11 |
+
---
|
| 12 |
+
|
| 13 |
+
# Model Card for Model ID
|
| 14 |
+
|
| 15 |
+
<!-- Provide a quick summary of what the model is/does. -->
|
| 16 |
+
|
| 17 |
+
|
| 18 |
+
|
| 19 |
+
## Model Details
|
| 20 |
+
|
| 21 |
+
### Model Description
|
| 22 |
+
|
| 23 |
+
<!-- Provide a longer summary of what this model is. -->
|
| 24 |
+
|
| 25 |
+
|
| 26 |
+
|
| 27 |
+
- **Developed by:** [More Information Needed]
|
| 28 |
+
- **Funded by [optional]:** [More Information Needed]
|
| 29 |
+
- **Shared by [optional]:** [More Information Needed]
|
| 30 |
+
- **Model type:** [More Information Needed]
|
| 31 |
+
- **Language(s) (NLP):** [More Information Needed]
|
| 32 |
+
- **License:** [More Information Needed]
|
| 33 |
+
- **Finetuned from model [optional]:** [More Information Needed]
|
| 34 |
+
|
| 35 |
+
### Model Sources [optional]
|
| 36 |
+
|
| 37 |
+
<!-- Provide the basic links for the model. -->
|
| 38 |
+
|
| 39 |
+
- **Repository:** [More Information Needed]
|
| 40 |
+
- **Paper [optional]:** [More Information Needed]
|
| 41 |
+
- **Demo [optional]:** [More Information Needed]
|
| 42 |
+
|
| 43 |
+
## Uses
|
| 44 |
+
|
| 45 |
+
<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
|
| 46 |
+
|
| 47 |
+
### Direct Use
|
| 48 |
+
|
| 49 |
+
<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
|
| 50 |
+
|
| 51 |
+
[More Information Needed]
|
| 52 |
+
|
| 53 |
+
### Downstream Use [optional]
|
| 54 |
+
|
| 55 |
+
<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
|
| 56 |
+
|
| 57 |
+
[More Information Needed]
|
| 58 |
+
|
| 59 |
+
### Out-of-Scope Use
|
| 60 |
+
|
| 61 |
+
<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
|
| 62 |
+
|
| 63 |
+
[More Information Needed]
|
| 64 |
+
|
| 65 |
+
## Bias, Risks, and Limitations
|
| 66 |
+
|
| 67 |
+
<!-- This section is meant to convey both technical and sociotechnical limitations. -->
|
| 68 |
+
|
| 69 |
+
[More Information Needed]
|
| 70 |
+
|
| 71 |
+
### Recommendations
|
| 72 |
+
|
| 73 |
+
<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
|
| 74 |
+
|
| 75 |
+
Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
|
| 76 |
+
|
| 77 |
+
## How to Get Started with the Model
|
| 78 |
+
|
| 79 |
+
Use the code below to get started with the model.
|
| 80 |
+
|
| 81 |
+
[More Information Needed]
|
| 82 |
+
|
| 83 |
+
## Training Details
|
| 84 |
+
|
| 85 |
+
### Training Data
|
| 86 |
+
|
| 87 |
+
<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
|
| 88 |
+
|
| 89 |
+
[More Information Needed]
|
| 90 |
+
|
| 91 |
+
### Training Procedure
|
| 92 |
+
|
| 93 |
+
<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
|
| 94 |
+
|
| 95 |
+
#### Preprocessing [optional]
|
| 96 |
+
|
| 97 |
+
[More Information Needed]
|
| 98 |
+
|
| 99 |
+
|
| 100 |
+
#### Training Hyperparameters
|
| 101 |
+
|
| 102 |
+
- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
|
| 103 |
+
|
| 104 |
+
#### Speeds, Sizes, Times [optional]
|
| 105 |
+
|
| 106 |
+
<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
|
| 107 |
+
|
| 108 |
+
[More Information Needed]
|
| 109 |
+
|
| 110 |
+
## Evaluation
|
| 111 |
+
|
| 112 |
+
<!-- This section describes the evaluation protocols and provides the results. -->
|
| 113 |
+
|
| 114 |
+
### Testing Data, Factors & Metrics
|
| 115 |
+
|
| 116 |
+
#### Testing Data
|
| 117 |
+
|
| 118 |
+
<!-- This should link to a Dataset Card if possible. -->
|
| 119 |
+
|
| 120 |
+
[More Information Needed]
|
| 121 |
+
|
| 122 |
+
#### Factors
|
| 123 |
+
|
| 124 |
+
<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
|
| 125 |
+
|
| 126 |
+
[More Information Needed]
|
| 127 |
+
|
| 128 |
+
#### Metrics
|
| 129 |
+
|
| 130 |
+
<!-- These are the evaluation metrics being used, ideally with a description of why. -->
|
| 131 |
+
|
| 132 |
+
[More Information Needed]
|
| 133 |
+
|
| 134 |
+
### Results
|
| 135 |
+
|
| 136 |
+
[More Information Needed]
|
| 137 |
+
|
| 138 |
+
#### Summary
|
| 139 |
+
|
| 140 |
+
|
| 141 |
+
|
| 142 |
+
## Model Examination [optional]
|
| 143 |
+
|
| 144 |
+
<!-- Relevant interpretability work for the model goes here -->
|
| 145 |
+
|
| 146 |
+
[More Information Needed]
|
| 147 |
+
|
| 148 |
+
## Environmental Impact
|
| 149 |
+
|
| 150 |
+
<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
|
| 151 |
+
|
| 152 |
+
Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
|
| 153 |
+
|
| 154 |
+
- **Hardware Type:** [More Information Needed]
|
| 155 |
+
- **Hours used:** [More Information Needed]
|
| 156 |
+
- **Cloud Provider:** [More Information Needed]
|
| 157 |
+
- **Compute Region:** [More Information Needed]
|
| 158 |
+
- **Carbon Emitted:** [More Information Needed]
|
| 159 |
+
|
| 160 |
+
## Technical Specifications [optional]
|
| 161 |
+
|
| 162 |
+
### Model Architecture and Objective
|
| 163 |
+
|
| 164 |
+
[More Information Needed]
|
| 165 |
+
|
| 166 |
+
### Compute Infrastructure
|
| 167 |
+
|
| 168 |
+
[More Information Needed]
|
| 169 |
+
|
| 170 |
+
#### Hardware
|
| 171 |
+
|
| 172 |
+
[More Information Needed]
|
| 173 |
+
|
| 174 |
+
#### Software
|
| 175 |
+
|
| 176 |
+
[More Information Needed]
|
| 177 |
+
|
| 178 |
+
## Citation [optional]
|
| 179 |
+
|
| 180 |
+
<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
|
| 181 |
+
|
| 182 |
+
**BibTeX:**
|
| 183 |
+
|
| 184 |
+
[More Information Needed]
|
| 185 |
+
|
| 186 |
+
**APA:**
|
| 187 |
+
|
| 188 |
+
[More Information Needed]
|
| 189 |
+
|
| 190 |
+
## Glossary [optional]
|
| 191 |
+
|
| 192 |
+
<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
|
| 193 |
+
|
| 194 |
+
[More Information Needed]
|
| 195 |
+
|
| 196 |
+
## More Information [optional]
|
| 197 |
+
|
| 198 |
+
[More Information Needed]
|
| 199 |
+
|
| 200 |
+
## Model Card Authors [optional]
|
| 201 |
+
|
| 202 |
+
[More Information Needed]
|
| 203 |
+
|
| 204 |
+
## Model Card Contact
|
| 205 |
+
|
| 206 |
+
[More Information Needed]
|
| 207 |
+
### Framework versions
|
| 208 |
+
|
| 209 |
+
- PEFT 0.19.1
|
DBCA_original_Estonian/Qwen3.5-2B-Base_original_features_structural_train_original_features_structural_test1/checkpoint-2996/adapter_config.json
ADDED
|
@@ -0,0 +1,48 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"alora_invocation_tokens": null,
|
| 3 |
+
"alpha_pattern": {},
|
| 4 |
+
"arrow_config": null,
|
| 5 |
+
"auto_mapping": null,
|
| 6 |
+
"base_model_name_or_path": "Qwen/Qwen3.5-2B-Base",
|
| 7 |
+
"bias": "none",
|
| 8 |
+
"corda_config": null,
|
| 9 |
+
"ensure_weight_tying": false,
|
| 10 |
+
"eva_config": null,
|
| 11 |
+
"exclude_modules": null,
|
| 12 |
+
"fan_in_fan_out": false,
|
| 13 |
+
"inference_mode": true,
|
| 14 |
+
"init_lora_weights": true,
|
| 15 |
+
"layer_replication": null,
|
| 16 |
+
"layers_pattern": null,
|
| 17 |
+
"layers_to_transform": null,
|
| 18 |
+
"loftq_config": {},
|
| 19 |
+
"lora_alpha": 64,
|
| 20 |
+
"lora_bias": false,
|
| 21 |
+
"lora_dropout": 0.04180832058159435,
|
| 22 |
+
"lora_ga_config": null,
|
| 23 |
+
"megatron_config": null,
|
| 24 |
+
"megatron_core": "megatron.core",
|
| 25 |
+
"modules_to_save": null,
|
| 26 |
+
"peft_type": "LORA",
|
| 27 |
+
"peft_version": "0.19.1",
|
| 28 |
+
"qalora_group_size": 16,
|
| 29 |
+
"r": 64,
|
| 30 |
+
"rank_pattern": {},
|
| 31 |
+
"revision": null,
|
| 32 |
+
"target_modules": [
|
| 33 |
+
"v_proj",
|
| 34 |
+
"o_proj",
|
| 35 |
+
"q_proj",
|
| 36 |
+
"gate_proj",
|
| 37 |
+
"k_proj",
|
| 38 |
+
"down_proj",
|
| 39 |
+
"up_proj"
|
| 40 |
+
],
|
| 41 |
+
"target_parameters": null,
|
| 42 |
+
"task_type": "CAUSAL_LM",
|
| 43 |
+
"trainable_token_indices": null,
|
| 44 |
+
"use_bdlora": null,
|
| 45 |
+
"use_dora": false,
|
| 46 |
+
"use_qalora": false,
|
| 47 |
+
"use_rslora": false
|
| 48 |
+
}
|
DBCA_original_Estonian/Qwen3.5-2B-Base_original_features_structural_train_original_features_structural_test1/checkpoint-2996/chat_template.jinja
ADDED
|
@@ -0,0 +1,154 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{%- set image_count = namespace(value=0) %}
|
| 2 |
+
{%- set video_count = namespace(value=0) %}
|
| 3 |
+
{%- macro render_content(content, do_vision_count, is_system_content=false) %}
|
| 4 |
+
{%- if content is string %}
|
| 5 |
+
{{- content }}
|
| 6 |
+
{%- elif content is iterable and content is not mapping %}
|
| 7 |
+
{%- for item in content %}
|
| 8 |
+
{%- if 'image' in item or 'image_url' in item or item.type == 'image' %}
|
| 9 |
+
{%- if is_system_content %}
|
| 10 |
+
{{- raise_exception('System message cannot contain images.') }}
|
| 11 |
+
{%- endif %}
|
| 12 |
+
{%- if do_vision_count %}
|
| 13 |
+
{%- set image_count.value = image_count.value + 1 %}
|
| 14 |
+
{%- endif %}
|
| 15 |
+
{%- if add_vision_id %}
|
| 16 |
+
{{- 'Picture ' ~ image_count.value ~ ': ' }}
|
| 17 |
+
{%- endif %}
|
| 18 |
+
{{- '<|vision_start|><|image_pad|><|vision_end|>' }}
|
| 19 |
+
{%- elif 'video' in item or item.type == 'video' %}
|
| 20 |
+
{%- if is_system_content %}
|
| 21 |
+
{{- raise_exception('System message cannot contain videos.') }}
|
| 22 |
+
{%- endif %}
|
| 23 |
+
{%- if do_vision_count %}
|
| 24 |
+
{%- set video_count.value = video_count.value + 1 %}
|
| 25 |
+
{%- endif %}
|
| 26 |
+
{%- if add_vision_id %}
|
| 27 |
+
{{- 'Video ' ~ video_count.value ~ ': ' }}
|
| 28 |
+
{%- endif %}
|
| 29 |
+
{{- '<|vision_start|><|video_pad|><|vision_end|>' }}
|
| 30 |
+
{%- elif 'text' in item %}
|
| 31 |
+
{{- item.text }}
|
| 32 |
+
{%- else %}
|
| 33 |
+
{{- raise_exception('Unexpected item type in content.') }}
|
| 34 |
+
{%- endif %}
|
| 35 |
+
{%- endfor %}
|
| 36 |
+
{%- elif content is none or content is undefined %}
|
| 37 |
+
{{- '' }}
|
| 38 |
+
{%- else %}
|
| 39 |
+
{{- raise_exception('Unexpected content type.') }}
|
| 40 |
+
{%- endif %}
|
| 41 |
+
{%- endmacro %}
|
| 42 |
+
{%- if not messages %}
|
| 43 |
+
{{- raise_exception('No messages provided.') }}
|
| 44 |
+
{%- endif %}
|
| 45 |
+
{%- if tools and tools is iterable and tools is not mapping %}
|
| 46 |
+
{{- '<|im_start|>system\n' }}
|
| 47 |
+
{{- "# Tools\n\nYou have access to the following functions:\n\n<tools>" }}
|
| 48 |
+
{%- for tool in tools %}
|
| 49 |
+
{{- "\n" }}
|
| 50 |
+
{{- tool | tojson }}
|
| 51 |
+
{%- endfor %}
|
| 52 |
+
{{- "\n</tools>" }}
|
| 53 |
+
{{- '\n\nIf you choose to call a function ONLY reply in the following format with NO suffix:\n\n<tool_call>\n<function=example_function_name>\n<parameter=example_parameter_1>\nvalue_1\n</parameter>\n<parameter=example_parameter_2>\nThis is the value for the second parameter\nthat can span\nmultiple lines\n</parameter>\n</function>\n</tool_call>\n\n<IMPORTANT>\nReminder:\n- Function calls MUST follow the specified format: an inner <function=...></function> block must be nested within <tool_call></tool_call> XML tags\n- Required parameters MUST be specified\n- You may provide optional reasoning for your function call in natural language BEFORE the function call, but NOT after\n- If there is no function call available, answer the question like normal with your current knowledge and do not tell the user about function calls\n</IMPORTANT>' }}
|
| 54 |
+
{%- if messages[0].role == 'system' %}
|
| 55 |
+
{%- set content = render_content(messages[0].content, false, true)|trim %}
|
| 56 |
+
{%- if content %}
|
| 57 |
+
{{- '\n\n' + content }}
|
| 58 |
+
{%- endif %}
|
| 59 |
+
{%- endif %}
|
| 60 |
+
{{- '<|im_end|>\n' }}
|
| 61 |
+
{%- else %}
|
| 62 |
+
{%- if messages[0].role == 'system' %}
|
| 63 |
+
{%- set content = render_content(messages[0].content, false, true)|trim %}
|
| 64 |
+
{{- '<|im_start|>system\n' + content + '<|im_end|>\n' }}
|
| 65 |
+
{%- endif %}
|
| 66 |
+
{%- endif %}
|
| 67 |
+
{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
|
| 68 |
+
{%- for message in messages[::-1] %}
|
| 69 |
+
{%- set index = (messages|length - 1) - loop.index0 %}
|
| 70 |
+
{%- if ns.multi_step_tool and message.role == "user" %}
|
| 71 |
+
{%- set content = render_content(message.content, false)|trim %}
|
| 72 |
+
{%- if not(content.startswith('<tool_response>') and content.endswith('</tool_response>')) %}
|
| 73 |
+
{%- set ns.multi_step_tool = false %}
|
| 74 |
+
{%- set ns.last_query_index = index %}
|
| 75 |
+
{%- endif %}
|
| 76 |
+
{%- endif %}
|
| 77 |
+
{%- endfor %}
|
| 78 |
+
{%- if ns.multi_step_tool %}
|
| 79 |
+
{{- raise_exception('No user query found in messages.') }}
|
| 80 |
+
{%- endif %}
|
| 81 |
+
{%- for message in messages %}
|
| 82 |
+
{%- set content = render_content(message.content, true)|trim %}
|
| 83 |
+
{%- if message.role == "system" %}
|
| 84 |
+
{%- if not loop.first %}
|
| 85 |
+
{{- raise_exception('System message must be at the beginning.') }}
|
| 86 |
+
{%- endif %}
|
| 87 |
+
{%- elif message.role == "user" %}
|
| 88 |
+
{{- '<|im_start|>' + message.role + '\n' + content + '<|im_end|>' + '\n' }}
|
| 89 |
+
{%- elif message.role == "assistant" %}
|
| 90 |
+
{%- set reasoning_content = '' %}
|
| 91 |
+
{%- if message.reasoning_content is string %}
|
| 92 |
+
{%- set reasoning_content = message.reasoning_content %}
|
| 93 |
+
{%- else %}
|
| 94 |
+
{%- if '</think>' in content %}
|
| 95 |
+
{%- set reasoning_content = content.split('</think>')[0].rstrip('\n').split('<think>')[-1].lstrip('\n') %}
|
| 96 |
+
{%- set content = content.split('</think>')[-1].lstrip('\n') %}
|
| 97 |
+
{%- endif %}
|
| 98 |
+
{%- endif %}
|
| 99 |
+
{%- set reasoning_content = reasoning_content|trim %}
|
| 100 |
+
{%- if loop.index0 > ns.last_query_index %}
|
| 101 |
+
{{- '<|im_start|>' + message.role + '\n<think>\n' + reasoning_content + '\n</think>\n\n' + content }}
|
| 102 |
+
{%- else %}
|
| 103 |
+
{{- '<|im_start|>' + message.role + '\n' + content }}
|
| 104 |
+
{%- endif %}
|
| 105 |
+
{%- if message.tool_calls and message.tool_calls is iterable and message.tool_calls is not mapping %}
|
| 106 |
+
{%- for tool_call in message.tool_calls %}
|
| 107 |
+
{%- if tool_call.function is defined %}
|
| 108 |
+
{%- set tool_call = tool_call.function %}
|
| 109 |
+
{%- endif %}
|
| 110 |
+
{%- if loop.first %}
|
| 111 |
+
{%- if content|trim %}
|
| 112 |
+
{{- '\n\n<tool_call>\n<function=' + tool_call.name + '>\n' }}
|
| 113 |
+
{%- else %}
|
| 114 |
+
{{- '<tool_call>\n<function=' + tool_call.name + '>\n' }}
|
| 115 |
+
{%- endif %}
|
| 116 |
+
{%- else %}
|
| 117 |
+
{{- '\n<tool_call>\n<function=' + tool_call.name + '>\n' }}
|
| 118 |
+
{%- endif %}
|
| 119 |
+
{%- if tool_call.arguments is defined %}
|
| 120 |
+
{%- for args_name, args_value in tool_call.arguments|items %}
|
| 121 |
+
{{- '<parameter=' + args_name + '>\n' }}
|
| 122 |
+
{%- set args_value = args_value | tojson | safe if args_value is mapping or (args_value is sequence and args_value is not string) else args_value | string %}
|
| 123 |
+
{{- args_value }}
|
| 124 |
+
{{- '\n</parameter>\n' }}
|
| 125 |
+
{%- endfor %}
|
| 126 |
+
{%- endif %}
|
| 127 |
+
{{- '</function>\n</tool_call>' }}
|
| 128 |
+
{%- endfor %}
|
| 129 |
+
{%- endif %}
|
| 130 |
+
{{- '<|im_end|>\n' }}
|
| 131 |
+
{%- elif message.role == "tool" %}
|
| 132 |
+
{%- if loop.previtem and loop.previtem.role != "tool" %}
|
| 133 |
+
{{- '<|im_start|>user' }}
|
| 134 |
+
{%- endif %}
|
| 135 |
+
{{- '\n<tool_response>\n' }}
|
| 136 |
+
{{- content }}
|
| 137 |
+
{{- '\n</tool_response>' }}
|
| 138 |
+
{%- if not loop.last and loop.nextitem.role != "tool" %}
|
| 139 |
+
{{- '<|im_end|>\n' }}
|
| 140 |
+
{%- elif loop.last %}
|
| 141 |
+
{{- '<|im_end|>\n' }}
|
| 142 |
+
{%- endif %}
|
| 143 |
+
{%- else %}
|
| 144 |
+
{{- raise_exception('Unexpected message role.') }}
|
| 145 |
+
{%- endif %}
|
| 146 |
+
{%- endfor %}
|
| 147 |
+
{%- if add_generation_prompt %}
|
| 148 |
+
{{- '<|im_start|>assistant\n' }}
|
| 149 |
+
{%- if enable_thinking is defined and enable_thinking is true %}
|
| 150 |
+
{{- '<think>\n' }}
|
| 151 |
+
{%- else %}
|
| 152 |
+
{{- '<think>\n\n</think>\n\n' }}
|
| 153 |
+
{%- endif %}
|
| 154 |
+
{%- endif %}
|
DBCA_original_Estonian/Qwen3.5-2B-Base_original_features_structural_train_original_features_structural_test1/checkpoint-2996/tokenizer_config.json
ADDED
|
@@ -0,0 +1,31 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"add_prefix_space": false,
|
| 3 |
+
"audio_bos_token": "<|audio_start|>",
|
| 4 |
+
"audio_eos_token": "<|audio_end|>",
|
| 5 |
+
"audio_token": "<|audio_pad|>",
|
| 6 |
+
"backend": "tokenizers",
|
| 7 |
+
"bos_token": null,
|
| 8 |
+
"clean_up_tokenization_spaces": false,
|
| 9 |
+
"eos_token": "<|endoftext|>",
|
| 10 |
+
"errors": "replace",
|
| 11 |
+
"image_token": "<|image_pad|>",
|
| 12 |
+
"is_local": false,
|
| 13 |
+
"model_max_length": 262144,
|
| 14 |
+
"model_specific_special_tokens": {
|
| 15 |
+
"audio_bos_token": "<|audio_start|>",
|
| 16 |
+
"audio_eos_token": "<|audio_end|>",
|
| 17 |
+
"audio_token": "<|audio_pad|>",
|
| 18 |
+
"image_token": "<|image_pad|>",
|
| 19 |
+
"video_token": "<|video_pad|>",
|
| 20 |
+
"vision_bos_token": "<|vision_start|>",
|
| 21 |
+
"vision_eos_token": "<|vision_end|>"
|
| 22 |
+
},
|
| 23 |
+
"pad_token": "<|endoftext|>",
|
| 24 |
+
"pretokenize_regex": "(?i:'s|'t|'re|'ve|'m|'ll|'d)|[^\\r\\n\\p{L}\\p{N}]?[\\p{L}\\p{M}]+|\\p{N}| ?[^\\s\\p{L}\\p{M}\\p{N}]+[\\r\\n]*|\\s*[\\r\\n]+|\\s+(?!\\S)|\\s+",
|
| 25 |
+
"split_special_tokens": false,
|
| 26 |
+
"tokenizer_class": "TokenizersBackend",
|
| 27 |
+
"unk_token": null,
|
| 28 |
+
"video_token": "<|video_pad|>",
|
| 29 |
+
"vision_bos_token": "<|vision_start|>",
|
| 30 |
+
"vision_eos_token": "<|vision_end|>"
|
| 31 |
+
}
|
DBCA_original_Estonian/Qwen3.5-2B-Base_original_features_structural_train_original_features_structural_test1/checkpoint-2996/trainer_state.json
ADDED
|
@@ -0,0 +1,701 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": null,
|
| 3 |
+
"best_metric": null,
|
| 4 |
+
"best_model_checkpoint": null,
|
| 5 |
+
"epoch": 7.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 2996,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"entropy": 1.8341716039180755,
|
| 14 |
+
"epoch": 0.11689070718877849,
|
| 15 |
+
"grad_norm": 1.1992322206497192,
|
| 16 |
+
"learning_rate": 5.366728168873013e-05,
|
| 17 |
+
"loss": 1.6918942260742187,
|
| 18 |
+
"mean_token_accuracy": 0.631798365265131,
|
| 19 |
+
"num_tokens": 153689.0,
|
| 20 |
+
"step": 50
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"entropy": 0.9513179230690002,
|
| 24 |
+
"epoch": 0.23378141437755698,
|
| 25 |
+
"grad_norm": 1.1290454864501953,
|
| 26 |
+
"learning_rate": 0.00010842981402416902,
|
| 27 |
+
"loss": 0.8867578887939453,
|
| 28 |
+
"mean_token_accuracy": 0.7587438315153122,
|
| 29 |
+
"num_tokens": 310548.0,
|
| 30 |
+
"step": 100
|
| 31 |
+
},
|
| 32 |
+
{
|
| 33 |
+
"entropy": 0.8510265004634857,
|
| 34 |
+
"epoch": 0.3506721215663355,
|
| 35 |
+
"grad_norm": 0.7732136845588684,
|
| 36 |
+
"learning_rate": 0.00016319234635960792,
|
| 37 |
+
"loss": 0.7856448364257812,
|
| 38 |
+
"mean_token_accuracy": 0.7769897204637527,
|
| 39 |
+
"num_tokens": 468560.0,
|
| 40 |
+
"step": 150
|
| 41 |
+
},
|
| 42 |
+
{
|
| 43 |
+
"entropy": 0.8142367601394653,
|
| 44 |
+
"epoch": 0.46756282875511396,
|
| 45 |
+
"grad_norm": 0.7580806016921997,
|
| 46 |
+
"learning_rate": 0.00021795487869504684,
|
| 47 |
+
"loss": 0.7461666870117187,
|
| 48 |
+
"mean_token_accuracy": 0.7888966089487076,
|
| 49 |
+
"num_tokens": 619552.0,
|
| 50 |
+
"step": 200
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"entropy": 0.772578022480011,
|
| 54 |
+
"epoch": 0.5844535359438925,
|
| 55 |
+
"grad_norm": 0.5330358147621155,
|
| 56 |
+
"learning_rate": 0.00027271741103048575,
|
| 57 |
+
"loss": 0.7142900848388671,
|
| 58 |
+
"mean_token_accuracy": 0.7959607627987861,
|
| 59 |
+
"num_tokens": 771972.0,
|
| 60 |
+
"step": 250
|
| 61 |
+
},
|
| 62 |
+
{
|
| 63 |
+
"entropy": 0.7481484657526016,
|
| 64 |
+
"epoch": 0.701344243132671,
|
| 65 |
+
"grad_norm": 0.8242517709732056,
|
| 66 |
+
"learning_rate": 0.00032747994336592464,
|
| 67 |
+
"loss": 0.6911422729492187,
|
| 68 |
+
"mean_token_accuracy": 0.8028205358982086,
|
| 69 |
+
"num_tokens": 928088.0,
|
| 70 |
+
"step": 300
|
| 71 |
+
},
|
| 72 |
+
{
|
| 73 |
+
"entropy": 0.741372903585434,
|
| 74 |
+
"epoch": 0.8182349503214494,
|
| 75 |
+
"grad_norm": 0.7722981572151184,
|
| 76 |
+
"learning_rate": 0.00038224247570136353,
|
| 77 |
+
"loss": 0.6908904266357422,
|
| 78 |
+
"mean_token_accuracy": 0.8018302822113037,
|
| 79 |
+
"num_tokens": 1086836.0,
|
| 80 |
+
"step": 350
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"entropy": 0.7257227802276611,
|
| 84 |
+
"epoch": 0.9351256575102279,
|
| 85 |
+
"grad_norm": 0.7969573140144348,
|
| 86 |
+
"learning_rate": 0.0004370050080368024,
|
| 87 |
+
"loss": 0.6714310455322265,
|
| 88 |
+
"mean_token_accuracy": 0.8057891410589219,
|
| 89 |
+
"num_tokens": 1241227.0,
|
| 90 |
+
"step": 400
|
| 91 |
+
},
|
| 92 |
+
{
|
| 93 |
+
"epoch": 1.0,
|
| 94 |
+
"eval_entropy": 0.8317563653766334,
|
| 95 |
+
"eval_loss": 0.8135058283805847,
|
| 96 |
+
"eval_mean_token_accuracy": 0.7755894215850087,
|
| 97 |
+
"eval_num_tokens": 1325643.0,
|
| 98 |
+
"eval_runtime": 37.4951,
|
| 99 |
+
"eval_samples_per_second": 32.831,
|
| 100 |
+
"eval_steps_per_second": 4.107,
|
| 101 |
+
"step": 428
|
| 102 |
+
},
|
| 103 |
+
{
|
| 104 |
+
"entropy": 0.7179872094087265,
|
| 105 |
+
"epoch": 1.0514319111630626,
|
| 106 |
+
"grad_norm": 0.6224768161773682,
|
| 107 |
+
"learning_rate": 0.00046873290101955413,
|
| 108 |
+
"loss": 0.6683222961425781,
|
| 109 |
+
"mean_token_accuracy": 0.8054787133207273,
|
| 110 |
+
"num_tokens": 1393922.0,
|
| 111 |
+
"step": 450
|
| 112 |
+
},
|
| 113 |
+
{
|
| 114 |
+
"entropy": 0.6991294291615486,
|
| 115 |
+
"epoch": 1.1683226183518411,
|
| 116 |
+
"grad_norm": 0.6813111901283264,
|
| 117 |
+
"learning_rate": 0.00046837443306086947,
|
| 118 |
+
"loss": 0.6493977355957031,
|
| 119 |
+
"mean_token_accuracy": 0.8089349576830864,
|
| 120 |
+
"num_tokens": 1551442.0,
|
| 121 |
+
"step": 500
|
| 122 |
+
},
|
| 123 |
+
{
|
| 124 |
+
"entropy": 0.6951853144168854,
|
| 125 |
+
"epoch": 1.2852133255406195,
|
| 126 |
+
"grad_norm": 0.617091953754425,
|
| 127 |
+
"learning_rate": 0.0004676269147738558,
|
| 128 |
+
"loss": 0.6484123229980469,
|
| 129 |
+
"mean_token_accuracy": 0.8088837671279907,
|
| 130 |
+
"num_tokens": 1706697.0,
|
| 131 |
+
"step": 550
|
| 132 |
+
},
|
| 133 |
+
{
|
| 134 |
+
"entropy": 0.6852494943141937,
|
| 135 |
+
"epoch": 1.4021040327293979,
|
| 136 |
+
"grad_norm": 0.4677598774433136,
|
| 137 |
+
"learning_rate": 0.0004664915890374708,
|
| 138 |
+
"loss": 0.6416233062744141,
|
| 139 |
+
"mean_token_accuracy": 0.8106312158703805,
|
| 140 |
+
"num_tokens": 1865246.0,
|
| 141 |
+
"step": 600
|
| 142 |
+
},
|
| 143 |
+
{
|
| 144 |
+
"entropy": 0.6703594943881035,
|
| 145 |
+
"epoch": 1.5189947399181765,
|
| 146 |
+
"grad_norm": 0.5708025693893433,
|
| 147 |
+
"learning_rate": 0.0004649703435278991,
|
| 148 |
+
"loss": 0.6273183441162109,
|
| 149 |
+
"mean_token_accuracy": 0.8155005398392677,
|
| 150 |
+
"num_tokens": 2022060.0,
|
| 151 |
+
"step": 650
|
| 152 |
+
},
|
| 153 |
+
{
|
| 154 |
+
"entropy": 0.6676286320388317,
|
| 155 |
+
"epoch": 1.635885447106955,
|
| 156 |
+
"grad_norm": 0.5555347800254822,
|
| 157 |
+
"learning_rate": 0.00046306570757996264,
|
| 158 |
+
"loss": 0.6276700973510743,
|
| 159 |
+
"mean_token_accuracy": 0.8154828292131424,
|
| 160 |
+
"num_tokens": 2178909.0,
|
| 161 |
+
"step": 700
|
| 162 |
+
},
|
| 163 |
+
{
|
| 164 |
+
"entropy": 0.6669242936372757,
|
| 165 |
+
"epoch": 1.7527761542957334,
|
| 166 |
+
"grad_norm": 0.5260112285614014,
|
| 167 |
+
"learning_rate": 0.0004607808479816624,
|
| 168 |
+
"loss": 0.6228141784667969,
|
| 169 |
+
"mean_token_accuracy": 0.8164840793609619,
|
| 170 |
+
"num_tokens": 2329314.0,
|
| 171 |
+
"step": 750
|
| 172 |
+
},
|
| 173 |
+
{
|
| 174 |
+
"entropy": 0.6598560312390327,
|
| 175 |
+
"epoch": 1.869666861484512,
|
| 176 |
+
"grad_norm": 0.5677736401557922,
|
| 177 |
+
"learning_rate": 0.0004581195637088436,
|
| 178 |
+
"loss": 0.6214292907714843,
|
| 179 |
+
"mean_token_accuracy": 0.8171795177459716,
|
| 180 |
+
"num_tokens": 2478559.0,
|
| 181 |
+
"step": 800
|
| 182 |
+
},
|
| 183 |
+
{
|
| 184 |
+
"entropy": 0.6687870016694069,
|
| 185 |
+
"epoch": 1.9865575686732906,
|
| 186 |
+
"grad_norm": 0.6449615955352783,
|
| 187 |
+
"learning_rate": 0.00045508627960873823,
|
| 188 |
+
"loss": 0.6243909454345703,
|
| 189 |
+
"mean_token_accuracy": 0.8147780740261078,
|
| 190 |
+
"num_tokens": 2631937.0,
|
| 191 |
+
"step": 850
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"epoch": 2.0,
|
| 195 |
+
"eval_entropy": 0.7786144471013701,
|
| 196 |
+
"eval_loss": 0.7599140405654907,
|
| 197 |
+
"eval_mean_token_accuracy": 0.7935443106409791,
|
| 198 |
+
"eval_num_tokens": 2651286.0,
|
| 199 |
+
"eval_runtime": 39.0048,
|
| 200 |
+
"eval_samples_per_second": 31.56,
|
| 201 |
+
"eval_steps_per_second": 3.948,
|
| 202 |
+
"step": 856
|
| 203 |
+
},
|
| 204 |
+
{
|
| 205 |
+
"entropy": 0.5821921329701966,
|
| 206 |
+
"epoch": 2.102863822326125,
|
| 207 |
+
"grad_norm": 0.6116960048675537,
|
| 208 |
+
"learning_rate": 0.00045168603904288863,
|
| 209 |
+
"loss": 0.5406004714965821,
|
| 210 |
+
"mean_token_accuracy": 0.8334921918921734,
|
| 211 |
+
"num_tokens": 2795200.0,
|
| 212 |
+
"step": 900
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"entropy": 0.5937331764400006,
|
| 216 |
+
"epoch": 2.2197545295149035,
|
| 217 |
+
"grad_norm": 0.5440122485160828,
|
| 218 |
+
"learning_rate": 0.00044792449550168286,
|
| 219 |
+
"loss": 0.5515737533569336,
|
| 220 |
+
"mean_token_accuracy": 0.8314691257476806,
|
| 221 |
+
"num_tokens": 2950534.0,
|
| 222 |
+
"step": 950
|
| 223 |
+
},
|
| 224 |
+
{
|
| 225 |
+
"entropy": 0.5877273553609847,
|
| 226 |
+
"epoch": 2.3366452367036823,
|
| 227 |
+
"grad_norm": 0.6505260467529297,
|
| 228 |
+
"learning_rate": 0.0004438079032044453,
|
| 229 |
+
"loss": 0.5507744979858399,
|
| 230 |
+
"mean_token_accuracy": 0.8314063146710395,
|
| 231 |
+
"num_tokens": 3101471.0,
|
| 232 |
+
"step": 1000
|
| 233 |
+
},
|
| 234 |
+
{
|
| 235 |
+
"entropy": 0.5897100016474723,
|
| 236 |
+
"epoch": 2.4535359438924607,
|
| 237 |
+
"grad_norm": 0.54567551612854,
|
| 238 |
+
"learning_rate": 0.0004393431067007111,
|
| 239 |
+
"loss": 0.5528768157958984,
|
| 240 |
+
"mean_token_accuracy": 0.8315245220065117,
|
| 241 |
+
"num_tokens": 3258554.0,
|
| 242 |
+
"step": 1050
|
| 243 |
+
},
|
| 244 |
+
{
|
| 245 |
+
"entropy": 0.5831590622663498,
|
| 246 |
+
"epoch": 2.570426651081239,
|
| 247 |
+
"grad_norm": 0.44826796650886536,
|
| 248 |
+
"learning_rate": 0.00043453752948997376,
|
| 249 |
+
"loss": 0.5442767333984375,
|
| 250 |
+
"mean_token_accuracy": 0.8327079233527184,
|
| 251 |
+
"num_tokens": 3416088.0,
|
| 252 |
+
"step": 1100
|
| 253 |
+
},
|
| 254 |
+
{
|
| 255 |
+
"entropy": 0.5839906217157841,
|
| 256 |
+
"epoch": 2.6873173582700174,
|
| 257 |
+
"grad_norm": 0.5067696571350098,
|
| 258 |
+
"learning_rate": 0.0004293991616788285,
|
| 259 |
+
"loss": 0.5485663604736328,
|
| 260 |
+
"mean_token_accuracy": 0.8324281191825866,
|
| 261 |
+
"num_tokens": 3567251.0,
|
| 262 |
+
"step": 1150
|
| 263 |
+
},
|
| 264 |
+
{
|
| 265 |
+
"entropy": 0.5870274990797043,
|
| 266 |
+
"epoch": 2.8042080654587958,
|
| 267 |
+
"grad_norm": 0.44032806158065796,
|
| 268 |
+
"learning_rate": 0.00042393654669603217,
|
| 269 |
+
"loss": 0.5524474716186524,
|
| 270 |
+
"mean_token_accuracy": 0.8313613015413285,
|
| 271 |
+
"num_tokens": 3719464.0,
|
| 272 |
+
"step": 1200
|
| 273 |
+
},
|
| 274 |
+
{
|
| 275 |
+
"entropy": 0.5864466108381748,
|
| 276 |
+
"epoch": 2.9210987726475746,
|
| 277 |
+
"grad_norm": 0.5520864129066467,
|
| 278 |
+
"learning_rate": 0.00041815876708756964,
|
| 279 |
+
"loss": 0.5490786361694336,
|
| 280 |
+
"mean_token_accuracy": 0.8314764249324799,
|
| 281 |
+
"num_tokens": 3874423.0,
|
| 282 |
+
"step": 1250
|
| 283 |
+
},
|
| 284 |
+
{
|
| 285 |
+
"epoch": 3.0,
|
| 286 |
+
"eval_entropy": 0.6819627613990338,
|
| 287 |
+
"eval_loss": 0.7593190670013428,
|
| 288 |
+
"eval_mean_token_accuracy": 0.7857603395914102,
|
| 289 |
+
"eval_num_tokens": 3976929.0,
|
| 290 |
+
"eval_runtime": 39.3654,
|
| 291 |
+
"eval_samples_per_second": 31.271,
|
| 292 |
+
"eval_steps_per_second": 3.912,
|
| 293 |
+
"step": 1284
|
| 294 |
+
},
|
| 295 |
+
{
|
| 296 |
+
"entropy": 0.5587641523411525,
|
| 297 |
+
"epoch": 3.037405026300409,
|
| 298 |
+
"grad_norm": 0.622401773929596,
|
| 299 |
+
"learning_rate": 0.0004120754294153441,
|
| 300 |
+
"loss": 0.518932762145996,
|
| 301 |
+
"mean_token_accuracy": 0.8399260457436643,
|
| 302 |
+
"num_tokens": 4027609.0,
|
| 303 |
+
"step": 1300
|
| 304 |
+
},
|
| 305 |
+
{
|
| 306 |
+
"entropy": 0.5028628017008304,
|
| 307 |
+
"epoch": 3.1542957334891875,
|
| 308 |
+
"grad_norm": 0.577396810054779,
|
| 309 |
+
"learning_rate": 0.00040569664828459917,
|
| 310 |
+
"loss": 0.4589382171630859,
|
| 311 |
+
"mean_token_accuracy": 0.8538418188691139,
|
| 312 |
+
"num_tokens": 4180560.0,
|
| 313 |
+
"step": 1350
|
| 314 |
+
},
|
| 315 |
+
{
|
| 316 |
+
"entropy": 0.5227407096326351,
|
| 317 |
+
"epoch": 3.2711864406779663,
|
| 318 |
+
"grad_norm": 0.5113071203231812,
|
| 319 |
+
"learning_rate": 0.00039903302952663176,
|
| 320 |
+
"loss": 0.47863777160644533,
|
| 321 |
+
"mean_token_accuracy": 0.847364938557148,
|
| 322 |
+
"num_tokens": 4329954.0,
|
| 323 |
+
"step": 1400
|
| 324 |
+
},
|
| 325 |
+
{
|
| 326 |
+
"entropy": 0.5050077450275421,
|
| 327 |
+
"epoch": 3.3880771478667446,
|
| 328 |
+
"grad_norm": 0.48977184295654297,
|
| 329 |
+
"learning_rate": 0.0003920956525647558,
|
| 330 |
+
"loss": 0.46841480255126955,
|
| 331 |
+
"mean_token_accuracy": 0.8519425508379936,
|
| 332 |
+
"num_tokens": 4484354.0,
|
| 333 |
+
"step": 1450
|
| 334 |
+
},
|
| 335 |
+
{
|
| 336 |
+
"entropy": 0.5163291451334954,
|
| 337 |
+
"epoch": 3.504967855055523,
|
| 338 |
+
"grad_norm": 0.509864330291748,
|
| 339 |
+
"learning_rate": 0.000384896051992837,
|
| 340 |
+
"loss": 0.47587432861328127,
|
| 341 |
+
"mean_token_accuracy": 0.8486990982294083,
|
| 342 |
+
"num_tokens": 4638919.0,
|
| 343 |
+
"step": 1500
|
| 344 |
+
},
|
| 345 |
+
{
|
| 346 |
+
"entropy": 0.5134807989001274,
|
| 347 |
+
"epoch": 3.6218585622443014,
|
| 348 |
+
"grad_norm": 0.46544620394706726,
|
| 349 |
+
"learning_rate": 0.00037744619839702735,
|
| 350 |
+
"loss": 0.47692710876464844,
|
| 351 |
+
"mean_token_accuracy": 0.8489899519085884,
|
| 352 |
+
"num_tokens": 4795540.0,
|
| 353 |
+
"step": 1550
|
| 354 |
+
},
|
| 355 |
+
{
|
| 356 |
+
"entropy": 0.5216251534223556,
|
| 357 |
+
"epoch": 3.73874926943308,
|
| 358 |
+
"grad_norm": 0.5167004466056824,
|
| 359 |
+
"learning_rate": 0.0003697584784525874,
|
| 360 |
+
"loss": 0.4837848663330078,
|
| 361 |
+
"mean_token_accuracy": 0.847066233754158,
|
| 362 |
+
"num_tokens": 4951177.0,
|
| 363 |
+
"step": 1600
|
| 364 |
+
},
|
| 365 |
+
{
|
| 366 |
+
"entropy": 0.5254101701080799,
|
| 367 |
+
"epoch": 3.8556399766218585,
|
| 368 |
+
"grad_norm": 0.5504006743431091,
|
| 369 |
+
"learning_rate": 0.00036184567432888745,
|
| 370 |
+
"loss": 0.48759506225585936,
|
| 371 |
+
"mean_token_accuracy": 0.8457375919818878,
|
| 372 |
+
"num_tokens": 5104814.0,
|
| 373 |
+
"step": 1650
|
| 374 |
+
},
|
| 375 |
+
{
|
| 376 |
+
"entropy": 0.5107753933966159,
|
| 377 |
+
"epoch": 3.972530683810637,
|
| 378 |
+
"grad_norm": 0.421975314617157,
|
| 379 |
+
"learning_rate": 0.0003537209424368311,
|
| 380 |
+
"loss": 0.4743759536743164,
|
| 381 |
+
"mean_token_accuracy": 0.8505110186338425,
|
| 382 |
+
"num_tokens": 5265185.0,
|
| 383 |
+
"step": 1700
|
| 384 |
+
},
|
| 385 |
+
{
|
| 386 |
+
"epoch": 4.0,
|
| 387 |
+
"eval_entropy": 0.665543915002377,
|
| 388 |
+
"eval_loss": 0.7487082481384277,
|
| 389 |
+
"eval_mean_token_accuracy": 0.7999678904359991,
|
| 390 |
+
"eval_num_tokens": 5302572.0,
|
| 391 |
+
"eval_runtime": 38.5439,
|
| 392 |
+
"eval_samples_per_second": 31.938,
|
| 393 |
+
"eval_steps_per_second": 3.995,
|
| 394 |
+
"step": 1712
|
| 395 |
+
},
|
| 396 |
+
{
|
| 397 |
+
"entropy": 0.4414465431891494,
|
| 398 |
+
"epoch": 4.088836937463472,
|
| 399 |
+
"grad_norm": 0.5308493971824646,
|
| 400 |
+
"learning_rate": 0.0003453977915540383,
|
| 401 |
+
"loss": 0.39650299072265627,
|
| 402 |
+
"mean_token_accuracy": 0.8703725772287378,
|
| 403 |
+
"num_tokens": 5419002.0,
|
| 404 |
+
"step": 1750
|
| 405 |
+
},
|
| 406 |
+
{
|
| 407 |
+
"entropy": 0.42591195791959763,
|
| 408 |
+
"epoch": 4.20572764465225,
|
| 409 |
+
"grad_norm": 0.6111563444137573,
|
| 410 |
+
"learning_rate": 0.00033689006036415585,
|
| 411 |
+
"loss": 0.37838283538818357,
|
| 412 |
+
"mean_token_accuracy": 0.874235480427742,
|
| 413 |
+
"num_tokens": 5572969.0,
|
| 414 |
+
"step": 1800
|
| 415 |
+
},
|
| 416 |
+
{
|
| 417 |
+
"entropy": 0.44086351931095125,
|
| 418 |
+
"epoch": 4.322618351841029,
|
| 419 |
+
"grad_norm": 0.5524880290031433,
|
| 420 |
+
"learning_rate": 0.0003282118944476435,
|
| 421 |
+
"loss": 0.3920352554321289,
|
| 422 |
+
"mean_token_accuracy": 0.8694414687156677,
|
| 423 |
+
"num_tokens": 5733805.0,
|
| 424 |
+
"step": 1850
|
| 425 |
+
},
|
| 426 |
+
{
|
| 427 |
+
"entropy": 0.4380896310508251,
|
| 428 |
+
"epoch": 4.439509059029807,
|
| 429 |
+
"grad_norm": 0.5023863315582275,
|
| 430 |
+
"learning_rate": 0.0003193777227622898,
|
| 431 |
+
"loss": 0.3916081237792969,
|
| 432 |
+
"mean_token_accuracy": 0.8697139009833336,
|
| 433 |
+
"num_tokens": 5888694.0,
|
| 434 |
+
"step": 1900
|
| 435 |
+
},
|
| 436 |
+
{
|
| 437 |
+
"entropy": 0.44005885019898416,
|
| 438 |
+
"epoch": 4.556399766218585,
|
| 439 |
+
"grad_norm": 0.455997109413147,
|
| 440 |
+
"learning_rate": 0.000310402233652564,
|
| 441 |
+
"loss": 0.3955466842651367,
|
| 442 |
+
"mean_token_accuracy": 0.8693471103906631,
|
| 443 |
+
"num_tokens": 6042655.0,
|
| 444 |
+
"step": 1950
|
| 445 |
+
},
|
| 446 |
+
{
|
| 447 |
+
"entropy": 0.444269048422575,
|
| 448 |
+
"epoch": 4.673290473407365,
|
| 449 |
+
"grad_norm": 0.4542011320590973,
|
| 450 |
+
"learning_rate": 0.00030130035042769316,
|
| 451 |
+
"loss": 0.3988466262817383,
|
| 452 |
+
"mean_token_accuracy": 0.8680117425322532,
|
| 453 |
+
"num_tokens": 6203980.0,
|
| 454 |
+
"step": 2000
|
| 455 |
+
},
|
| 456 |
+
{
|
| 457 |
+
"entropy": 0.43520899042487143,
|
| 458 |
+
"epoch": 4.790181180596143,
|
| 459 |
+
"grad_norm": 0.49051445722579956,
|
| 460 |
+
"learning_rate": 0.0002920872065490688,
|
| 461 |
+
"loss": 0.39819797515869143,
|
| 462 |
+
"mean_token_accuracy": 0.8678940117359162,
|
| 463 |
+
"num_tokens": 6354604.0,
|
| 464 |
+
"step": 2050
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"entropy": 0.429456724524498,
|
| 468 |
+
"epoch": 4.907071887784921,
|
| 469 |
+
"grad_norm": 0.5267980098724365,
|
| 470 |
+
"learning_rate": 0.0002827781204682396,
|
| 471 |
+
"loss": 0.39413917541503907,
|
| 472 |
+
"mean_token_accuracy": 0.870051506459713,
|
| 473 |
+
"num_tokens": 6507484.0,
|
| 474 |
+
"step": 2100
|
| 475 |
+
},
|
| 476 |
+
{
|
| 477 |
+
"epoch": 5.0,
|
| 478 |
+
"eval_entropy": 0.5826305385146823,
|
| 479 |
+
"eval_loss": 0.7694042921066284,
|
| 480 |
+
"eval_mean_token_accuracy": 0.7950637824349589,
|
| 481 |
+
"eval_num_tokens": 6628215.0,
|
| 482 |
+
"eval_runtime": 38.5641,
|
| 483 |
+
"eval_samples_per_second": 31.921,
|
| 484 |
+
"eval_steps_per_second": 3.993,
|
| 485 |
+
"step": 2140
|
| 486 |
+
},
|
| 487 |
+
{
|
| 488 |
+
"entropy": 0.41423329688496324,
|
| 489 |
+
"epoch": 5.023378141437756,
|
| 490 |
+
"grad_norm": 0.6393507719039917,
|
| 491 |
+
"learning_rate": 0.0002733885701573256,
|
| 492 |
+
"loss": 0.3733938598632813,
|
| 493 |
+
"mean_token_accuracy": 0.8762160557598325,
|
| 494 |
+
"num_tokens": 6662015.0,
|
| 495 |
+
"step": 2150
|
| 496 |
+
},
|
| 497 |
+
{
|
| 498 |
+
"entropy": 0.32116023637354374,
|
| 499 |
+
"epoch": 5.140268848626534,
|
| 500 |
+
"grad_norm": 0.5128306150436401,
|
| 501 |
+
"learning_rate": 0.00026393416737420213,
|
| 502 |
+
"loss": 0.2751211166381836,
|
| 503 |
+
"mean_token_accuracy": 0.9052674892544746,
|
| 504 |
+
"num_tokens": 6817989.0,
|
| 505 |
+
"step": 2200
|
| 506 |
+
},
|
| 507 |
+
{
|
| 508 |
+
"entropy": 0.33186347484588624,
|
| 509 |
+
"epoch": 5.257159555815313,
|
| 510 |
+
"grad_norm": 0.6007080078125,
|
| 511 |
+
"learning_rate": 0.0002544306317052408,
|
| 512 |
+
"loss": 0.2850212097167969,
|
| 513 |
+
"mean_token_accuracy": 0.9016379952430725,
|
| 514 |
+
"num_tokens": 6977021.0,
|
| 515 |
+
"step": 2250
|
| 516 |
+
},
|
| 517 |
+
{
|
| 518 |
+
"entropy": 0.3474555689096451,
|
| 519 |
+
"epoch": 5.374050263004091,
|
| 520 |
+
"grad_norm": 0.6231185793876648,
|
| 521 |
+
"learning_rate": 0.0002448937644287679,
|
| 522 |
+
"loss": 0.29555959701538087,
|
| 523 |
+
"mean_token_accuracy": 0.8966628012061119,
|
| 524 |
+
"num_tokens": 7133360.0,
|
| 525 |
+
"step": 2300
|
| 526 |
+
},
|
| 527 |
+
{
|
| 528 |
+
"entropy": 0.346554354429245,
|
| 529 |
+
"epoch": 5.490940970192869,
|
| 530 |
+
"grad_norm": 0.5146846175193787,
|
| 531 |
+
"learning_rate": 0.0002353394222426952,
|
| 532 |
+
"loss": 0.29524572372436525,
|
| 533 |
+
"mean_token_accuracy": 0.8978548383712769,
|
| 534 |
+
"num_tokens": 7282372.0,
|
| 535 |
+
"step": 2350
|
| 536 |
+
},
|
| 537 |
+
{
|
| 538 |
+
"entropy": 0.3467242659628391,
|
| 539 |
+
"epoch": 5.607831677381649,
|
| 540 |
+
"grad_norm": 0.6118131875991821,
|
| 541 |
+
"learning_rate": 0.00022578349090000624,
|
| 542 |
+
"loss": 0.2970552635192871,
|
| 543 |
+
"mean_token_accuracy": 0.8985717830061912,
|
| 544 |
+
"num_tokens": 7435766.0,
|
| 545 |
+
"step": 2400
|
| 546 |
+
},
|
| 547 |
+
{
|
| 548 |
+
"entropy": 0.34968974225223065,
|
| 549 |
+
"epoch": 5.724722384570427,
|
| 550 |
+
"grad_norm": 0.6346496939659119,
|
| 551 |
+
"learning_rate": 0.0002162418587959333,
|
| 552 |
+
"loss": 0.29940626144409177,
|
| 553 |
+
"mean_token_accuracy": 0.8972864350676537,
|
| 554 |
+
"num_tokens": 7578408.0,
|
| 555 |
+
"step": 2450
|
| 556 |
+
},
|
| 557 |
+
{
|
| 558 |
+
"entropy": 0.35084108062088487,
|
| 559 |
+
"epoch": 5.841613091759205,
|
| 560 |
+
"grad_norm": 0.6566210389137268,
|
| 561 |
+
"learning_rate": 0.00020673039055074148,
|
| 562 |
+
"loss": 0.302426700592041,
|
| 563 |
+
"mean_token_accuracy": 0.8953999072313309,
|
| 564 |
+
"num_tokens": 7734599.0,
|
| 565 |
+
"step": 2500
|
| 566 |
+
},
|
| 567 |
+
{
|
| 568 |
+
"entropy": 0.33895430400967597,
|
| 569 |
+
"epoch": 5.958503798947984,
|
| 570 |
+
"grad_norm": 0.517574667930603,
|
| 571 |
+
"learning_rate": 0.00019726490063204247,
|
| 572 |
+
"loss": 0.292838020324707,
|
| 573 |
+
"mean_token_accuracy": 0.8984868070483207,
|
| 574 |
+
"num_tokens": 7898028.0,
|
| 575 |
+
"step": 2550
|
| 576 |
+
},
|
| 577 |
+
{
|
| 578 |
+
"epoch": 6.0,
|
| 579 |
+
"eval_entropy": 0.5211211553254684,
|
| 580 |
+
"eval_loss": 0.8269560933113098,
|
| 581 |
+
"eval_mean_token_accuracy": 0.7899065927251593,
|
| 582 |
+
"eval_num_tokens": 7953858.0,
|
| 583 |
+
"eval_runtime": 39.369,
|
| 584 |
+
"eval_samples_per_second": 31.268,
|
| 585 |
+
"eval_steps_per_second": 3.912,
|
| 586 |
+
"step": 2568
|
| 587 |
+
},
|
| 588 |
+
{
|
| 589 |
+
"entropy": 0.2763742347009218,
|
| 590 |
+
"epoch": 6.074810052600818,
|
| 591 |
+
"grad_norm": 0.5116816163063049,
|
| 592 |
+
"learning_rate": 0.00018786112706049623,
|
| 593 |
+
"loss": 0.22524795532226563,
|
| 594 |
+
"mean_token_accuracy": 0.9218554844209297,
|
| 595 |
+
"num_tokens": 8052646.0,
|
| 596 |
+
"step": 2600
|
| 597 |
+
},
|
| 598 |
+
{
|
| 599 |
+
"entropy": 0.23341606348752975,
|
| 600 |
+
"epoch": 6.1917007597895966,
|
| 601 |
+
"grad_norm": 0.6335962414741516,
|
| 602 |
+
"learning_rate": 0.00017853470524261842,
|
| 603 |
+
"loss": 0.18153846740722657,
|
| 604 |
+
"mean_token_accuracy": 0.935965863764286,
|
| 605 |
+
"num_tokens": 8212323.0,
|
| 606 |
+
"step": 2650
|
| 607 |
+
},
|
| 608 |
+
{
|
| 609 |
+
"entropy": 0.23515729174017908,
|
| 610 |
+
"epoch": 6.308591466978375,
|
| 611 |
+
"grad_norm": 0.6312636733055115,
|
| 612 |
+
"learning_rate": 0.0001693011419742025,
|
| 613 |
+
"loss": 0.18562854766845704,
|
| 614 |
+
"mean_token_accuracy": 0.9342144966125489,
|
| 615 |
+
"num_tokens": 8366715.0,
|
| 616 |
+
"step": 2700
|
| 617 |
+
},
|
| 618 |
+
{
|
| 619 |
+
"entropy": 0.23566059060394765,
|
| 620 |
+
"epoch": 6.425482174167154,
|
| 621 |
+
"grad_norm": 0.5869485139846802,
|
| 622 |
+
"learning_rate": 0.0001601757896575784,
|
| 623 |
+
"loss": 0.1867989158630371,
|
| 624 |
+
"mean_token_accuracy": 0.934665755033493,
|
| 625 |
+
"num_tokens": 8515955.0,
|
| 626 |
+
"step": 2750
|
| 627 |
+
},
|
| 628 |
+
{
|
| 629 |
+
"entropy": 0.2446490554511547,
|
| 630 |
+
"epoch": 6.5423728813559325,
|
| 631 |
+
"grad_norm": 0.5957017540931702,
|
| 632 |
+
"learning_rate": 0.00015117382077557817,
|
| 633 |
+
"loss": 0.19312274932861329,
|
| 634 |
+
"mean_token_accuracy": 0.9308532625436783,
|
| 635 |
+
"num_tokens": 8672361.0,
|
| 636 |
+
"step": 2800
|
| 637 |
+
},
|
| 638 |
+
{
|
| 639 |
+
"entropy": 0.24066772796213626,
|
| 640 |
+
"epoch": 6.659263588544711,
|
| 641 |
+
"grad_norm": 0.6069761514663696,
|
| 642 |
+
"learning_rate": 0.0001423102026646483,
|
| 643 |
+
"loss": 0.19540096282958985,
|
| 644 |
+
"mean_token_accuracy": 0.9300534284114838,
|
| 645 |
+
"num_tokens": 8827819.0,
|
| 646 |
+
"step": 2850
|
| 647 |
+
},
|
| 648 |
+
{
|
| 649 |
+
"entropy": 0.2378876845538616,
|
| 650 |
+
"epoch": 6.776154295733489,
|
| 651 |
+
"grad_norm": 0.6069548726081848,
|
| 652 |
+
"learning_rate": 0.00013359967262905405,
|
| 653 |
+
"loss": 0.1942250633239746,
|
| 654 |
+
"mean_token_accuracy": 0.9310428243875504,
|
| 655 |
+
"num_tokens": 8979535.0,
|
| 656 |
+
"step": 2900
|
| 657 |
+
},
|
| 658 |
+
{
|
| 659 |
+
"entropy": 0.23810458809137344,
|
| 660 |
+
"epoch": 6.893045002922268,
|
| 661 |
+
"grad_norm": 0.5807433128356934,
|
| 662 |
+
"learning_rate": 0.00012505671343755173,
|
| 663 |
+
"loss": 0.1908321762084961,
|
| 664 |
+
"mean_token_accuracy": 0.9318096882104874,
|
| 665 |
+
"num_tokens": 9134134.0,
|
| 666 |
+
"step": 2950
|
| 667 |
+
},
|
| 668 |
+
{
|
| 669 |
+
"epoch": 7.0,
|
| 670 |
+
"eval_entropy": 0.4395190301266583,
|
| 671 |
+
"eval_loss": 0.9185124635696411,
|
| 672 |
+
"eval_mean_token_accuracy": 0.7936263409527865,
|
| 673 |
+
"eval_num_tokens": 9279501.0,
|
| 674 |
+
"eval_runtime": 37.9924,
|
| 675 |
+
"eval_samples_per_second": 32.401,
|
| 676 |
+
"eval_steps_per_second": 4.053,
|
| 677 |
+
"step": 2996
|
| 678 |
+
}
|
| 679 |
+
],
|
| 680 |
+
"logging_steps": 50,
|
| 681 |
+
"max_steps": 4280,
|
| 682 |
+
"num_input_tokens_seen": 0,
|
| 683 |
+
"num_train_epochs": 10,
|
| 684 |
+
"save_steps": 500,
|
| 685 |
+
"stateful_callbacks": {
|
| 686 |
+
"TrainerControl": {
|
| 687 |
+
"args": {
|
| 688 |
+
"should_epoch_stop": false,
|
| 689 |
+
"should_evaluate": false,
|
| 690 |
+
"should_log": false,
|
| 691 |
+
"should_save": true,
|
| 692 |
+
"should_training_stop": false
|
| 693 |
+
},
|
| 694 |
+
"attributes": {}
|
| 695 |
+
}
|
| 696 |
+
},
|
| 697 |
+
"total_flos": 1.3912420998060749e+17,
|
| 698 |
+
"train_batch_size": 4,
|
| 699 |
+
"trial_name": null,
|
| 700 |
+
"trial_params": null
|
| 701 |
+
}
|
DBCA_original_Estonian/Qwen3.5-2B-Base_original_features_structural_train_original_features_structural_test1/checkpoint-3424/README.md
ADDED
|
@@ -0,0 +1,209 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
base_model: Qwen/Qwen3.5-2B-Base
|
| 3 |
+
library_name: peft
|
| 4 |
+
pipeline_tag: text-generation
|
| 5 |
+
tags:
|
| 6 |
+
- base_model:adapter:Qwen/Qwen3.5-2B-Base
|
| 7 |
+
- lora
|
| 8 |
+
- sft
|
| 9 |
+
- transformers
|
| 10 |
+
- trl
|
| 11 |
+
---
|
| 12 |
+
|
| 13 |
+
# Model Card for Model ID
|
| 14 |
+
|
| 15 |
+
<!-- Provide a quick summary of what the model is/does. -->
|
| 16 |
+
|
| 17 |
+
|
| 18 |
+
|
| 19 |
+
## Model Details
|
| 20 |
+
|
| 21 |
+
### Model Description
|
| 22 |
+
|
| 23 |
+
<!-- Provide a longer summary of what this model is. -->
|
| 24 |
+
|
| 25 |
+
|
| 26 |
+
|
| 27 |
+
- **Developed by:** [More Information Needed]
|
| 28 |
+
- **Funded by [optional]:** [More Information Needed]
|
| 29 |
+
- **Shared by [optional]:** [More Information Needed]
|
| 30 |
+
- **Model type:** [More Information Needed]
|
| 31 |
+
- **Language(s) (NLP):** [More Information Needed]
|
| 32 |
+
- **License:** [More Information Needed]
|
| 33 |
+
- **Finetuned from model [optional]:** [More Information Needed]
|
| 34 |
+
|
| 35 |
+
### Model Sources [optional]
|
| 36 |
+
|
| 37 |
+
<!-- Provide the basic links for the model. -->
|
| 38 |
+
|
| 39 |
+
- **Repository:** [More Information Needed]
|
| 40 |
+
- **Paper [optional]:** [More Information Needed]
|
| 41 |
+
- **Demo [optional]:** [More Information Needed]
|
| 42 |
+
|
| 43 |
+
## Uses
|
| 44 |
+
|
| 45 |
+
<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
|
| 46 |
+
|
| 47 |
+
### Direct Use
|
| 48 |
+
|
| 49 |
+
<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
|
| 50 |
+
|
| 51 |
+
[More Information Needed]
|
| 52 |
+
|
| 53 |
+
### Downstream Use [optional]
|
| 54 |
+
|
| 55 |
+
<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
|
| 56 |
+
|
| 57 |
+
[More Information Needed]
|
| 58 |
+
|
| 59 |
+
### Out-of-Scope Use
|
| 60 |
+
|
| 61 |
+
<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
|
| 62 |
+
|
| 63 |
+
[More Information Needed]
|
| 64 |
+
|
| 65 |
+
## Bias, Risks, and Limitations
|
| 66 |
+
|
| 67 |
+
<!-- This section is meant to convey both technical and sociotechnical limitations. -->
|
| 68 |
+
|
| 69 |
+
[More Information Needed]
|
| 70 |
+
|
| 71 |
+
### Recommendations
|
| 72 |
+
|
| 73 |
+
<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
|
| 74 |
+
|
| 75 |
+
Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
|
| 76 |
+
|
| 77 |
+
## How to Get Started with the Model
|
| 78 |
+
|
| 79 |
+
Use the code below to get started with the model.
|
| 80 |
+
|
| 81 |
+
[More Information Needed]
|
| 82 |
+
|
| 83 |
+
## Training Details
|
| 84 |
+
|
| 85 |
+
### Training Data
|
| 86 |
+
|
| 87 |
+
<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
|
| 88 |
+
|
| 89 |
+
[More Information Needed]
|
| 90 |
+
|
| 91 |
+
### Training Procedure
|
| 92 |
+
|
| 93 |
+
<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
|
| 94 |
+
|
| 95 |
+
#### Preprocessing [optional]
|
| 96 |
+
|
| 97 |
+
[More Information Needed]
|
| 98 |
+
|
| 99 |
+
|
| 100 |
+
#### Training Hyperparameters
|
| 101 |
+
|
| 102 |
+
- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
|
| 103 |
+
|
| 104 |
+
#### Speeds, Sizes, Times [optional]
|
| 105 |
+
|
| 106 |
+
<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
|
| 107 |
+
|
| 108 |
+
[More Information Needed]
|
| 109 |
+
|
| 110 |
+
## Evaluation
|
| 111 |
+
|
| 112 |
+
<!-- This section describes the evaluation protocols and provides the results. -->
|
| 113 |
+
|
| 114 |
+
### Testing Data, Factors & Metrics
|
| 115 |
+
|
| 116 |
+
#### Testing Data
|
| 117 |
+
|
| 118 |
+
<!-- This should link to a Dataset Card if possible. -->
|
| 119 |
+
|
| 120 |
+
[More Information Needed]
|
| 121 |
+
|
| 122 |
+
#### Factors
|
| 123 |
+
|
| 124 |
+
<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
|
| 125 |
+
|
| 126 |
+
[More Information Needed]
|
| 127 |
+
|
| 128 |
+
#### Metrics
|
| 129 |
+
|
| 130 |
+
<!-- These are the evaluation metrics being used, ideally with a description of why. -->
|
| 131 |
+
|
| 132 |
+
[More Information Needed]
|
| 133 |
+
|
| 134 |
+
### Results
|
| 135 |
+
|
| 136 |
+
[More Information Needed]
|
| 137 |
+
|
| 138 |
+
#### Summary
|
| 139 |
+
|
| 140 |
+
|
| 141 |
+
|
| 142 |
+
## Model Examination [optional]
|
| 143 |
+
|
| 144 |
+
<!-- Relevant interpretability work for the model goes here -->
|
| 145 |
+
|
| 146 |
+
[More Information Needed]
|
| 147 |
+
|
| 148 |
+
## Environmental Impact
|
| 149 |
+
|
| 150 |
+
<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
|
| 151 |
+
|
| 152 |
+
Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
|
| 153 |
+
|
| 154 |
+
- **Hardware Type:** [More Information Needed]
|
| 155 |
+
- **Hours used:** [More Information Needed]
|
| 156 |
+
- **Cloud Provider:** [More Information Needed]
|
| 157 |
+
- **Compute Region:** [More Information Needed]
|
| 158 |
+
- **Carbon Emitted:** [More Information Needed]
|
| 159 |
+
|
| 160 |
+
## Technical Specifications [optional]
|
| 161 |
+
|
| 162 |
+
### Model Architecture and Objective
|
| 163 |
+
|
| 164 |
+
[More Information Needed]
|
| 165 |
+
|
| 166 |
+
### Compute Infrastructure
|
| 167 |
+
|
| 168 |
+
[More Information Needed]
|
| 169 |
+
|
| 170 |
+
#### Hardware
|
| 171 |
+
|
| 172 |
+
[More Information Needed]
|
| 173 |
+
|
| 174 |
+
#### Software
|
| 175 |
+
|
| 176 |
+
[More Information Needed]
|
| 177 |
+
|
| 178 |
+
## Citation [optional]
|
| 179 |
+
|
| 180 |
+
<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
|
| 181 |
+
|
| 182 |
+
**BibTeX:**
|
| 183 |
+
|
| 184 |
+
[More Information Needed]
|
| 185 |
+
|
| 186 |
+
**APA:**
|
| 187 |
+
|
| 188 |
+
[More Information Needed]
|
| 189 |
+
|
| 190 |
+
## Glossary [optional]
|
| 191 |
+
|
| 192 |
+
<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
|
| 193 |
+
|
| 194 |
+
[More Information Needed]
|
| 195 |
+
|
| 196 |
+
## More Information [optional]
|
| 197 |
+
|
| 198 |
+
[More Information Needed]
|
| 199 |
+
|
| 200 |
+
## Model Card Authors [optional]
|
| 201 |
+
|
| 202 |
+
[More Information Needed]
|
| 203 |
+
|
| 204 |
+
## Model Card Contact
|
| 205 |
+
|
| 206 |
+
[More Information Needed]
|
| 207 |
+
### Framework versions
|
| 208 |
+
|
| 209 |
+
- PEFT 0.19.1
|
DBCA_original_Estonian/Qwen3.5-2B-Base_original_features_structural_train_original_features_structural_test1/checkpoint-3424/adapter_config.json
ADDED
|
@@ -0,0 +1,48 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"alora_invocation_tokens": null,
|
| 3 |
+
"alpha_pattern": {},
|
| 4 |
+
"arrow_config": null,
|
| 5 |
+
"auto_mapping": null,
|
| 6 |
+
"base_model_name_or_path": "Qwen/Qwen3.5-2B-Base",
|
| 7 |
+
"bias": "none",
|
| 8 |
+
"corda_config": null,
|
| 9 |
+
"ensure_weight_tying": false,
|
| 10 |
+
"eva_config": null,
|
| 11 |
+
"exclude_modules": null,
|
| 12 |
+
"fan_in_fan_out": false,
|
| 13 |
+
"inference_mode": true,
|
| 14 |
+
"init_lora_weights": true,
|
| 15 |
+
"layer_replication": null,
|
| 16 |
+
"layers_pattern": null,
|
| 17 |
+
"layers_to_transform": null,
|
| 18 |
+
"loftq_config": {},
|
| 19 |
+
"lora_alpha": 64,
|
| 20 |
+
"lora_bias": false,
|
| 21 |
+
"lora_dropout": 0.04180832058159435,
|
| 22 |
+
"lora_ga_config": null,
|
| 23 |
+
"megatron_config": null,
|
| 24 |
+
"megatron_core": "megatron.core",
|
| 25 |
+
"modules_to_save": null,
|
| 26 |
+
"peft_type": "LORA",
|
| 27 |
+
"peft_version": "0.19.1",
|
| 28 |
+
"qalora_group_size": 16,
|
| 29 |
+
"r": 64,
|
| 30 |
+
"rank_pattern": {},
|
| 31 |
+
"revision": null,
|
| 32 |
+
"target_modules": [
|
| 33 |
+
"v_proj",
|
| 34 |
+
"o_proj",
|
| 35 |
+
"q_proj",
|
| 36 |
+
"gate_proj",
|
| 37 |
+
"k_proj",
|
| 38 |
+
"down_proj",
|
| 39 |
+
"up_proj"
|
| 40 |
+
],
|
| 41 |
+
"target_parameters": null,
|
| 42 |
+
"task_type": "CAUSAL_LM",
|
| 43 |
+
"trainable_token_indices": null,
|
| 44 |
+
"use_bdlora": null,
|
| 45 |
+
"use_dora": false,
|
| 46 |
+
"use_qalora": false,
|
| 47 |
+
"use_rslora": false
|
| 48 |
+
}
|
DBCA_original_Estonian/Qwen3.5-2B-Base_original_features_structural_train_original_features_structural_test1/checkpoint-3424/chat_template.jinja
ADDED
|
@@ -0,0 +1,154 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{%- set image_count = namespace(value=0) %}
|
| 2 |
+
{%- set video_count = namespace(value=0) %}
|
| 3 |
+
{%- macro render_content(content, do_vision_count, is_system_content=false) %}
|
| 4 |
+
{%- if content is string %}
|
| 5 |
+
{{- content }}
|
| 6 |
+
{%- elif content is iterable and content is not mapping %}
|
| 7 |
+
{%- for item in content %}
|
| 8 |
+
{%- if 'image' in item or 'image_url' in item or item.type == 'image' %}
|
| 9 |
+
{%- if is_system_content %}
|
| 10 |
+
{{- raise_exception('System message cannot contain images.') }}
|
| 11 |
+
{%- endif %}
|
| 12 |
+
{%- if do_vision_count %}
|
| 13 |
+
{%- set image_count.value = image_count.value + 1 %}
|
| 14 |
+
{%- endif %}
|
| 15 |
+
{%- if add_vision_id %}
|
| 16 |
+
{{- 'Picture ' ~ image_count.value ~ ': ' }}
|
| 17 |
+
{%- endif %}
|
| 18 |
+
{{- '<|vision_start|><|image_pad|><|vision_end|>' }}
|
| 19 |
+
{%- elif 'video' in item or item.type == 'video' %}
|
| 20 |
+
{%- if is_system_content %}
|
| 21 |
+
{{- raise_exception('System message cannot contain videos.') }}
|
| 22 |
+
{%- endif %}
|
| 23 |
+
{%- if do_vision_count %}
|
| 24 |
+
{%- set video_count.value = video_count.value + 1 %}
|
| 25 |
+
{%- endif %}
|
| 26 |
+
{%- if add_vision_id %}
|
| 27 |
+
{{- 'Video ' ~ video_count.value ~ ': ' }}
|
| 28 |
+
{%- endif %}
|
| 29 |
+
{{- '<|vision_start|><|video_pad|><|vision_end|>' }}
|
| 30 |
+
{%- elif 'text' in item %}
|
| 31 |
+
{{- item.text }}
|
| 32 |
+
{%- else %}
|
| 33 |
+
{{- raise_exception('Unexpected item type in content.') }}
|
| 34 |
+
{%- endif %}
|
| 35 |
+
{%- endfor %}
|
| 36 |
+
{%- elif content is none or content is undefined %}
|
| 37 |
+
{{- '' }}
|
| 38 |
+
{%- else %}
|
| 39 |
+
{{- raise_exception('Unexpected content type.') }}
|
| 40 |
+
{%- endif %}
|
| 41 |
+
{%- endmacro %}
|
| 42 |
+
{%- if not messages %}
|
| 43 |
+
{{- raise_exception('No messages provided.') }}
|
| 44 |
+
{%- endif %}
|
| 45 |
+
{%- if tools and tools is iterable and tools is not mapping %}
|
| 46 |
+
{{- '<|im_start|>system\n' }}
|
| 47 |
+
{{- "# Tools\n\nYou have access to the following functions:\n\n<tools>" }}
|
| 48 |
+
{%- for tool in tools %}
|
| 49 |
+
{{- "\n" }}
|
| 50 |
+
{{- tool | tojson }}
|
| 51 |
+
{%- endfor %}
|
| 52 |
+
{{- "\n</tools>" }}
|
| 53 |
+
{{- '\n\nIf you choose to call a function ONLY reply in the following format with NO suffix:\n\n<tool_call>\n<function=example_function_name>\n<parameter=example_parameter_1>\nvalue_1\n</parameter>\n<parameter=example_parameter_2>\nThis is the value for the second parameter\nthat can span\nmultiple lines\n</parameter>\n</function>\n</tool_call>\n\n<IMPORTANT>\nReminder:\n- Function calls MUST follow the specified format: an inner <function=...></function> block must be nested within <tool_call></tool_call> XML tags\n- Required parameters MUST be specified\n- You may provide optional reasoning for your function call in natural language BEFORE the function call, but NOT after\n- If there is no function call available, answer the question like normal with your current knowledge and do not tell the user about function calls\n</IMPORTANT>' }}
|
| 54 |
+
{%- if messages[0].role == 'system' %}
|
| 55 |
+
{%- set content = render_content(messages[0].content, false, true)|trim %}
|
| 56 |
+
{%- if content %}
|
| 57 |
+
{{- '\n\n' + content }}
|
| 58 |
+
{%- endif %}
|
| 59 |
+
{%- endif %}
|
| 60 |
+
{{- '<|im_end|>\n' }}
|
| 61 |
+
{%- else %}
|
| 62 |
+
{%- if messages[0].role == 'system' %}
|
| 63 |
+
{%- set content = render_content(messages[0].content, false, true)|trim %}
|
| 64 |
+
{{- '<|im_start|>system\n' + content + '<|im_end|>\n' }}
|
| 65 |
+
{%- endif %}
|
| 66 |
+
{%- endif %}
|
| 67 |
+
{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
|
| 68 |
+
{%- for message in messages[::-1] %}
|
| 69 |
+
{%- set index = (messages|length - 1) - loop.index0 %}
|
| 70 |
+
{%- if ns.multi_step_tool and message.role == "user" %}
|
| 71 |
+
{%- set content = render_content(message.content, false)|trim %}
|
| 72 |
+
{%- if not(content.startswith('<tool_response>') and content.endswith('</tool_response>')) %}
|
| 73 |
+
{%- set ns.multi_step_tool = false %}
|
| 74 |
+
{%- set ns.last_query_index = index %}
|
| 75 |
+
{%- endif %}
|
| 76 |
+
{%- endif %}
|
| 77 |
+
{%- endfor %}
|
| 78 |
+
{%- if ns.multi_step_tool %}
|
| 79 |
+
{{- raise_exception('No user query found in messages.') }}
|
| 80 |
+
{%- endif %}
|
| 81 |
+
{%- for message in messages %}
|
| 82 |
+
{%- set content = render_content(message.content, true)|trim %}
|
| 83 |
+
{%- if message.role == "system" %}
|
| 84 |
+
{%- if not loop.first %}
|
| 85 |
+
{{- raise_exception('System message must be at the beginning.') }}
|
| 86 |
+
{%- endif %}
|
| 87 |
+
{%- elif message.role == "user" %}
|
| 88 |
+
{{- '<|im_start|>' + message.role + '\n' + content + '<|im_end|>' + '\n' }}
|
| 89 |
+
{%- elif message.role == "assistant" %}
|
| 90 |
+
{%- set reasoning_content = '' %}
|
| 91 |
+
{%- if message.reasoning_content is string %}
|
| 92 |
+
{%- set reasoning_content = message.reasoning_content %}
|
| 93 |
+
{%- else %}
|
| 94 |
+
{%- if '</think>' in content %}
|
| 95 |
+
{%- set reasoning_content = content.split('</think>')[0].rstrip('\n').split('<think>')[-1].lstrip('\n') %}
|
| 96 |
+
{%- set content = content.split('</think>')[-1].lstrip('\n') %}
|
| 97 |
+
{%- endif %}
|
| 98 |
+
{%- endif %}
|
| 99 |
+
{%- set reasoning_content = reasoning_content|trim %}
|
| 100 |
+
{%- if loop.index0 > ns.last_query_index %}
|
| 101 |
+
{{- '<|im_start|>' + message.role + '\n<think>\n' + reasoning_content + '\n</think>\n\n' + content }}
|
| 102 |
+
{%- else %}
|
| 103 |
+
{{- '<|im_start|>' + message.role + '\n' + content }}
|
| 104 |
+
{%- endif %}
|
| 105 |
+
{%- if message.tool_calls and message.tool_calls is iterable and message.tool_calls is not mapping %}
|
| 106 |
+
{%- for tool_call in message.tool_calls %}
|
| 107 |
+
{%- if tool_call.function is defined %}
|
| 108 |
+
{%- set tool_call = tool_call.function %}
|
| 109 |
+
{%- endif %}
|
| 110 |
+
{%- if loop.first %}
|
| 111 |
+
{%- if content|trim %}
|
| 112 |
+
{{- '\n\n<tool_call>\n<function=' + tool_call.name + '>\n' }}
|
| 113 |
+
{%- else %}
|
| 114 |
+
{{- '<tool_call>\n<function=' + tool_call.name + '>\n' }}
|
| 115 |
+
{%- endif %}
|
| 116 |
+
{%- else %}
|
| 117 |
+
{{- '\n<tool_call>\n<function=' + tool_call.name + '>\n' }}
|
| 118 |
+
{%- endif %}
|
| 119 |
+
{%- if tool_call.arguments is defined %}
|
| 120 |
+
{%- for args_name, args_value in tool_call.arguments|items %}
|
| 121 |
+
{{- '<parameter=' + args_name + '>\n' }}
|
| 122 |
+
{%- set args_value = args_value | tojson | safe if args_value is mapping or (args_value is sequence and args_value is not string) else args_value | string %}
|
| 123 |
+
{{- args_value }}
|
| 124 |
+
{{- '\n</parameter>\n' }}
|
| 125 |
+
{%- endfor %}
|
| 126 |
+
{%- endif %}
|
| 127 |
+
{{- '</function>\n</tool_call>' }}
|
| 128 |
+
{%- endfor %}
|
| 129 |
+
{%- endif %}
|
| 130 |
+
{{- '<|im_end|>\n' }}
|
| 131 |
+
{%- elif message.role == "tool" %}
|
| 132 |
+
{%- if loop.previtem and loop.previtem.role != "tool" %}
|
| 133 |
+
{{- '<|im_start|>user' }}
|
| 134 |
+
{%- endif %}
|
| 135 |
+
{{- '\n<tool_response>\n' }}
|
| 136 |
+
{{- content }}
|
| 137 |
+
{{- '\n</tool_response>' }}
|
| 138 |
+
{%- if not loop.last and loop.nextitem.role != "tool" %}
|
| 139 |
+
{{- '<|im_end|>\n' }}
|
| 140 |
+
{%- elif loop.last %}
|
| 141 |
+
{{- '<|im_end|>\n' }}
|
| 142 |
+
{%- endif %}
|
| 143 |
+
{%- else %}
|
| 144 |
+
{{- raise_exception('Unexpected message role.') }}
|
| 145 |
+
{%- endif %}
|
| 146 |
+
{%- endfor %}
|
| 147 |
+
{%- if add_generation_prompt %}
|
| 148 |
+
{{- '<|im_start|>assistant\n' }}
|
| 149 |
+
{%- if enable_thinking is defined and enable_thinking is true %}
|
| 150 |
+
{{- '<think>\n' }}
|
| 151 |
+
{%- else %}
|
| 152 |
+
{{- '<think>\n\n</think>\n\n' }}
|
| 153 |
+
{%- endif %}
|
| 154 |
+
{%- endif %}
|
DBCA_original_Estonian/Qwen3.5-2B-Base_original_features_structural_train_original_features_structural_test1/checkpoint-3424/tokenizer_config.json
ADDED
|
@@ -0,0 +1,31 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"add_prefix_space": false,
|
| 3 |
+
"audio_bos_token": "<|audio_start|>",
|
| 4 |
+
"audio_eos_token": "<|audio_end|>",
|
| 5 |
+
"audio_token": "<|audio_pad|>",
|
| 6 |
+
"backend": "tokenizers",
|
| 7 |
+
"bos_token": null,
|
| 8 |
+
"clean_up_tokenization_spaces": false,
|
| 9 |
+
"eos_token": "<|endoftext|>",
|
| 10 |
+
"errors": "replace",
|
| 11 |
+
"image_token": "<|image_pad|>",
|
| 12 |
+
"is_local": false,
|
| 13 |
+
"model_max_length": 262144,
|
| 14 |
+
"model_specific_special_tokens": {
|
| 15 |
+
"audio_bos_token": "<|audio_start|>",
|
| 16 |
+
"audio_eos_token": "<|audio_end|>",
|
| 17 |
+
"audio_token": "<|audio_pad|>",
|
| 18 |
+
"image_token": "<|image_pad|>",
|
| 19 |
+
"video_token": "<|video_pad|>",
|
| 20 |
+
"vision_bos_token": "<|vision_start|>",
|
| 21 |
+
"vision_eos_token": "<|vision_end|>"
|
| 22 |
+
},
|
| 23 |
+
"pad_token": "<|endoftext|>",
|
| 24 |
+
"pretokenize_regex": "(?i:'s|'t|'re|'ve|'m|'ll|'d)|[^\\r\\n\\p{L}\\p{N}]?[\\p{L}\\p{M}]+|\\p{N}| ?[^\\s\\p{L}\\p{M}\\p{N}]+[\\r\\n]*|\\s*[\\r\\n]+|\\s+(?!\\S)|\\s+",
|
| 25 |
+
"split_special_tokens": false,
|
| 26 |
+
"tokenizer_class": "TokenizersBackend",
|
| 27 |
+
"unk_token": null,
|
| 28 |
+
"video_token": "<|video_pad|>",
|
| 29 |
+
"vision_bos_token": "<|vision_start|>",
|
| 30 |
+
"vision_eos_token": "<|vision_end|>"
|
| 31 |
+
}
|
DBCA_original_Estonian/Qwen3.5-2B-Base_original_features_structural_train_original_features_structural_test1/checkpoint-3424/trainer_state.json
ADDED
|
@@ -0,0 +1,802 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": null,
|
| 3 |
+
"best_metric": null,
|
| 4 |
+
"best_model_checkpoint": null,
|
| 5 |
+
"epoch": 8.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 3424,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"entropy": 1.8341716039180755,
|
| 14 |
+
"epoch": 0.11689070718877849,
|
| 15 |
+
"grad_norm": 1.1992322206497192,
|
| 16 |
+
"learning_rate": 5.366728168873013e-05,
|
| 17 |
+
"loss": 1.6918942260742187,
|
| 18 |
+
"mean_token_accuracy": 0.631798365265131,
|
| 19 |
+
"num_tokens": 153689.0,
|
| 20 |
+
"step": 50
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"entropy": 0.9513179230690002,
|
| 24 |
+
"epoch": 0.23378141437755698,
|
| 25 |
+
"grad_norm": 1.1290454864501953,
|
| 26 |
+
"learning_rate": 0.00010842981402416902,
|
| 27 |
+
"loss": 0.8867578887939453,
|
| 28 |
+
"mean_token_accuracy": 0.7587438315153122,
|
| 29 |
+
"num_tokens": 310548.0,
|
| 30 |
+
"step": 100
|
| 31 |
+
},
|
| 32 |
+
{
|
| 33 |
+
"entropy": 0.8510265004634857,
|
| 34 |
+
"epoch": 0.3506721215663355,
|
| 35 |
+
"grad_norm": 0.7732136845588684,
|
| 36 |
+
"learning_rate": 0.00016319234635960792,
|
| 37 |
+
"loss": 0.7856448364257812,
|
| 38 |
+
"mean_token_accuracy": 0.7769897204637527,
|
| 39 |
+
"num_tokens": 468560.0,
|
| 40 |
+
"step": 150
|
| 41 |
+
},
|
| 42 |
+
{
|
| 43 |
+
"entropy": 0.8142367601394653,
|
| 44 |
+
"epoch": 0.46756282875511396,
|
| 45 |
+
"grad_norm": 0.7580806016921997,
|
| 46 |
+
"learning_rate": 0.00021795487869504684,
|
| 47 |
+
"loss": 0.7461666870117187,
|
| 48 |
+
"mean_token_accuracy": 0.7888966089487076,
|
| 49 |
+
"num_tokens": 619552.0,
|
| 50 |
+
"step": 200
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"entropy": 0.772578022480011,
|
| 54 |
+
"epoch": 0.5844535359438925,
|
| 55 |
+
"grad_norm": 0.5330358147621155,
|
| 56 |
+
"learning_rate": 0.00027271741103048575,
|
| 57 |
+
"loss": 0.7142900848388671,
|
| 58 |
+
"mean_token_accuracy": 0.7959607627987861,
|
| 59 |
+
"num_tokens": 771972.0,
|
| 60 |
+
"step": 250
|
| 61 |
+
},
|
| 62 |
+
{
|
| 63 |
+
"entropy": 0.7481484657526016,
|
| 64 |
+
"epoch": 0.701344243132671,
|
| 65 |
+
"grad_norm": 0.8242517709732056,
|
| 66 |
+
"learning_rate": 0.00032747994336592464,
|
| 67 |
+
"loss": 0.6911422729492187,
|
| 68 |
+
"mean_token_accuracy": 0.8028205358982086,
|
| 69 |
+
"num_tokens": 928088.0,
|
| 70 |
+
"step": 300
|
| 71 |
+
},
|
| 72 |
+
{
|
| 73 |
+
"entropy": 0.741372903585434,
|
| 74 |
+
"epoch": 0.8182349503214494,
|
| 75 |
+
"grad_norm": 0.7722981572151184,
|
| 76 |
+
"learning_rate": 0.00038224247570136353,
|
| 77 |
+
"loss": 0.6908904266357422,
|
| 78 |
+
"mean_token_accuracy": 0.8018302822113037,
|
| 79 |
+
"num_tokens": 1086836.0,
|
| 80 |
+
"step": 350
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"entropy": 0.7257227802276611,
|
| 84 |
+
"epoch": 0.9351256575102279,
|
| 85 |
+
"grad_norm": 0.7969573140144348,
|
| 86 |
+
"learning_rate": 0.0004370050080368024,
|
| 87 |
+
"loss": 0.6714310455322265,
|
| 88 |
+
"mean_token_accuracy": 0.8057891410589219,
|
| 89 |
+
"num_tokens": 1241227.0,
|
| 90 |
+
"step": 400
|
| 91 |
+
},
|
| 92 |
+
{
|
| 93 |
+
"epoch": 1.0,
|
| 94 |
+
"eval_entropy": 0.8317563653766334,
|
| 95 |
+
"eval_loss": 0.8135058283805847,
|
| 96 |
+
"eval_mean_token_accuracy": 0.7755894215850087,
|
| 97 |
+
"eval_num_tokens": 1325643.0,
|
| 98 |
+
"eval_runtime": 37.4951,
|
| 99 |
+
"eval_samples_per_second": 32.831,
|
| 100 |
+
"eval_steps_per_second": 4.107,
|
| 101 |
+
"step": 428
|
| 102 |
+
},
|
| 103 |
+
{
|
| 104 |
+
"entropy": 0.7179872094087265,
|
| 105 |
+
"epoch": 1.0514319111630626,
|
| 106 |
+
"grad_norm": 0.6224768161773682,
|
| 107 |
+
"learning_rate": 0.00046873290101955413,
|
| 108 |
+
"loss": 0.6683222961425781,
|
| 109 |
+
"mean_token_accuracy": 0.8054787133207273,
|
| 110 |
+
"num_tokens": 1393922.0,
|
| 111 |
+
"step": 450
|
| 112 |
+
},
|
| 113 |
+
{
|
| 114 |
+
"entropy": 0.6991294291615486,
|
| 115 |
+
"epoch": 1.1683226183518411,
|
| 116 |
+
"grad_norm": 0.6813111901283264,
|
| 117 |
+
"learning_rate": 0.00046837443306086947,
|
| 118 |
+
"loss": 0.6493977355957031,
|
| 119 |
+
"mean_token_accuracy": 0.8089349576830864,
|
| 120 |
+
"num_tokens": 1551442.0,
|
| 121 |
+
"step": 500
|
| 122 |
+
},
|
| 123 |
+
{
|
| 124 |
+
"entropy": 0.6951853144168854,
|
| 125 |
+
"epoch": 1.2852133255406195,
|
| 126 |
+
"grad_norm": 0.617091953754425,
|
| 127 |
+
"learning_rate": 0.0004676269147738558,
|
| 128 |
+
"loss": 0.6484123229980469,
|
| 129 |
+
"mean_token_accuracy": 0.8088837671279907,
|
| 130 |
+
"num_tokens": 1706697.0,
|
| 131 |
+
"step": 550
|
| 132 |
+
},
|
| 133 |
+
{
|
| 134 |
+
"entropy": 0.6852494943141937,
|
| 135 |
+
"epoch": 1.4021040327293979,
|
| 136 |
+
"grad_norm": 0.4677598774433136,
|
| 137 |
+
"learning_rate": 0.0004664915890374708,
|
| 138 |
+
"loss": 0.6416233062744141,
|
| 139 |
+
"mean_token_accuracy": 0.8106312158703805,
|
| 140 |
+
"num_tokens": 1865246.0,
|
| 141 |
+
"step": 600
|
| 142 |
+
},
|
| 143 |
+
{
|
| 144 |
+
"entropy": 0.6703594943881035,
|
| 145 |
+
"epoch": 1.5189947399181765,
|
| 146 |
+
"grad_norm": 0.5708025693893433,
|
| 147 |
+
"learning_rate": 0.0004649703435278991,
|
| 148 |
+
"loss": 0.6273183441162109,
|
| 149 |
+
"mean_token_accuracy": 0.8155005398392677,
|
| 150 |
+
"num_tokens": 2022060.0,
|
| 151 |
+
"step": 650
|
| 152 |
+
},
|
| 153 |
+
{
|
| 154 |
+
"entropy": 0.6676286320388317,
|
| 155 |
+
"epoch": 1.635885447106955,
|
| 156 |
+
"grad_norm": 0.5555347800254822,
|
| 157 |
+
"learning_rate": 0.00046306570757996264,
|
| 158 |
+
"loss": 0.6276700973510743,
|
| 159 |
+
"mean_token_accuracy": 0.8154828292131424,
|
| 160 |
+
"num_tokens": 2178909.0,
|
| 161 |
+
"step": 700
|
| 162 |
+
},
|
| 163 |
+
{
|
| 164 |
+
"entropy": 0.6669242936372757,
|
| 165 |
+
"epoch": 1.7527761542957334,
|
| 166 |
+
"grad_norm": 0.5260112285614014,
|
| 167 |
+
"learning_rate": 0.0004607808479816624,
|
| 168 |
+
"loss": 0.6228141784667969,
|
| 169 |
+
"mean_token_accuracy": 0.8164840793609619,
|
| 170 |
+
"num_tokens": 2329314.0,
|
| 171 |
+
"step": 750
|
| 172 |
+
},
|
| 173 |
+
{
|
| 174 |
+
"entropy": 0.6598560312390327,
|
| 175 |
+
"epoch": 1.869666861484512,
|
| 176 |
+
"grad_norm": 0.5677736401557922,
|
| 177 |
+
"learning_rate": 0.0004581195637088436,
|
| 178 |
+
"loss": 0.6214292907714843,
|
| 179 |
+
"mean_token_accuracy": 0.8171795177459716,
|
| 180 |
+
"num_tokens": 2478559.0,
|
| 181 |
+
"step": 800
|
| 182 |
+
},
|
| 183 |
+
{
|
| 184 |
+
"entropy": 0.6687870016694069,
|
| 185 |
+
"epoch": 1.9865575686732906,
|
| 186 |
+
"grad_norm": 0.6449615955352783,
|
| 187 |
+
"learning_rate": 0.00045508627960873823,
|
| 188 |
+
"loss": 0.6243909454345703,
|
| 189 |
+
"mean_token_accuracy": 0.8147780740261078,
|
| 190 |
+
"num_tokens": 2631937.0,
|
| 191 |
+
"step": 850
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"epoch": 2.0,
|
| 195 |
+
"eval_entropy": 0.7786144471013701,
|
| 196 |
+
"eval_loss": 0.7599140405654907,
|
| 197 |
+
"eval_mean_token_accuracy": 0.7935443106409791,
|
| 198 |
+
"eval_num_tokens": 2651286.0,
|
| 199 |
+
"eval_runtime": 39.0048,
|
| 200 |
+
"eval_samples_per_second": 31.56,
|
| 201 |
+
"eval_steps_per_second": 3.948,
|
| 202 |
+
"step": 856
|
| 203 |
+
},
|
| 204 |
+
{
|
| 205 |
+
"entropy": 0.5821921329701966,
|
| 206 |
+
"epoch": 2.102863822326125,
|
| 207 |
+
"grad_norm": 0.6116960048675537,
|
| 208 |
+
"learning_rate": 0.00045168603904288863,
|
| 209 |
+
"loss": 0.5406004714965821,
|
| 210 |
+
"mean_token_accuracy": 0.8334921918921734,
|
| 211 |
+
"num_tokens": 2795200.0,
|
| 212 |
+
"step": 900
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"entropy": 0.5937331764400006,
|
| 216 |
+
"epoch": 2.2197545295149035,
|
| 217 |
+
"grad_norm": 0.5440122485160828,
|
| 218 |
+
"learning_rate": 0.00044792449550168286,
|
| 219 |
+
"loss": 0.5515737533569336,
|
| 220 |
+
"mean_token_accuracy": 0.8314691257476806,
|
| 221 |
+
"num_tokens": 2950534.0,
|
| 222 |
+
"step": 950
|
| 223 |
+
},
|
| 224 |
+
{
|
| 225 |
+
"entropy": 0.5877273553609847,
|
| 226 |
+
"epoch": 2.3366452367036823,
|
| 227 |
+
"grad_norm": 0.6505260467529297,
|
| 228 |
+
"learning_rate": 0.0004438079032044453,
|
| 229 |
+
"loss": 0.5507744979858399,
|
| 230 |
+
"mean_token_accuracy": 0.8314063146710395,
|
| 231 |
+
"num_tokens": 3101471.0,
|
| 232 |
+
"step": 1000
|
| 233 |
+
},
|
| 234 |
+
{
|
| 235 |
+
"entropy": 0.5897100016474723,
|
| 236 |
+
"epoch": 2.4535359438924607,
|
| 237 |
+
"grad_norm": 0.54567551612854,
|
| 238 |
+
"learning_rate": 0.0004393431067007111,
|
| 239 |
+
"loss": 0.5528768157958984,
|
| 240 |
+
"mean_token_accuracy": 0.8315245220065117,
|
| 241 |
+
"num_tokens": 3258554.0,
|
| 242 |
+
"step": 1050
|
| 243 |
+
},
|
| 244 |
+
{
|
| 245 |
+
"entropy": 0.5831590622663498,
|
| 246 |
+
"epoch": 2.570426651081239,
|
| 247 |
+
"grad_norm": 0.44826796650886536,
|
| 248 |
+
"learning_rate": 0.00043453752948997376,
|
| 249 |
+
"loss": 0.5442767333984375,
|
| 250 |
+
"mean_token_accuracy": 0.8327079233527184,
|
| 251 |
+
"num_tokens": 3416088.0,
|
| 252 |
+
"step": 1100
|
| 253 |
+
},
|
| 254 |
+
{
|
| 255 |
+
"entropy": 0.5839906217157841,
|
| 256 |
+
"epoch": 2.6873173582700174,
|
| 257 |
+
"grad_norm": 0.5067696571350098,
|
| 258 |
+
"learning_rate": 0.0004293991616788285,
|
| 259 |
+
"loss": 0.5485663604736328,
|
| 260 |
+
"mean_token_accuracy": 0.8324281191825866,
|
| 261 |
+
"num_tokens": 3567251.0,
|
| 262 |
+
"step": 1150
|
| 263 |
+
},
|
| 264 |
+
{
|
| 265 |
+
"entropy": 0.5870274990797043,
|
| 266 |
+
"epoch": 2.8042080654587958,
|
| 267 |
+
"grad_norm": 0.44032806158065796,
|
| 268 |
+
"learning_rate": 0.00042393654669603217,
|
| 269 |
+
"loss": 0.5524474716186524,
|
| 270 |
+
"mean_token_accuracy": 0.8313613015413285,
|
| 271 |
+
"num_tokens": 3719464.0,
|
| 272 |
+
"step": 1200
|
| 273 |
+
},
|
| 274 |
+
{
|
| 275 |
+
"entropy": 0.5864466108381748,
|
| 276 |
+
"epoch": 2.9210987726475746,
|
| 277 |
+
"grad_norm": 0.5520864129066467,
|
| 278 |
+
"learning_rate": 0.00041815876708756964,
|
| 279 |
+
"loss": 0.5490786361694336,
|
| 280 |
+
"mean_token_accuracy": 0.8314764249324799,
|
| 281 |
+
"num_tokens": 3874423.0,
|
| 282 |
+
"step": 1250
|
| 283 |
+
},
|
| 284 |
+
{
|
| 285 |
+
"epoch": 3.0,
|
| 286 |
+
"eval_entropy": 0.6819627613990338,
|
| 287 |
+
"eval_loss": 0.7593190670013428,
|
| 288 |
+
"eval_mean_token_accuracy": 0.7857603395914102,
|
| 289 |
+
"eval_num_tokens": 3976929.0,
|
| 290 |
+
"eval_runtime": 39.3654,
|
| 291 |
+
"eval_samples_per_second": 31.271,
|
| 292 |
+
"eval_steps_per_second": 3.912,
|
| 293 |
+
"step": 1284
|
| 294 |
+
},
|
| 295 |
+
{
|
| 296 |
+
"entropy": 0.5587641523411525,
|
| 297 |
+
"epoch": 3.037405026300409,
|
| 298 |
+
"grad_norm": 0.622401773929596,
|
| 299 |
+
"learning_rate": 0.0004120754294153441,
|
| 300 |
+
"loss": 0.518932762145996,
|
| 301 |
+
"mean_token_accuracy": 0.8399260457436643,
|
| 302 |
+
"num_tokens": 4027609.0,
|
| 303 |
+
"step": 1300
|
| 304 |
+
},
|
| 305 |
+
{
|
| 306 |
+
"entropy": 0.5028628017008304,
|
| 307 |
+
"epoch": 3.1542957334891875,
|
| 308 |
+
"grad_norm": 0.577396810054779,
|
| 309 |
+
"learning_rate": 0.00040569664828459917,
|
| 310 |
+
"loss": 0.4589382171630859,
|
| 311 |
+
"mean_token_accuracy": 0.8538418188691139,
|
| 312 |
+
"num_tokens": 4180560.0,
|
| 313 |
+
"step": 1350
|
| 314 |
+
},
|
| 315 |
+
{
|
| 316 |
+
"entropy": 0.5227407096326351,
|
| 317 |
+
"epoch": 3.2711864406779663,
|
| 318 |
+
"grad_norm": 0.5113071203231812,
|
| 319 |
+
"learning_rate": 0.00039903302952663176,
|
| 320 |
+
"loss": 0.47863777160644533,
|
| 321 |
+
"mean_token_accuracy": 0.847364938557148,
|
| 322 |
+
"num_tokens": 4329954.0,
|
| 323 |
+
"step": 1400
|
| 324 |
+
},
|
| 325 |
+
{
|
| 326 |
+
"entropy": 0.5050077450275421,
|
| 327 |
+
"epoch": 3.3880771478667446,
|
| 328 |
+
"grad_norm": 0.48977184295654297,
|
| 329 |
+
"learning_rate": 0.0003920956525647558,
|
| 330 |
+
"loss": 0.46841480255126955,
|
| 331 |
+
"mean_token_accuracy": 0.8519425508379936,
|
| 332 |
+
"num_tokens": 4484354.0,
|
| 333 |
+
"step": 1450
|
| 334 |
+
},
|
| 335 |
+
{
|
| 336 |
+
"entropy": 0.5163291451334954,
|
| 337 |
+
"epoch": 3.504967855055523,
|
| 338 |
+
"grad_norm": 0.509864330291748,
|
| 339 |
+
"learning_rate": 0.000384896051992837,
|
| 340 |
+
"loss": 0.47587432861328127,
|
| 341 |
+
"mean_token_accuracy": 0.8486990982294083,
|
| 342 |
+
"num_tokens": 4638919.0,
|
| 343 |
+
"step": 1500
|
| 344 |
+
},
|
| 345 |
+
{
|
| 346 |
+
"entropy": 0.5134807989001274,
|
| 347 |
+
"epoch": 3.6218585622443014,
|
| 348 |
+
"grad_norm": 0.46544620394706726,
|
| 349 |
+
"learning_rate": 0.00037744619839702735,
|
| 350 |
+
"loss": 0.47692710876464844,
|
| 351 |
+
"mean_token_accuracy": 0.8489899519085884,
|
| 352 |
+
"num_tokens": 4795540.0,
|
| 353 |
+
"step": 1550
|
| 354 |
+
},
|
| 355 |
+
{
|
| 356 |
+
"entropy": 0.5216251534223556,
|
| 357 |
+
"epoch": 3.73874926943308,
|
| 358 |
+
"grad_norm": 0.5167004466056824,
|
| 359 |
+
"learning_rate": 0.0003697584784525874,
|
| 360 |
+
"loss": 0.4837848663330078,
|
| 361 |
+
"mean_token_accuracy": 0.847066233754158,
|
| 362 |
+
"num_tokens": 4951177.0,
|
| 363 |
+
"step": 1600
|
| 364 |
+
},
|
| 365 |
+
{
|
| 366 |
+
"entropy": 0.5254101701080799,
|
| 367 |
+
"epoch": 3.8556399766218585,
|
| 368 |
+
"grad_norm": 0.5504006743431091,
|
| 369 |
+
"learning_rate": 0.00036184567432888745,
|
| 370 |
+
"loss": 0.48759506225585936,
|
| 371 |
+
"mean_token_accuracy": 0.8457375919818878,
|
| 372 |
+
"num_tokens": 5104814.0,
|
| 373 |
+
"step": 1650
|
| 374 |
+
},
|
| 375 |
+
{
|
| 376 |
+
"entropy": 0.5107753933966159,
|
| 377 |
+
"epoch": 3.972530683810637,
|
| 378 |
+
"grad_norm": 0.421975314617157,
|
| 379 |
+
"learning_rate": 0.0003537209424368311,
|
| 380 |
+
"loss": 0.4743759536743164,
|
| 381 |
+
"mean_token_accuracy": 0.8505110186338425,
|
| 382 |
+
"num_tokens": 5265185.0,
|
| 383 |
+
"step": 1700
|
| 384 |
+
},
|
| 385 |
+
{
|
| 386 |
+
"epoch": 4.0,
|
| 387 |
+
"eval_entropy": 0.665543915002377,
|
| 388 |
+
"eval_loss": 0.7487082481384277,
|
| 389 |
+
"eval_mean_token_accuracy": 0.7999678904359991,
|
| 390 |
+
"eval_num_tokens": 5302572.0,
|
| 391 |
+
"eval_runtime": 38.5439,
|
| 392 |
+
"eval_samples_per_second": 31.938,
|
| 393 |
+
"eval_steps_per_second": 3.995,
|
| 394 |
+
"step": 1712
|
| 395 |
+
},
|
| 396 |
+
{
|
| 397 |
+
"entropy": 0.4414465431891494,
|
| 398 |
+
"epoch": 4.088836937463472,
|
| 399 |
+
"grad_norm": 0.5308493971824646,
|
| 400 |
+
"learning_rate": 0.0003453977915540383,
|
| 401 |
+
"loss": 0.39650299072265627,
|
| 402 |
+
"mean_token_accuracy": 0.8703725772287378,
|
| 403 |
+
"num_tokens": 5419002.0,
|
| 404 |
+
"step": 1750
|
| 405 |
+
},
|
| 406 |
+
{
|
| 407 |
+
"entropy": 0.42591195791959763,
|
| 408 |
+
"epoch": 4.20572764465225,
|
| 409 |
+
"grad_norm": 0.6111563444137573,
|
| 410 |
+
"learning_rate": 0.00033689006036415585,
|
| 411 |
+
"loss": 0.37838283538818357,
|
| 412 |
+
"mean_token_accuracy": 0.874235480427742,
|
| 413 |
+
"num_tokens": 5572969.0,
|
| 414 |
+
"step": 1800
|
| 415 |
+
},
|
| 416 |
+
{
|
| 417 |
+
"entropy": 0.44086351931095125,
|
| 418 |
+
"epoch": 4.322618351841029,
|
| 419 |
+
"grad_norm": 0.5524880290031433,
|
| 420 |
+
"learning_rate": 0.0003282118944476435,
|
| 421 |
+
"loss": 0.3920352554321289,
|
| 422 |
+
"mean_token_accuracy": 0.8694414687156677,
|
| 423 |
+
"num_tokens": 5733805.0,
|
| 424 |
+
"step": 1850
|
| 425 |
+
},
|
| 426 |
+
{
|
| 427 |
+
"entropy": 0.4380896310508251,
|
| 428 |
+
"epoch": 4.439509059029807,
|
| 429 |
+
"grad_norm": 0.5023863315582275,
|
| 430 |
+
"learning_rate": 0.0003193777227622898,
|
| 431 |
+
"loss": 0.3916081237792969,
|
| 432 |
+
"mean_token_accuracy": 0.8697139009833336,
|
| 433 |
+
"num_tokens": 5888694.0,
|
| 434 |
+
"step": 1900
|
| 435 |
+
},
|
| 436 |
+
{
|
| 437 |
+
"entropy": 0.44005885019898416,
|
| 438 |
+
"epoch": 4.556399766218585,
|
| 439 |
+
"grad_norm": 0.455997109413147,
|
| 440 |
+
"learning_rate": 0.000310402233652564,
|
| 441 |
+
"loss": 0.3955466842651367,
|
| 442 |
+
"mean_token_accuracy": 0.8693471103906631,
|
| 443 |
+
"num_tokens": 6042655.0,
|
| 444 |
+
"step": 1950
|
| 445 |
+
},
|
| 446 |
+
{
|
| 447 |
+
"entropy": 0.444269048422575,
|
| 448 |
+
"epoch": 4.673290473407365,
|
| 449 |
+
"grad_norm": 0.4542011320590973,
|
| 450 |
+
"learning_rate": 0.00030130035042769316,
|
| 451 |
+
"loss": 0.3988466262817383,
|
| 452 |
+
"mean_token_accuracy": 0.8680117425322532,
|
| 453 |
+
"num_tokens": 6203980.0,
|
| 454 |
+
"step": 2000
|
| 455 |
+
},
|
| 456 |
+
{
|
| 457 |
+
"entropy": 0.43520899042487143,
|
| 458 |
+
"epoch": 4.790181180596143,
|
| 459 |
+
"grad_norm": 0.49051445722579956,
|
| 460 |
+
"learning_rate": 0.0002920872065490688,
|
| 461 |
+
"loss": 0.39819797515869143,
|
| 462 |
+
"mean_token_accuracy": 0.8678940117359162,
|
| 463 |
+
"num_tokens": 6354604.0,
|
| 464 |
+
"step": 2050
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"entropy": 0.429456724524498,
|
| 468 |
+
"epoch": 4.907071887784921,
|
| 469 |
+
"grad_norm": 0.5267980098724365,
|
| 470 |
+
"learning_rate": 0.0002827781204682396,
|
| 471 |
+
"loss": 0.39413917541503907,
|
| 472 |
+
"mean_token_accuracy": 0.870051506459713,
|
| 473 |
+
"num_tokens": 6507484.0,
|
| 474 |
+
"step": 2100
|
| 475 |
+
},
|
| 476 |
+
{
|
| 477 |
+
"epoch": 5.0,
|
| 478 |
+
"eval_entropy": 0.5826305385146823,
|
| 479 |
+
"eval_loss": 0.7694042921066284,
|
| 480 |
+
"eval_mean_token_accuracy": 0.7950637824349589,
|
| 481 |
+
"eval_num_tokens": 6628215.0,
|
| 482 |
+
"eval_runtime": 38.5641,
|
| 483 |
+
"eval_samples_per_second": 31.921,
|
| 484 |
+
"eval_steps_per_second": 3.993,
|
| 485 |
+
"step": 2140
|
| 486 |
+
},
|
| 487 |
+
{
|
| 488 |
+
"entropy": 0.41423329688496324,
|
| 489 |
+
"epoch": 5.023378141437756,
|
| 490 |
+
"grad_norm": 0.6393507719039917,
|
| 491 |
+
"learning_rate": 0.0002733885701573256,
|
| 492 |
+
"loss": 0.3733938598632813,
|
| 493 |
+
"mean_token_accuracy": 0.8762160557598325,
|
| 494 |
+
"num_tokens": 6662015.0,
|
| 495 |
+
"step": 2150
|
| 496 |
+
},
|
| 497 |
+
{
|
| 498 |
+
"entropy": 0.32116023637354374,
|
| 499 |
+
"epoch": 5.140268848626534,
|
| 500 |
+
"grad_norm": 0.5128306150436401,
|
| 501 |
+
"learning_rate": 0.00026393416737420213,
|
| 502 |
+
"loss": 0.2751211166381836,
|
| 503 |
+
"mean_token_accuracy": 0.9052674892544746,
|
| 504 |
+
"num_tokens": 6817989.0,
|
| 505 |
+
"step": 2200
|
| 506 |
+
},
|
| 507 |
+
{
|
| 508 |
+
"entropy": 0.33186347484588624,
|
| 509 |
+
"epoch": 5.257159555815313,
|
| 510 |
+
"grad_norm": 0.6007080078125,
|
| 511 |
+
"learning_rate": 0.0002544306317052408,
|
| 512 |
+
"loss": 0.2850212097167969,
|
| 513 |
+
"mean_token_accuracy": 0.9016379952430725,
|
| 514 |
+
"num_tokens": 6977021.0,
|
| 515 |
+
"step": 2250
|
| 516 |
+
},
|
| 517 |
+
{
|
| 518 |
+
"entropy": 0.3474555689096451,
|
| 519 |
+
"epoch": 5.374050263004091,
|
| 520 |
+
"grad_norm": 0.6231185793876648,
|
| 521 |
+
"learning_rate": 0.0002448937644287679,
|
| 522 |
+
"loss": 0.29555959701538087,
|
| 523 |
+
"mean_token_accuracy": 0.8966628012061119,
|
| 524 |
+
"num_tokens": 7133360.0,
|
| 525 |
+
"step": 2300
|
| 526 |
+
},
|
| 527 |
+
{
|
| 528 |
+
"entropy": 0.346554354429245,
|
| 529 |
+
"epoch": 5.490940970192869,
|
| 530 |
+
"grad_norm": 0.5146846175193787,
|
| 531 |
+
"learning_rate": 0.0002353394222426952,
|
| 532 |
+
"loss": 0.29524572372436525,
|
| 533 |
+
"mean_token_accuracy": 0.8978548383712769,
|
| 534 |
+
"num_tokens": 7282372.0,
|
| 535 |
+
"step": 2350
|
| 536 |
+
},
|
| 537 |
+
{
|
| 538 |
+
"entropy": 0.3467242659628391,
|
| 539 |
+
"epoch": 5.607831677381649,
|
| 540 |
+
"grad_norm": 0.6118131875991821,
|
| 541 |
+
"learning_rate": 0.00022578349090000624,
|
| 542 |
+
"loss": 0.2970552635192871,
|
| 543 |
+
"mean_token_accuracy": 0.8985717830061912,
|
| 544 |
+
"num_tokens": 7435766.0,
|
| 545 |
+
"step": 2400
|
| 546 |
+
},
|
| 547 |
+
{
|
| 548 |
+
"entropy": 0.34968974225223065,
|
| 549 |
+
"epoch": 5.724722384570427,
|
| 550 |
+
"grad_norm": 0.6346496939659119,
|
| 551 |
+
"learning_rate": 0.0002162418587959333,
|
| 552 |
+
"loss": 0.29940626144409177,
|
| 553 |
+
"mean_token_accuracy": 0.8972864350676537,
|
| 554 |
+
"num_tokens": 7578408.0,
|
| 555 |
+
"step": 2450
|
| 556 |
+
},
|
| 557 |
+
{
|
| 558 |
+
"entropy": 0.35084108062088487,
|
| 559 |
+
"epoch": 5.841613091759205,
|
| 560 |
+
"grad_norm": 0.6566210389137268,
|
| 561 |
+
"learning_rate": 0.00020673039055074148,
|
| 562 |
+
"loss": 0.302426700592041,
|
| 563 |
+
"mean_token_accuracy": 0.8953999072313309,
|
| 564 |
+
"num_tokens": 7734599.0,
|
| 565 |
+
"step": 2500
|
| 566 |
+
},
|
| 567 |
+
{
|
| 568 |
+
"entropy": 0.33895430400967597,
|
| 569 |
+
"epoch": 5.958503798947984,
|
| 570 |
+
"grad_norm": 0.517574667930603,
|
| 571 |
+
"learning_rate": 0.00019726490063204247,
|
| 572 |
+
"loss": 0.292838020324707,
|
| 573 |
+
"mean_token_accuracy": 0.8984868070483207,
|
| 574 |
+
"num_tokens": 7898028.0,
|
| 575 |
+
"step": 2550
|
| 576 |
+
},
|
| 577 |
+
{
|
| 578 |
+
"epoch": 6.0,
|
| 579 |
+
"eval_entropy": 0.5211211553254684,
|
| 580 |
+
"eval_loss": 0.8269560933113098,
|
| 581 |
+
"eval_mean_token_accuracy": 0.7899065927251593,
|
| 582 |
+
"eval_num_tokens": 7953858.0,
|
| 583 |
+
"eval_runtime": 39.369,
|
| 584 |
+
"eval_samples_per_second": 31.268,
|
| 585 |
+
"eval_steps_per_second": 3.912,
|
| 586 |
+
"step": 2568
|
| 587 |
+
},
|
| 588 |
+
{
|
| 589 |
+
"entropy": 0.2763742347009218,
|
| 590 |
+
"epoch": 6.074810052600818,
|
| 591 |
+
"grad_norm": 0.5116816163063049,
|
| 592 |
+
"learning_rate": 0.00018786112706049623,
|
| 593 |
+
"loss": 0.22524795532226563,
|
| 594 |
+
"mean_token_accuracy": 0.9218554844209297,
|
| 595 |
+
"num_tokens": 8052646.0,
|
| 596 |
+
"step": 2600
|
| 597 |
+
},
|
| 598 |
+
{
|
| 599 |
+
"entropy": 0.23341606348752975,
|
| 600 |
+
"epoch": 6.1917007597895966,
|
| 601 |
+
"grad_norm": 0.6335962414741516,
|
| 602 |
+
"learning_rate": 0.00017853470524261842,
|
| 603 |
+
"loss": 0.18153846740722657,
|
| 604 |
+
"mean_token_accuracy": 0.935965863764286,
|
| 605 |
+
"num_tokens": 8212323.0,
|
| 606 |
+
"step": 2650
|
| 607 |
+
},
|
| 608 |
+
{
|
| 609 |
+
"entropy": 0.23515729174017908,
|
| 610 |
+
"epoch": 6.308591466978375,
|
| 611 |
+
"grad_norm": 0.6312636733055115,
|
| 612 |
+
"learning_rate": 0.0001693011419742025,
|
| 613 |
+
"loss": 0.18562854766845704,
|
| 614 |
+
"mean_token_accuracy": 0.9342144966125489,
|
| 615 |
+
"num_tokens": 8366715.0,
|
| 616 |
+
"step": 2700
|
| 617 |
+
},
|
| 618 |
+
{
|
| 619 |
+
"entropy": 0.23566059060394765,
|
| 620 |
+
"epoch": 6.425482174167154,
|
| 621 |
+
"grad_norm": 0.5869485139846802,
|
| 622 |
+
"learning_rate": 0.0001601757896575784,
|
| 623 |
+
"loss": 0.1867989158630371,
|
| 624 |
+
"mean_token_accuracy": 0.934665755033493,
|
| 625 |
+
"num_tokens": 8515955.0,
|
| 626 |
+
"step": 2750
|
| 627 |
+
},
|
| 628 |
+
{
|
| 629 |
+
"entropy": 0.2446490554511547,
|
| 630 |
+
"epoch": 6.5423728813559325,
|
| 631 |
+
"grad_norm": 0.5957017540931702,
|
| 632 |
+
"learning_rate": 0.00015117382077557817,
|
| 633 |
+
"loss": 0.19312274932861329,
|
| 634 |
+
"mean_token_accuracy": 0.9308532625436783,
|
| 635 |
+
"num_tokens": 8672361.0,
|
| 636 |
+
"step": 2800
|
| 637 |
+
},
|
| 638 |
+
{
|
| 639 |
+
"entropy": 0.24066772796213626,
|
| 640 |
+
"epoch": 6.659263588544711,
|
| 641 |
+
"grad_norm": 0.6069761514663696,
|
| 642 |
+
"learning_rate": 0.0001423102026646483,
|
| 643 |
+
"loss": 0.19540096282958985,
|
| 644 |
+
"mean_token_accuracy": 0.9300534284114838,
|
| 645 |
+
"num_tokens": 8827819.0,
|
| 646 |
+
"step": 2850
|
| 647 |
+
},
|
| 648 |
+
{
|
| 649 |
+
"entropy": 0.2378876845538616,
|
| 650 |
+
"epoch": 6.776154295733489,
|
| 651 |
+
"grad_norm": 0.6069548726081848,
|
| 652 |
+
"learning_rate": 0.00013359967262905405,
|
| 653 |
+
"loss": 0.1942250633239746,
|
| 654 |
+
"mean_token_accuracy": 0.9310428243875504,
|
| 655 |
+
"num_tokens": 8979535.0,
|
| 656 |
+
"step": 2900
|
| 657 |
+
},
|
| 658 |
+
{
|
| 659 |
+
"entropy": 0.23810458809137344,
|
| 660 |
+
"epoch": 6.893045002922268,
|
| 661 |
+
"grad_norm": 0.5807433128356934,
|
| 662 |
+
"learning_rate": 0.00012505671343755173,
|
| 663 |
+
"loss": 0.1908321762084961,
|
| 664 |
+
"mean_token_accuracy": 0.9318096882104874,
|
| 665 |
+
"num_tokens": 9134134.0,
|
| 666 |
+
"step": 2950
|
| 667 |
+
},
|
| 668 |
+
{
|
| 669 |
+
"epoch": 7.0,
|
| 670 |
+
"eval_entropy": 0.4395190301266583,
|
| 671 |
+
"eval_loss": 0.9185124635696411,
|
| 672 |
+
"eval_mean_token_accuracy": 0.7936263409527865,
|
| 673 |
+
"eval_num_tokens": 9279501.0,
|
| 674 |
+
"eval_runtime": 37.9924,
|
| 675 |
+
"eval_samples_per_second": 32.401,
|
| 676 |
+
"eval_steps_per_second": 4.053,
|
| 677 |
+
"step": 2996
|
| 678 |
+
},
|
| 679 |
+
{
|
| 680 |
+
"entropy": 0.22885651856511083,
|
| 681 |
+
"epoch": 7.009351256575102,
|
| 682 |
+
"grad_norm": 0.49437037110328674,
|
| 683 |
+
"learning_rate": 0.00011669552924327112,
|
| 684 |
+
"loss": 0.18458213806152343,
|
| 685 |
+
"mean_token_accuracy": 0.9342338658457425,
|
| 686 |
+
"num_tokens": 9290416.0,
|
| 687 |
+
"step": 3000
|
| 688 |
+
},
|
| 689 |
+
{
|
| 690 |
+
"entropy": 0.1516158328205347,
|
| 691 |
+
"epoch": 7.1262419637638805,
|
| 692 |
+
"grad_norm": 0.5319057703018188,
|
| 693 |
+
"learning_rate": 0.00010853002196684399,
|
| 694 |
+
"loss": 0.10484781265258789,
|
| 695 |
+
"mean_token_accuracy": 0.9638857007026672,
|
| 696 |
+
"num_tokens": 9444705.0,
|
| 697 |
+
"step": 3050
|
| 698 |
+
},
|
| 699 |
+
{
|
| 700 |
+
"entropy": 0.14421835087239743,
|
| 701 |
+
"epoch": 7.243132670952659,
|
| 702 |
+
"grad_norm": 0.7959536910057068,
|
| 703 |
+
"learning_rate": 0.00010057376818204711,
|
| 704 |
+
"loss": 0.10359338760375976,
|
| 705 |
+
"mean_token_accuracy": 0.964658388197422,
|
| 706 |
+
"num_tokens": 9601476.0,
|
| 707 |
+
"step": 3100
|
| 708 |
+
},
|
| 709 |
+
{
|
| 710 |
+
"entropy": 0.14138914909213782,
|
| 711 |
+
"epoch": 7.360023378141438,
|
| 712 |
+
"grad_norm": 0.5152567028999329,
|
| 713 |
+
"learning_rate": 9.283999654239098e-05,
|
| 714 |
+
"loss": 0.10241118431091309,
|
| 715 |
+
"mean_token_accuracy": 0.9644565668702125,
|
| 716 |
+
"num_tokens": 9757913.0,
|
| 717 |
+
"step": 3150
|
| 718 |
+
},
|
| 719 |
+
{
|
| 720 |
+
"entropy": 0.14779060527682306,
|
| 721 |
+
"epoch": 7.4769140853302165,
|
| 722 |
+
"grad_norm": 0.5124489068984985,
|
| 723 |
+
"learning_rate": 8.534156578618579e-05,
|
| 724 |
+
"loss": 0.10744266510009766,
|
| 725 |
+
"mean_token_accuracy": 0.9628658410906792,
|
| 726 |
+
"num_tokens": 9905790.0,
|
| 727 |
+
"step": 3200
|
| 728 |
+
},
|
| 729 |
+
{
|
| 730 |
+
"entropy": 0.1453312272951007,
|
| 731 |
+
"epoch": 7.593804792518995,
|
| 732 |
+
"grad_norm": 0.4871680438518524,
|
| 733 |
+
"learning_rate": 7.809094335665658e-05,
|
| 734 |
+
"loss": 0.10566988945007325,
|
| 735 |
+
"mean_token_accuracy": 0.9635729387402534,
|
| 736 |
+
"num_tokens": 10064234.0,
|
| 737 |
+
"step": 3250
|
| 738 |
+
},
|
| 739 |
+
{
|
| 740 |
+
"entropy": 0.14601945489645005,
|
| 741 |
+
"epoch": 7.710695499707773,
|
| 742 |
+
"grad_norm": 0.5538403987884521,
|
| 743 |
+
"learning_rate": 7.110018467265364e-05,
|
| 744 |
+
"loss": 0.10567939758300782,
|
| 745 |
+
"mean_token_accuracy": 0.9625948050618172,
|
| 746 |
+
"num_tokens": 10221089.0,
|
| 747 |
+
"step": 3300
|
| 748 |
+
},
|
| 749 |
+
{
|
| 750 |
+
"entropy": 0.14106836922466756,
|
| 751 |
+
"epoch": 7.827586206896552,
|
| 752 |
+
"grad_norm": 0.4848695695400238,
|
| 753 |
+
"learning_rate": 6.438091308442527e-05,
|
| 754 |
+
"loss": 0.10246496200561524,
|
| 755 |
+
"mean_token_accuracy": 0.9643726572394371,
|
| 756 |
+
"num_tokens": 10379655.0,
|
| 757 |
+
"step": 3350
|
| 758 |
+
},
|
| 759 |
+
{
|
| 760 |
+
"entropy": 0.14235207289457322,
|
| 761 |
+
"epoch": 7.94447691408533,
|
| 762 |
+
"grad_norm": 0.5310181379318237,
|
| 763 |
+
"learning_rate": 5.7944300547779826e-05,
|
| 764 |
+
"loss": 0.10354788780212403,
|
| 765 |
+
"mean_token_accuracy": 0.9633960571885108,
|
| 766 |
+
"num_tokens": 10531301.0,
|
| 767 |
+
"step": 3400
|
| 768 |
+
},
|
| 769 |
+
{
|
| 770 |
+
"epoch": 8.0,
|
| 771 |
+
"eval_entropy": 0.3818505619253431,
|
| 772 |
+
"eval_loss": 1.0742957592010498,
|
| 773 |
+
"eval_mean_token_accuracy": 0.7910223340059256,
|
| 774 |
+
"eval_num_tokens": 10605144.0,
|
| 775 |
+
"eval_runtime": 38.0708,
|
| 776 |
+
"eval_samples_per_second": 32.335,
|
| 777 |
+
"eval_steps_per_second": 4.045,
|
| 778 |
+
"step": 3424
|
| 779 |
+
}
|
| 780 |
+
],
|
| 781 |
+
"logging_steps": 50,
|
| 782 |
+
"max_steps": 4280,
|
| 783 |
+
"num_input_tokens_seen": 0,
|
| 784 |
+
"num_train_epochs": 10,
|
| 785 |
+
"save_steps": 500,
|
| 786 |
+
"stateful_callbacks": {
|
| 787 |
+
"TrainerControl": {
|
| 788 |
+
"args": {
|
| 789 |
+
"should_epoch_stop": false,
|
| 790 |
+
"should_evaluate": false,
|
| 791 |
+
"should_log": false,
|
| 792 |
+
"should_save": true,
|
| 793 |
+
"should_training_stop": false
|
| 794 |
+
},
|
| 795 |
+
"attributes": {}
|
| 796 |
+
}
|
| 797 |
+
},
|
| 798 |
+
"total_flos": 1.589156466300795e+17,
|
| 799 |
+
"train_batch_size": 4,
|
| 800 |
+
"trial_name": null,
|
| 801 |
+
"trial_params": null
|
| 802 |
+
}
|
DBCA_original_Estonian/Qwen3.5-2B-Base_original_features_structural_train_original_features_structural_test1/checkpoint-3852/README.md
ADDED
|
@@ -0,0 +1,209 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
base_model: Qwen/Qwen3.5-2B-Base
|
| 3 |
+
library_name: peft
|
| 4 |
+
pipeline_tag: text-generation
|
| 5 |
+
tags:
|
| 6 |
+
- base_model:adapter:Qwen/Qwen3.5-2B-Base
|
| 7 |
+
- lora
|
| 8 |
+
- sft
|
| 9 |
+
- transformers
|
| 10 |
+
- trl
|
| 11 |
+
---
|
| 12 |
+
|
| 13 |
+
# Model Card for Model ID
|
| 14 |
+
|
| 15 |
+
<!-- Provide a quick summary of what the model is/does. -->
|
| 16 |
+
|
| 17 |
+
|
| 18 |
+
|
| 19 |
+
## Model Details
|
| 20 |
+
|
| 21 |
+
### Model Description
|
| 22 |
+
|
| 23 |
+
<!-- Provide a longer summary of what this model is. -->
|
| 24 |
+
|
| 25 |
+
|
| 26 |
+
|
| 27 |
+
- **Developed by:** [More Information Needed]
|
| 28 |
+
- **Funded by [optional]:** [More Information Needed]
|
| 29 |
+
- **Shared by [optional]:** [More Information Needed]
|
| 30 |
+
- **Model type:** [More Information Needed]
|
| 31 |
+
- **Language(s) (NLP):** [More Information Needed]
|
| 32 |
+
- **License:** [More Information Needed]
|
| 33 |
+
- **Finetuned from model [optional]:** [More Information Needed]
|
| 34 |
+
|
| 35 |
+
### Model Sources [optional]
|
| 36 |
+
|
| 37 |
+
<!-- Provide the basic links for the model. -->
|
| 38 |
+
|
| 39 |
+
- **Repository:** [More Information Needed]
|
| 40 |
+
- **Paper [optional]:** [More Information Needed]
|
| 41 |
+
- **Demo [optional]:** [More Information Needed]
|
| 42 |
+
|
| 43 |
+
## Uses
|
| 44 |
+
|
| 45 |
+
<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
|
| 46 |
+
|
| 47 |
+
### Direct Use
|
| 48 |
+
|
| 49 |
+
<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
|
| 50 |
+
|
| 51 |
+
[More Information Needed]
|
| 52 |
+
|
| 53 |
+
### Downstream Use [optional]
|
| 54 |
+
|
| 55 |
+
<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
|
| 56 |
+
|
| 57 |
+
[More Information Needed]
|
| 58 |
+
|
| 59 |
+
### Out-of-Scope Use
|
| 60 |
+
|
| 61 |
+
<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
|
| 62 |
+
|
| 63 |
+
[More Information Needed]
|
| 64 |
+
|
| 65 |
+
## Bias, Risks, and Limitations
|
| 66 |
+
|
| 67 |
+
<!-- This section is meant to convey both technical and sociotechnical limitations. -->
|
| 68 |
+
|
| 69 |
+
[More Information Needed]
|
| 70 |
+
|
| 71 |
+
### Recommendations
|
| 72 |
+
|
| 73 |
+
<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
|
| 74 |
+
|
| 75 |
+
Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
|
| 76 |
+
|
| 77 |
+
## How to Get Started with the Model
|
| 78 |
+
|
| 79 |
+
Use the code below to get started with the model.
|
| 80 |
+
|
| 81 |
+
[More Information Needed]
|
| 82 |
+
|
| 83 |
+
## Training Details
|
| 84 |
+
|
| 85 |
+
### Training Data
|
| 86 |
+
|
| 87 |
+
<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
|
| 88 |
+
|
| 89 |
+
[More Information Needed]
|
| 90 |
+
|
| 91 |
+
### Training Procedure
|
| 92 |
+
|
| 93 |
+
<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
|
| 94 |
+
|
| 95 |
+
#### Preprocessing [optional]
|
| 96 |
+
|
| 97 |
+
[More Information Needed]
|
| 98 |
+
|
| 99 |
+
|
| 100 |
+
#### Training Hyperparameters
|
| 101 |
+
|
| 102 |
+
- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
|
| 103 |
+
|
| 104 |
+
#### Speeds, Sizes, Times [optional]
|
| 105 |
+
|
| 106 |
+
<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
|
| 107 |
+
|
| 108 |
+
[More Information Needed]
|
| 109 |
+
|
| 110 |
+
## Evaluation
|
| 111 |
+
|
| 112 |
+
<!-- This section describes the evaluation protocols and provides the results. -->
|
| 113 |
+
|
| 114 |
+
### Testing Data, Factors & Metrics
|
| 115 |
+
|
| 116 |
+
#### Testing Data
|
| 117 |
+
|
| 118 |
+
<!-- This should link to a Dataset Card if possible. -->
|
| 119 |
+
|
| 120 |
+
[More Information Needed]
|
| 121 |
+
|
| 122 |
+
#### Factors
|
| 123 |
+
|
| 124 |
+
<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
|
| 125 |
+
|
| 126 |
+
[More Information Needed]
|
| 127 |
+
|
| 128 |
+
#### Metrics
|
| 129 |
+
|
| 130 |
+
<!-- These are the evaluation metrics being used, ideally with a description of why. -->
|
| 131 |
+
|
| 132 |
+
[More Information Needed]
|
| 133 |
+
|
| 134 |
+
### Results
|
| 135 |
+
|
| 136 |
+
[More Information Needed]
|
| 137 |
+
|
| 138 |
+
#### Summary
|
| 139 |
+
|
| 140 |
+
|
| 141 |
+
|
| 142 |
+
## Model Examination [optional]
|
| 143 |
+
|
| 144 |
+
<!-- Relevant interpretability work for the model goes here -->
|
| 145 |
+
|
| 146 |
+
[More Information Needed]
|
| 147 |
+
|
| 148 |
+
## Environmental Impact
|
| 149 |
+
|
| 150 |
+
<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
|
| 151 |
+
|
| 152 |
+
Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
|
| 153 |
+
|
| 154 |
+
- **Hardware Type:** [More Information Needed]
|
| 155 |
+
- **Hours used:** [More Information Needed]
|
| 156 |
+
- **Cloud Provider:** [More Information Needed]
|
| 157 |
+
- **Compute Region:** [More Information Needed]
|
| 158 |
+
- **Carbon Emitted:** [More Information Needed]
|
| 159 |
+
|
| 160 |
+
## Technical Specifications [optional]
|
| 161 |
+
|
| 162 |
+
### Model Architecture and Objective
|
| 163 |
+
|
| 164 |
+
[More Information Needed]
|
| 165 |
+
|
| 166 |
+
### Compute Infrastructure
|
| 167 |
+
|
| 168 |
+
[More Information Needed]
|
| 169 |
+
|
| 170 |
+
#### Hardware
|
| 171 |
+
|
| 172 |
+
[More Information Needed]
|
| 173 |
+
|
| 174 |
+
#### Software
|
| 175 |
+
|
| 176 |
+
[More Information Needed]
|
| 177 |
+
|
| 178 |
+
## Citation [optional]
|
| 179 |
+
|
| 180 |
+
<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
|
| 181 |
+
|
| 182 |
+
**BibTeX:**
|
| 183 |
+
|
| 184 |
+
[More Information Needed]
|
| 185 |
+
|
| 186 |
+
**APA:**
|
| 187 |
+
|
| 188 |
+
[More Information Needed]
|
| 189 |
+
|
| 190 |
+
## Glossary [optional]
|
| 191 |
+
|
| 192 |
+
<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
|
| 193 |
+
|
| 194 |
+
[More Information Needed]
|
| 195 |
+
|
| 196 |
+
## More Information [optional]
|
| 197 |
+
|
| 198 |
+
[More Information Needed]
|
| 199 |
+
|
| 200 |
+
## Model Card Authors [optional]
|
| 201 |
+
|
| 202 |
+
[More Information Needed]
|
| 203 |
+
|
| 204 |
+
## Model Card Contact
|
| 205 |
+
|
| 206 |
+
[More Information Needed]
|
| 207 |
+
### Framework versions
|
| 208 |
+
|
| 209 |
+
- PEFT 0.19.1
|
DBCA_original_Estonian/Qwen3.5-2B-Base_original_features_structural_train_original_features_structural_test1/checkpoint-3852/adapter_config.json
ADDED
|
@@ -0,0 +1,48 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"alora_invocation_tokens": null,
|
| 3 |
+
"alpha_pattern": {},
|
| 4 |
+
"arrow_config": null,
|
| 5 |
+
"auto_mapping": null,
|
| 6 |
+
"base_model_name_or_path": "Qwen/Qwen3.5-2B-Base",
|
| 7 |
+
"bias": "none",
|
| 8 |
+
"corda_config": null,
|
| 9 |
+
"ensure_weight_tying": false,
|
| 10 |
+
"eva_config": null,
|
| 11 |
+
"exclude_modules": null,
|
| 12 |
+
"fan_in_fan_out": false,
|
| 13 |
+
"inference_mode": true,
|
| 14 |
+
"init_lora_weights": true,
|
| 15 |
+
"layer_replication": null,
|
| 16 |
+
"layers_pattern": null,
|
| 17 |
+
"layers_to_transform": null,
|
| 18 |
+
"loftq_config": {},
|
| 19 |
+
"lora_alpha": 64,
|
| 20 |
+
"lora_bias": false,
|
| 21 |
+
"lora_dropout": 0.04180832058159435,
|
| 22 |
+
"lora_ga_config": null,
|
| 23 |
+
"megatron_config": null,
|
| 24 |
+
"megatron_core": "megatron.core",
|
| 25 |
+
"modules_to_save": null,
|
| 26 |
+
"peft_type": "LORA",
|
| 27 |
+
"peft_version": "0.19.1",
|
| 28 |
+
"qalora_group_size": 16,
|
| 29 |
+
"r": 64,
|
| 30 |
+
"rank_pattern": {},
|
| 31 |
+
"revision": null,
|
| 32 |
+
"target_modules": [
|
| 33 |
+
"v_proj",
|
| 34 |
+
"o_proj",
|
| 35 |
+
"q_proj",
|
| 36 |
+
"gate_proj",
|
| 37 |
+
"k_proj",
|
| 38 |
+
"down_proj",
|
| 39 |
+
"up_proj"
|
| 40 |
+
],
|
| 41 |
+
"target_parameters": null,
|
| 42 |
+
"task_type": "CAUSAL_LM",
|
| 43 |
+
"trainable_token_indices": null,
|
| 44 |
+
"use_bdlora": null,
|
| 45 |
+
"use_dora": false,
|
| 46 |
+
"use_qalora": false,
|
| 47 |
+
"use_rslora": false
|
| 48 |
+
}
|
DBCA_original_Estonian/Qwen3.5-2B-Base_original_features_structural_train_original_features_structural_test1/checkpoint-4280/tokenizer_config.json
ADDED
|
@@ -0,0 +1,31 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"add_prefix_space": false,
|
| 3 |
+
"audio_bos_token": "<|audio_start|>",
|
| 4 |
+
"audio_eos_token": "<|audio_end|>",
|
| 5 |
+
"audio_token": "<|audio_pad|>",
|
| 6 |
+
"backend": "tokenizers",
|
| 7 |
+
"bos_token": null,
|
| 8 |
+
"clean_up_tokenization_spaces": false,
|
| 9 |
+
"eos_token": "<|endoftext|>",
|
| 10 |
+
"errors": "replace",
|
| 11 |
+
"image_token": "<|image_pad|>",
|
| 12 |
+
"is_local": false,
|
| 13 |
+
"model_max_length": 262144,
|
| 14 |
+
"model_specific_special_tokens": {
|
| 15 |
+
"audio_bos_token": "<|audio_start|>",
|
| 16 |
+
"audio_eos_token": "<|audio_end|>",
|
| 17 |
+
"audio_token": "<|audio_pad|>",
|
| 18 |
+
"image_token": "<|image_pad|>",
|
| 19 |
+
"video_token": "<|video_pad|>",
|
| 20 |
+
"vision_bos_token": "<|vision_start|>",
|
| 21 |
+
"vision_eos_token": "<|vision_end|>"
|
| 22 |
+
},
|
| 23 |
+
"pad_token": "<|endoftext|>",
|
| 24 |
+
"pretokenize_regex": "(?i:'s|'t|'re|'ve|'m|'ll|'d)|[^\\r\\n\\p{L}\\p{N}]?[\\p{L}\\p{M}]+|\\p{N}| ?[^\\s\\p{L}\\p{M}\\p{N}]+[\\r\\n]*|\\s*[\\r\\n]+|\\s+(?!\\S)|\\s+",
|
| 25 |
+
"split_special_tokens": false,
|
| 26 |
+
"tokenizer_class": "TokenizersBackend",
|
| 27 |
+
"unk_token": null,
|
| 28 |
+
"video_token": "<|video_pad|>",
|
| 29 |
+
"vision_bos_token": "<|vision_start|>",
|
| 30 |
+
"vision_eos_token": "<|vision_end|>"
|
| 31 |
+
}
|
substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/README.md
ADDED
|
@@ -0,0 +1,58 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
base_model: Qwen/Qwen3.5-2B-Base
|
| 3 |
+
library_name: transformers
|
| 4 |
+
model_name: Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1
|
| 5 |
+
tags:
|
| 6 |
+
- generated_from_trainer
|
| 7 |
+
- sft
|
| 8 |
+
- trl
|
| 9 |
+
licence: license
|
| 10 |
+
---
|
| 11 |
+
|
| 12 |
+
# Model Card for Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1
|
| 13 |
+
|
| 14 |
+
This model is a fine-tuned version of [Qwen/Qwen3.5-2B-Base](https://huggingface.co/Qwen/Qwen3.5-2B-Base).
|
| 15 |
+
It has been trained using [TRL](https://github.com/huggingface/trl).
|
| 16 |
+
|
| 17 |
+
## Quick start
|
| 18 |
+
|
| 19 |
+
```python
|
| 20 |
+
from transformers import pipeline
|
| 21 |
+
|
| 22 |
+
question = "If you had a time machine, but could only go to the past or the future once and never return, which would you choose and why?"
|
| 23 |
+
generator = pipeline("text-generation", model="None", device="cuda")
|
| 24 |
+
output = generator([{"role": "user", "content": question}], max_new_tokens=128, return_full_text=False)[0]
|
| 25 |
+
print(output["generated_text"])
|
| 26 |
+
```
|
| 27 |
+
|
| 28 |
+
## Training procedure
|
| 29 |
+
|
| 30 |
+
[<img src="https://raw.githubusercontent.com/wandb/assets/main/wandb-github-badge-28.svg" alt="Visualize in Weights & Biases" width="150" height="24"/>](https://wandb.ai/katriin-kukk/Cross_lingual_morphological_generalization/runs/9asazldh)
|
| 31 |
+
|
| 32 |
+
|
| 33 |
+
|
| 34 |
+
This model was trained with SFT.
|
| 35 |
+
|
| 36 |
+
### Framework versions
|
| 37 |
+
|
| 38 |
+
- TRL: 0.29.0
|
| 39 |
+
- Transformers: 5.5.4
|
| 40 |
+
- Pytorch: 2.10.0
|
| 41 |
+
- Datasets: 4.6.1
|
| 42 |
+
- Tokenizers: 0.22.2
|
| 43 |
+
|
| 44 |
+
## Citations
|
| 45 |
+
|
| 46 |
+
|
| 47 |
+
|
| 48 |
+
Cite TRL as:
|
| 49 |
+
|
| 50 |
+
```bibtex
|
| 51 |
+
@software{vonwerra2020trl,
|
| 52 |
+
title = {{TRL: Transformers Reinforcement Learning}},
|
| 53 |
+
author = {von Werra, Leandro and Belkada, Younes and Tunstall, Lewis and Beeching, Edward and Thrush, Tristan and Lambert, Nathan and Huang, Shengyi and Rasul, Kashif and Gallouédec, Quentin},
|
| 54 |
+
license = {Apache-2.0},
|
| 55 |
+
url = {https://github.com/huggingface/trl},
|
| 56 |
+
year = {2020}
|
| 57 |
+
}
|
| 58 |
+
```
|
substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-1224/README.md
ADDED
|
@@ -0,0 +1,209 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
base_model: Qwen/Qwen3.5-2B-Base
|
| 3 |
+
library_name: peft
|
| 4 |
+
pipeline_tag: text-generation
|
| 5 |
+
tags:
|
| 6 |
+
- base_model:adapter:Qwen/Qwen3.5-2B-Base
|
| 7 |
+
- lora
|
| 8 |
+
- sft
|
| 9 |
+
- transformers
|
| 10 |
+
- trl
|
| 11 |
+
---
|
| 12 |
+
|
| 13 |
+
# Model Card for Model ID
|
| 14 |
+
|
| 15 |
+
<!-- Provide a quick summary of what the model is/does. -->
|
| 16 |
+
|
| 17 |
+
|
| 18 |
+
|
| 19 |
+
## Model Details
|
| 20 |
+
|
| 21 |
+
### Model Description
|
| 22 |
+
|
| 23 |
+
<!-- Provide a longer summary of what this model is. -->
|
| 24 |
+
|
| 25 |
+
|
| 26 |
+
|
| 27 |
+
- **Developed by:** [More Information Needed]
|
| 28 |
+
- **Funded by [optional]:** [More Information Needed]
|
| 29 |
+
- **Shared by [optional]:** [More Information Needed]
|
| 30 |
+
- **Model type:** [More Information Needed]
|
| 31 |
+
- **Language(s) (NLP):** [More Information Needed]
|
| 32 |
+
- **License:** [More Information Needed]
|
| 33 |
+
- **Finetuned from model [optional]:** [More Information Needed]
|
| 34 |
+
|
| 35 |
+
### Model Sources [optional]
|
| 36 |
+
|
| 37 |
+
<!-- Provide the basic links for the model. -->
|
| 38 |
+
|
| 39 |
+
- **Repository:** [More Information Needed]
|
| 40 |
+
- **Paper [optional]:** [More Information Needed]
|
| 41 |
+
- **Demo [optional]:** [More Information Needed]
|
| 42 |
+
|
| 43 |
+
## Uses
|
| 44 |
+
|
| 45 |
+
<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
|
| 46 |
+
|
| 47 |
+
### Direct Use
|
| 48 |
+
|
| 49 |
+
<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
|
| 50 |
+
|
| 51 |
+
[More Information Needed]
|
| 52 |
+
|
| 53 |
+
### Downstream Use [optional]
|
| 54 |
+
|
| 55 |
+
<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
|
| 56 |
+
|
| 57 |
+
[More Information Needed]
|
| 58 |
+
|
| 59 |
+
### Out-of-Scope Use
|
| 60 |
+
|
| 61 |
+
<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
|
| 62 |
+
|
| 63 |
+
[More Information Needed]
|
| 64 |
+
|
| 65 |
+
## Bias, Risks, and Limitations
|
| 66 |
+
|
| 67 |
+
<!-- This section is meant to convey both technical and sociotechnical limitations. -->
|
| 68 |
+
|
| 69 |
+
[More Information Needed]
|
| 70 |
+
|
| 71 |
+
### Recommendations
|
| 72 |
+
|
| 73 |
+
<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
|
| 74 |
+
|
| 75 |
+
Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
|
| 76 |
+
|
| 77 |
+
## How to Get Started with the Model
|
| 78 |
+
|
| 79 |
+
Use the code below to get started with the model.
|
| 80 |
+
|
| 81 |
+
[More Information Needed]
|
| 82 |
+
|
| 83 |
+
## Training Details
|
| 84 |
+
|
| 85 |
+
### Training Data
|
| 86 |
+
|
| 87 |
+
<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
|
| 88 |
+
|
| 89 |
+
[More Information Needed]
|
| 90 |
+
|
| 91 |
+
### Training Procedure
|
| 92 |
+
|
| 93 |
+
<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
|
| 94 |
+
|
| 95 |
+
#### Preprocessing [optional]
|
| 96 |
+
|
| 97 |
+
[More Information Needed]
|
| 98 |
+
|
| 99 |
+
|
| 100 |
+
#### Training Hyperparameters
|
| 101 |
+
|
| 102 |
+
- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
|
| 103 |
+
|
| 104 |
+
#### Speeds, Sizes, Times [optional]
|
| 105 |
+
|
| 106 |
+
<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
|
| 107 |
+
|
| 108 |
+
[More Information Needed]
|
| 109 |
+
|
| 110 |
+
## Evaluation
|
| 111 |
+
|
| 112 |
+
<!-- This section describes the evaluation protocols and provides the results. -->
|
| 113 |
+
|
| 114 |
+
### Testing Data, Factors & Metrics
|
| 115 |
+
|
| 116 |
+
#### Testing Data
|
| 117 |
+
|
| 118 |
+
<!-- This should link to a Dataset Card if possible. -->
|
| 119 |
+
|
| 120 |
+
[More Information Needed]
|
| 121 |
+
|
| 122 |
+
#### Factors
|
| 123 |
+
|
| 124 |
+
<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
|
| 125 |
+
|
| 126 |
+
[More Information Needed]
|
| 127 |
+
|
| 128 |
+
#### Metrics
|
| 129 |
+
|
| 130 |
+
<!-- These are the evaluation metrics being used, ideally with a description of why. -->
|
| 131 |
+
|
| 132 |
+
[More Information Needed]
|
| 133 |
+
|
| 134 |
+
### Results
|
| 135 |
+
|
| 136 |
+
[More Information Needed]
|
| 137 |
+
|
| 138 |
+
#### Summary
|
| 139 |
+
|
| 140 |
+
|
| 141 |
+
|
| 142 |
+
## Model Examination [optional]
|
| 143 |
+
|
| 144 |
+
<!-- Relevant interpretability work for the model goes here -->
|
| 145 |
+
|
| 146 |
+
[More Information Needed]
|
| 147 |
+
|
| 148 |
+
## Environmental Impact
|
| 149 |
+
|
| 150 |
+
<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
|
| 151 |
+
|
| 152 |
+
Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
|
| 153 |
+
|
| 154 |
+
- **Hardware Type:** [More Information Needed]
|
| 155 |
+
- **Hours used:** [More Information Needed]
|
| 156 |
+
- **Cloud Provider:** [More Information Needed]
|
| 157 |
+
- **Compute Region:** [More Information Needed]
|
| 158 |
+
- **Carbon Emitted:** [More Information Needed]
|
| 159 |
+
|
| 160 |
+
## Technical Specifications [optional]
|
| 161 |
+
|
| 162 |
+
### Model Architecture and Objective
|
| 163 |
+
|
| 164 |
+
[More Information Needed]
|
| 165 |
+
|
| 166 |
+
### Compute Infrastructure
|
| 167 |
+
|
| 168 |
+
[More Information Needed]
|
| 169 |
+
|
| 170 |
+
#### Hardware
|
| 171 |
+
|
| 172 |
+
[More Information Needed]
|
| 173 |
+
|
| 174 |
+
#### Software
|
| 175 |
+
|
| 176 |
+
[More Information Needed]
|
| 177 |
+
|
| 178 |
+
## Citation [optional]
|
| 179 |
+
|
| 180 |
+
<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
|
| 181 |
+
|
| 182 |
+
**BibTeX:**
|
| 183 |
+
|
| 184 |
+
[More Information Needed]
|
| 185 |
+
|
| 186 |
+
**APA:**
|
| 187 |
+
|
| 188 |
+
[More Information Needed]
|
| 189 |
+
|
| 190 |
+
## Glossary [optional]
|
| 191 |
+
|
| 192 |
+
<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
|
| 193 |
+
|
| 194 |
+
[More Information Needed]
|
| 195 |
+
|
| 196 |
+
## More Information [optional]
|
| 197 |
+
|
| 198 |
+
[More Information Needed]
|
| 199 |
+
|
| 200 |
+
## Model Card Authors [optional]
|
| 201 |
+
|
| 202 |
+
[More Information Needed]
|
| 203 |
+
|
| 204 |
+
## Model Card Contact
|
| 205 |
+
|
| 206 |
+
[More Information Needed]
|
| 207 |
+
### Framework versions
|
| 208 |
+
|
| 209 |
+
- PEFT 0.19.1
|
substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-1224/adapter_config.json
ADDED
|
@@ -0,0 +1,48 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"alora_invocation_tokens": null,
|
| 3 |
+
"alpha_pattern": {},
|
| 4 |
+
"arrow_config": null,
|
| 5 |
+
"auto_mapping": null,
|
| 6 |
+
"base_model_name_or_path": "Qwen/Qwen3.5-2B-Base",
|
| 7 |
+
"bias": "none",
|
| 8 |
+
"corda_config": null,
|
| 9 |
+
"ensure_weight_tying": false,
|
| 10 |
+
"eva_config": null,
|
| 11 |
+
"exclude_modules": null,
|
| 12 |
+
"fan_in_fan_out": false,
|
| 13 |
+
"inference_mode": true,
|
| 14 |
+
"init_lora_weights": true,
|
| 15 |
+
"layer_replication": null,
|
| 16 |
+
"layers_pattern": null,
|
| 17 |
+
"layers_to_transform": null,
|
| 18 |
+
"loftq_config": {},
|
| 19 |
+
"lora_alpha": 32,
|
| 20 |
+
"lora_bias": false,
|
| 21 |
+
"lora_dropout": 0.08602048083239339,
|
| 22 |
+
"lora_ga_config": null,
|
| 23 |
+
"megatron_config": null,
|
| 24 |
+
"megatron_core": "megatron.core",
|
| 25 |
+
"modules_to_save": null,
|
| 26 |
+
"peft_type": "LORA",
|
| 27 |
+
"peft_version": "0.19.1",
|
| 28 |
+
"qalora_group_size": 16,
|
| 29 |
+
"r": 16,
|
| 30 |
+
"rank_pattern": {},
|
| 31 |
+
"revision": null,
|
| 32 |
+
"target_modules": [
|
| 33 |
+
"q_proj",
|
| 34 |
+
"k_proj",
|
| 35 |
+
"gate_proj",
|
| 36 |
+
"up_proj",
|
| 37 |
+
"o_proj",
|
| 38 |
+
"down_proj",
|
| 39 |
+
"v_proj"
|
| 40 |
+
],
|
| 41 |
+
"target_parameters": null,
|
| 42 |
+
"task_type": "CAUSAL_LM",
|
| 43 |
+
"trainable_token_indices": null,
|
| 44 |
+
"use_bdlora": null,
|
| 45 |
+
"use_dora": false,
|
| 46 |
+
"use_qalora": false,
|
| 47 |
+
"use_rslora": false
|
| 48 |
+
}
|
substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-1224/chat_template.jinja
ADDED
|
@@ -0,0 +1,154 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{%- set image_count = namespace(value=0) %}
|
| 2 |
+
{%- set video_count = namespace(value=0) %}
|
| 3 |
+
{%- macro render_content(content, do_vision_count, is_system_content=false) %}
|
| 4 |
+
{%- if content is string %}
|
| 5 |
+
{{- content }}
|
| 6 |
+
{%- elif content is iterable and content is not mapping %}
|
| 7 |
+
{%- for item in content %}
|
| 8 |
+
{%- if 'image' in item or 'image_url' in item or item.type == 'image' %}
|
| 9 |
+
{%- if is_system_content %}
|
| 10 |
+
{{- raise_exception('System message cannot contain images.') }}
|
| 11 |
+
{%- endif %}
|
| 12 |
+
{%- if do_vision_count %}
|
| 13 |
+
{%- set image_count.value = image_count.value + 1 %}
|
| 14 |
+
{%- endif %}
|
| 15 |
+
{%- if add_vision_id %}
|
| 16 |
+
{{- 'Picture ' ~ image_count.value ~ ': ' }}
|
| 17 |
+
{%- endif %}
|
| 18 |
+
{{- '<|vision_start|><|image_pad|><|vision_end|>' }}
|
| 19 |
+
{%- elif 'video' in item or item.type == 'video' %}
|
| 20 |
+
{%- if is_system_content %}
|
| 21 |
+
{{- raise_exception('System message cannot contain videos.') }}
|
| 22 |
+
{%- endif %}
|
| 23 |
+
{%- if do_vision_count %}
|
| 24 |
+
{%- set video_count.value = video_count.value + 1 %}
|
| 25 |
+
{%- endif %}
|
| 26 |
+
{%- if add_vision_id %}
|
| 27 |
+
{{- 'Video ' ~ video_count.value ~ ': ' }}
|
| 28 |
+
{%- endif %}
|
| 29 |
+
{{- '<|vision_start|><|video_pad|><|vision_end|>' }}
|
| 30 |
+
{%- elif 'text' in item %}
|
| 31 |
+
{{- item.text }}
|
| 32 |
+
{%- else %}
|
| 33 |
+
{{- raise_exception('Unexpected item type in content.') }}
|
| 34 |
+
{%- endif %}
|
| 35 |
+
{%- endfor %}
|
| 36 |
+
{%- elif content is none or content is undefined %}
|
| 37 |
+
{{- '' }}
|
| 38 |
+
{%- else %}
|
| 39 |
+
{{- raise_exception('Unexpected content type.') }}
|
| 40 |
+
{%- endif %}
|
| 41 |
+
{%- endmacro %}
|
| 42 |
+
{%- if not messages %}
|
| 43 |
+
{{- raise_exception('No messages provided.') }}
|
| 44 |
+
{%- endif %}
|
| 45 |
+
{%- if tools and tools is iterable and tools is not mapping %}
|
| 46 |
+
{{- '<|im_start|>system\n' }}
|
| 47 |
+
{{- "# Tools\n\nYou have access to the following functions:\n\n<tools>" }}
|
| 48 |
+
{%- for tool in tools %}
|
| 49 |
+
{{- "\n" }}
|
| 50 |
+
{{- tool | tojson }}
|
| 51 |
+
{%- endfor %}
|
| 52 |
+
{{- "\n</tools>" }}
|
| 53 |
+
{{- '\n\nIf you choose to call a function ONLY reply in the following format with NO suffix:\n\n<tool_call>\n<function=example_function_name>\n<parameter=example_parameter_1>\nvalue_1\n</parameter>\n<parameter=example_parameter_2>\nThis is the value for the second parameter\nthat can span\nmultiple lines\n</parameter>\n</function>\n</tool_call>\n\n<IMPORTANT>\nReminder:\n- Function calls MUST follow the specified format: an inner <function=...></function> block must be nested within <tool_call></tool_call> XML tags\n- Required parameters MUST be specified\n- You may provide optional reasoning for your function call in natural language BEFORE the function call, but NOT after\n- If there is no function call available, answer the question like normal with your current knowledge and do not tell the user about function calls\n</IMPORTANT>' }}
|
| 54 |
+
{%- if messages[0].role == 'system' %}
|
| 55 |
+
{%- set content = render_content(messages[0].content, false, true)|trim %}
|
| 56 |
+
{%- if content %}
|
| 57 |
+
{{- '\n\n' + content }}
|
| 58 |
+
{%- endif %}
|
| 59 |
+
{%- endif %}
|
| 60 |
+
{{- '<|im_end|>\n' }}
|
| 61 |
+
{%- else %}
|
| 62 |
+
{%- if messages[0].role == 'system' %}
|
| 63 |
+
{%- set content = render_content(messages[0].content, false, true)|trim %}
|
| 64 |
+
{{- '<|im_start|>system\n' + content + '<|im_end|>\n' }}
|
| 65 |
+
{%- endif %}
|
| 66 |
+
{%- endif %}
|
| 67 |
+
{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
|
| 68 |
+
{%- for message in messages[::-1] %}
|
| 69 |
+
{%- set index = (messages|length - 1) - loop.index0 %}
|
| 70 |
+
{%- if ns.multi_step_tool and message.role == "user" %}
|
| 71 |
+
{%- set content = render_content(message.content, false)|trim %}
|
| 72 |
+
{%- if not(content.startswith('<tool_response>') and content.endswith('</tool_response>')) %}
|
| 73 |
+
{%- set ns.multi_step_tool = false %}
|
| 74 |
+
{%- set ns.last_query_index = index %}
|
| 75 |
+
{%- endif %}
|
| 76 |
+
{%- endif %}
|
| 77 |
+
{%- endfor %}
|
| 78 |
+
{%- if ns.multi_step_tool %}
|
| 79 |
+
{{- raise_exception('No user query found in messages.') }}
|
| 80 |
+
{%- endif %}
|
| 81 |
+
{%- for message in messages %}
|
| 82 |
+
{%- set content = render_content(message.content, true)|trim %}
|
| 83 |
+
{%- if message.role == "system" %}
|
| 84 |
+
{%- if not loop.first %}
|
| 85 |
+
{{- raise_exception('System message must be at the beginning.') }}
|
| 86 |
+
{%- endif %}
|
| 87 |
+
{%- elif message.role == "user" %}
|
| 88 |
+
{{- '<|im_start|>' + message.role + '\n' + content + '<|im_end|>' + '\n' }}
|
| 89 |
+
{%- elif message.role == "assistant" %}
|
| 90 |
+
{%- set reasoning_content = '' %}
|
| 91 |
+
{%- if message.reasoning_content is string %}
|
| 92 |
+
{%- set reasoning_content = message.reasoning_content %}
|
| 93 |
+
{%- else %}
|
| 94 |
+
{%- if '</think>' in content %}
|
| 95 |
+
{%- set reasoning_content = content.split('</think>')[0].rstrip('\n').split('<think>')[-1].lstrip('\n') %}
|
| 96 |
+
{%- set content = content.split('</think>')[-1].lstrip('\n') %}
|
| 97 |
+
{%- endif %}
|
| 98 |
+
{%- endif %}
|
| 99 |
+
{%- set reasoning_content = reasoning_content|trim %}
|
| 100 |
+
{%- if loop.index0 > ns.last_query_index %}
|
| 101 |
+
{{- '<|im_start|>' + message.role + '\n<think>\n' + reasoning_content + '\n</think>\n\n' + content }}
|
| 102 |
+
{%- else %}
|
| 103 |
+
{{- '<|im_start|>' + message.role + '\n' + content }}
|
| 104 |
+
{%- endif %}
|
| 105 |
+
{%- if message.tool_calls and message.tool_calls is iterable and message.tool_calls is not mapping %}
|
| 106 |
+
{%- for tool_call in message.tool_calls %}
|
| 107 |
+
{%- if tool_call.function is defined %}
|
| 108 |
+
{%- set tool_call = tool_call.function %}
|
| 109 |
+
{%- endif %}
|
| 110 |
+
{%- if loop.first %}
|
| 111 |
+
{%- if content|trim %}
|
| 112 |
+
{{- '\n\n<tool_call>\n<function=' + tool_call.name + '>\n' }}
|
| 113 |
+
{%- else %}
|
| 114 |
+
{{- '<tool_call>\n<function=' + tool_call.name + '>\n' }}
|
| 115 |
+
{%- endif %}
|
| 116 |
+
{%- else %}
|
| 117 |
+
{{- '\n<tool_call>\n<function=' + tool_call.name + '>\n' }}
|
| 118 |
+
{%- endif %}
|
| 119 |
+
{%- if tool_call.arguments is defined %}
|
| 120 |
+
{%- for args_name, args_value in tool_call.arguments|items %}
|
| 121 |
+
{{- '<parameter=' + args_name + '>\n' }}
|
| 122 |
+
{%- set args_value = args_value | tojson | safe if args_value is mapping or (args_value is sequence and args_value is not string) else args_value | string %}
|
| 123 |
+
{{- args_value }}
|
| 124 |
+
{{- '\n</parameter>\n' }}
|
| 125 |
+
{%- endfor %}
|
| 126 |
+
{%- endif %}
|
| 127 |
+
{{- '</function>\n</tool_call>' }}
|
| 128 |
+
{%- endfor %}
|
| 129 |
+
{%- endif %}
|
| 130 |
+
{{- '<|im_end|>\n' }}
|
| 131 |
+
{%- elif message.role == "tool" %}
|
| 132 |
+
{%- if loop.previtem and loop.previtem.role != "tool" %}
|
| 133 |
+
{{- '<|im_start|>user' }}
|
| 134 |
+
{%- endif %}
|
| 135 |
+
{{- '\n<tool_response>\n' }}
|
| 136 |
+
{{- content }}
|
| 137 |
+
{{- '\n</tool_response>' }}
|
| 138 |
+
{%- if not loop.last and loop.nextitem.role != "tool" %}
|
| 139 |
+
{{- '<|im_end|>\n' }}
|
| 140 |
+
{%- elif loop.last %}
|
| 141 |
+
{{- '<|im_end|>\n' }}
|
| 142 |
+
{%- endif %}
|
| 143 |
+
{%- else %}
|
| 144 |
+
{{- raise_exception('Unexpected message role.') }}
|
| 145 |
+
{%- endif %}
|
| 146 |
+
{%- endfor %}
|
| 147 |
+
{%- if add_generation_prompt %}
|
| 148 |
+
{{- '<|im_start|>assistant\n' }}
|
| 149 |
+
{%- if enable_thinking is defined and enable_thinking is true %}
|
| 150 |
+
{{- '<think>\n' }}
|
| 151 |
+
{%- else %}
|
| 152 |
+
{{- '<think>\n\n</think>\n\n' }}
|
| 153 |
+
{%- endif %}
|
| 154 |
+
{%- endif %}
|
substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-1224/tokenizer_config.json
ADDED
|
@@ -0,0 +1,31 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"add_prefix_space": false,
|
| 3 |
+
"audio_bos_token": "<|audio_start|>",
|
| 4 |
+
"audio_eos_token": "<|audio_end|>",
|
| 5 |
+
"audio_token": "<|audio_pad|>",
|
| 6 |
+
"backend": "tokenizers",
|
| 7 |
+
"bos_token": null,
|
| 8 |
+
"clean_up_tokenization_spaces": false,
|
| 9 |
+
"eos_token": "<|endoftext|>",
|
| 10 |
+
"errors": "replace",
|
| 11 |
+
"image_token": "<|image_pad|>",
|
| 12 |
+
"is_local": false,
|
| 13 |
+
"model_max_length": 262144,
|
| 14 |
+
"model_specific_special_tokens": {
|
| 15 |
+
"audio_bos_token": "<|audio_start|>",
|
| 16 |
+
"audio_eos_token": "<|audio_end|>",
|
| 17 |
+
"audio_token": "<|audio_pad|>",
|
| 18 |
+
"image_token": "<|image_pad|>",
|
| 19 |
+
"video_token": "<|video_pad|>",
|
| 20 |
+
"vision_bos_token": "<|vision_start|>",
|
| 21 |
+
"vision_eos_token": "<|vision_end|>"
|
| 22 |
+
},
|
| 23 |
+
"pad_token": "<|endoftext|>",
|
| 24 |
+
"pretokenize_regex": "(?i:'s|'t|'re|'ve|'m|'ll|'d)|[^\\r\\n\\p{L}\\p{N}]?[\\p{L}\\p{M}]+|\\p{N}| ?[^\\s\\p{L}\\p{M}\\p{N}]+[\\r\\n]*|\\s*[\\r\\n]+|\\s+(?!\\S)|\\s+",
|
| 25 |
+
"split_special_tokens": false,
|
| 26 |
+
"tokenizer_class": "TokenizersBackend",
|
| 27 |
+
"unk_token": null,
|
| 28 |
+
"video_token": "<|video_pad|>",
|
| 29 |
+
"vision_bos_token": "<|vision_start|>",
|
| 30 |
+
"vision_eos_token": "<|vision_end|>"
|
| 31 |
+
}
|
substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-1224/trainer_state.json
ADDED
|
@@ -0,0 +1,340 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": null,
|
| 3 |
+
"best_metric": null,
|
| 4 |
+
"best_model_checkpoint": null,
|
| 5 |
+
"epoch": 3.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 1224,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"entropy": 2.2408065378665922,
|
| 14 |
+
"epoch": 0.12269938650306748,
|
| 15 |
+
"grad_norm": 1.6718658208847046,
|
| 16 |
+
"learning_rate": 4.487804970912036e-05,
|
| 17 |
+
"loss": 2.129624786376953,
|
| 18 |
+
"mean_token_accuracy": 0.5743644836544991,
|
| 19 |
+
"num_tokens": 134329.0,
|
| 20 |
+
"step": 50
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"entropy": 1.1078504550457,
|
| 24 |
+
"epoch": 0.24539877300613497,
|
| 25 |
+
"grad_norm": 2.4091813564300537,
|
| 26 |
+
"learning_rate": 9.067197798373296e-05,
|
| 27 |
+
"loss": 1.0346390533447265,
|
| 28 |
+
"mean_token_accuracy": 0.7085381114482879,
|
| 29 |
+
"num_tokens": 262407.0,
|
| 30 |
+
"step": 100
|
| 31 |
+
},
|
| 32 |
+
{
|
| 33 |
+
"entropy": 0.9421598988771439,
|
| 34 |
+
"epoch": 0.36809815950920244,
|
| 35 |
+
"grad_norm": 1.3913646936416626,
|
| 36 |
+
"learning_rate": 0.00013646590625834558,
|
| 37 |
+
"loss": 0.8716551208496094,
|
| 38 |
+
"mean_token_accuracy": 0.7449040985107422,
|
| 39 |
+
"num_tokens": 401136.0,
|
| 40 |
+
"step": 150
|
| 41 |
+
},
|
| 42 |
+
{
|
| 43 |
+
"entropy": 0.8723550146818161,
|
| 44 |
+
"epoch": 0.49079754601226994,
|
| 45 |
+
"grad_norm": 1.1664468050003052,
|
| 46 |
+
"learning_rate": 0.0001822598345329582,
|
| 47 |
+
"loss": 0.8095977783203125,
|
| 48 |
+
"mean_token_accuracy": 0.7585240352153778,
|
| 49 |
+
"num_tokens": 531647.0,
|
| 50 |
+
"step": 200
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"entropy": 0.8466519421339035,
|
| 54 |
+
"epoch": 0.6134969325153374,
|
| 55 |
+
"grad_norm": 1.0145421028137207,
|
| 56 |
+
"learning_rate": 0.00022805376280757083,
|
| 57 |
+
"loss": 0.7830857849121093,
|
| 58 |
+
"mean_token_accuracy": 0.7661705195903779,
|
| 59 |
+
"num_tokens": 666127.0,
|
| 60 |
+
"step": 250
|
| 61 |
+
},
|
| 62 |
+
{
|
| 63 |
+
"entropy": 0.8180295366048813,
|
| 64 |
+
"epoch": 0.7361963190184049,
|
| 65 |
+
"grad_norm": 0.9731705784797668,
|
| 66 |
+
"learning_rate": 0.00027384769108218336,
|
| 67 |
+
"loss": 0.7573603820800782,
|
| 68 |
+
"mean_token_accuracy": 0.7704453605413437,
|
| 69 |
+
"num_tokens": 799292.0,
|
| 70 |
+
"step": 300
|
| 71 |
+
},
|
| 72 |
+
{
|
| 73 |
+
"entropy": 0.7990227991342544,
|
| 74 |
+
"epoch": 0.8588957055214724,
|
| 75 |
+
"grad_norm": 1.0642374753952026,
|
| 76 |
+
"learning_rate": 0.000319641619356796,
|
| 77 |
+
"loss": 0.73709228515625,
|
| 78 |
+
"mean_token_accuracy": 0.7752733880281448,
|
| 79 |
+
"num_tokens": 930614.0,
|
| 80 |
+
"step": 350
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"entropy": 0.7899087685346603,
|
| 84 |
+
"epoch": 0.9815950920245399,
|
| 85 |
+
"grad_norm": 0.880613386631012,
|
| 86 |
+
"learning_rate": 0.0003654355476314086,
|
| 87 |
+
"loss": 0.7347319793701171,
|
| 88 |
+
"mean_token_accuracy": 0.7766731631755829,
|
| 89 |
+
"num_tokens": 1061654.0,
|
| 90 |
+
"step": 400
|
| 91 |
+
},
|
| 92 |
+
{
|
| 93 |
+
"epoch": 1.0,
|
| 94 |
+
"eval_entropy": 0.83522472347532,
|
| 95 |
+
"eval_mean_token_accuracy": 0.7484098076820374,
|
| 96 |
+
"eval_not_syn_loss": 0.7872886061668396,
|
| 97 |
+
"eval_not_syn_runtime": 54.8697,
|
| 98 |
+
"eval_not_syn_samples_per_second": 25.424,
|
| 99 |
+
"eval_not_syn_steps_per_second": 3.189,
|
| 100 |
+
"eval_num_tokens": 1080196.0,
|
| 101 |
+
"step": 408
|
| 102 |
+
},
|
| 103 |
+
{
|
| 104 |
+
"epoch": 1.0,
|
| 105 |
+
"eval_entropy": 0.7784442271505083,
|
| 106 |
+
"eval_mean_token_accuracy": 0.8021791185651507,
|
| 107 |
+
"eval_num_tokens": 1080196.0,
|
| 108 |
+
"eval_syn_loss": 0.6980098485946655,
|
| 109 |
+
"eval_syn_runtime": 56.6342,
|
| 110 |
+
"eval_syn_samples_per_second": 24.632,
|
| 111 |
+
"eval_syn_steps_per_second": 3.09,
|
| 112 |
+
"step": 408
|
| 113 |
+
},
|
| 114 |
+
{
|
| 115 |
+
"entropy": 0.75672573694075,
|
| 116 |
+
"epoch": 1.1030674846625768,
|
| 117 |
+
"grad_norm": 0.9244198203086853,
|
| 118 |
+
"learning_rate": 0.00037356351883456205,
|
| 119 |
+
"loss": 0.7048745727539063,
|
| 120 |
+
"mean_token_accuracy": 0.783016896609104,
|
| 121 |
+
"num_tokens": 1199170.0,
|
| 122 |
+
"step": 450
|
| 123 |
+
},
|
| 124 |
+
{
|
| 125 |
+
"entropy": 0.7406797379255294,
|
| 126 |
+
"epoch": 1.2257668711656442,
|
| 127 |
+
"grad_norm": 0.6744896769523621,
|
| 128 |
+
"learning_rate": 0.00037311248151765587,
|
| 129 |
+
"loss": 0.6860730743408203,
|
| 130 |
+
"mean_token_accuracy": 0.787369327545166,
|
| 131 |
+
"num_tokens": 1334298.0,
|
| 132 |
+
"step": 500
|
| 133 |
+
},
|
| 134 |
+
{
|
| 135 |
+
"entropy": 0.7424973237514496,
|
| 136 |
+
"epoch": 1.3484662576687116,
|
| 137 |
+
"grad_norm": 0.92305588722229,
|
| 138 |
+
"learning_rate": 0.000372320629215228,
|
| 139 |
+
"loss": 0.6855724334716797,
|
| 140 |
+
"mean_token_accuracy": 0.7871134179830551,
|
| 141 |
+
"num_tokens": 1463189.0,
|
| 142 |
+
"step": 550
|
| 143 |
+
},
|
| 144 |
+
{
|
| 145 |
+
"entropy": 0.7502243828773498,
|
| 146 |
+
"epoch": 1.471165644171779,
|
| 147 |
+
"grad_norm": 0.8259047269821167,
|
| 148 |
+
"learning_rate": 0.00037118941074037944,
|
| 149 |
+
"loss": 0.6949834442138672,
|
| 150 |
+
"mean_token_accuracy": 0.7851914083957672,
|
| 151 |
+
"num_tokens": 1590310.0,
|
| 152 |
+
"step": 600
|
| 153 |
+
},
|
| 154 |
+
{
|
| 155 |
+
"entropy": 0.7233260518312454,
|
| 156 |
+
"epoch": 1.5938650306748468,
|
| 157 |
+
"grad_norm": 0.882722795009613,
|
| 158 |
+
"learning_rate": 0.00036972089582775814,
|
| 159 |
+
"loss": 0.6649005889892579,
|
| 160 |
+
"mean_token_accuracy": 0.7898879665136337,
|
| 161 |
+
"num_tokens": 1721836.0,
|
| 162 |
+
"step": 650
|
| 163 |
+
},
|
| 164 |
+
{
|
| 165 |
+
"entropy": 0.7162708967924118,
|
| 166 |
+
"epoch": 1.716564417177914,
|
| 167 |
+
"grad_norm": 0.6593027114868164,
|
| 168 |
+
"learning_rate": 0.0003679177713466678,
|
| 169 |
+
"loss": 0.6653710174560546,
|
| 170 |
+
"mean_token_accuracy": 0.7905678844451904,
|
| 171 |
+
"num_tokens": 1853507.0,
|
| 172 |
+
"step": 700
|
| 173 |
+
},
|
| 174 |
+
{
|
| 175 |
+
"entropy": 0.7150520020723343,
|
| 176 |
+
"epoch": 1.8392638036809816,
|
| 177 |
+
"grad_norm": 0.665340781211853,
|
| 178 |
+
"learning_rate": 0.0003657833363850354,
|
| 179 |
+
"loss": 0.6600938415527344,
|
| 180 |
+
"mean_token_accuracy": 0.7918649202585221,
|
| 181 |
+
"num_tokens": 1988113.0,
|
| 182 |
+
"step": 750
|
| 183 |
+
},
|
| 184 |
+
{
|
| 185 |
+
"entropy": 0.6993386310338974,
|
| 186 |
+
"epoch": 1.961963190184049,
|
| 187 |
+
"grad_norm": 0.6502187252044678,
|
| 188 |
+
"learning_rate": 0.00036332149621323294,
|
| 189 |
+
"loss": 0.6469607543945313,
|
| 190 |
+
"mean_token_accuracy": 0.7950778317451477,
|
| 191 |
+
"num_tokens": 2121314.0,
|
| 192 |
+
"step": 800
|
| 193 |
+
},
|
| 194 |
+
{
|
| 195 |
+
"epoch": 2.0,
|
| 196 |
+
"eval_entropy": 0.7315139893123082,
|
| 197 |
+
"eval_mean_token_accuracy": 0.7713504004478454,
|
| 198 |
+
"eval_not_syn_loss": 0.699893593788147,
|
| 199 |
+
"eval_not_syn_runtime": 55.4135,
|
| 200 |
+
"eval_not_syn_samples_per_second": 25.174,
|
| 201 |
+
"eval_not_syn_steps_per_second": 3.158,
|
| 202 |
+
"eval_num_tokens": 2160392.0,
|
| 203 |
+
"step": 816
|
| 204 |
+
},
|
| 205 |
+
{
|
| 206 |
+
"epoch": 2.0,
|
| 207 |
+
"eval_entropy": 0.6848451798302787,
|
| 208 |
+
"eval_mean_token_accuracy": 0.814895657471248,
|
| 209 |
+
"eval_num_tokens": 2160392.0,
|
| 210 |
+
"eval_syn_loss": 0.6415576934814453,
|
| 211 |
+
"eval_syn_runtime": 56.4735,
|
| 212 |
+
"eval_syn_samples_per_second": 24.702,
|
| 213 |
+
"eval_syn_steps_per_second": 3.099,
|
| 214 |
+
"step": 816
|
| 215 |
+
},
|
| 216 |
+
{
|
| 217 |
+
"entropy": 0.6596363644407253,
|
| 218 |
+
"epoch": 2.083435582822086,
|
| 219 |
+
"grad_norm": 0.6647700071334839,
|
| 220 |
+
"learning_rate": 0.00036053675513879666,
|
| 221 |
+
"loss": 0.599151611328125,
|
| 222 |
+
"mean_token_accuracy": 0.8071986864311527,
|
| 223 |
+
"num_tokens": 2253179.0,
|
| 224 |
+
"step": 850
|
| 225 |
+
},
|
| 226 |
+
{
|
| 227 |
+
"entropy": 0.6626429110765457,
|
| 228 |
+
"epoch": 2.2061349693251535,
|
| 229 |
+
"grad_norm": 0.7356868386268616,
|
| 230 |
+
"learning_rate": 0.00035743420826511783,
|
| 231 |
+
"loss": 0.6004195404052735,
|
| 232 |
+
"mean_token_accuracy": 0.805362731218338,
|
| 233 |
+
"num_tokens": 2379965.0,
|
| 234 |
+
"step": 900
|
| 235 |
+
},
|
| 236 |
+
{
|
| 237 |
+
"entropy": 0.6503470712900161,
|
| 238 |
+
"epoch": 2.3288343558282207,
|
| 239 |
+
"grad_norm": 1.095608115196228,
|
| 240 |
+
"learning_rate": 0.0003540195321691835,
|
| 241 |
+
"loss": 0.5951333999633789,
|
| 242 |
+
"mean_token_accuracy": 0.807324025630951,
|
| 243 |
+
"num_tokens": 2515593.0,
|
| 244 |
+
"step": 950
|
| 245 |
+
},
|
| 246 |
+
{
|
| 247 |
+
"entropy": 0.660577870607376,
|
| 248 |
+
"epoch": 2.4515337423312884,
|
| 249 |
+
"grad_norm": 0.6602179408073425,
|
| 250 |
+
"learning_rate": 0.00035029897451542325,
|
| 251 |
+
"loss": 0.6056365585327148,
|
| 252 |
+
"mean_token_accuracy": 0.8057438349723816,
|
| 253 |
+
"num_tokens": 2642790.0,
|
| 254 |
+
"step": 1000
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"entropy": 0.6555853113532066,
|
| 258 |
+
"epoch": 2.574233128834356,
|
| 259 |
+
"grad_norm": 0.7459151148796082,
|
| 260 |
+
"learning_rate": 0.00034627934262466606,
|
| 261 |
+
"loss": 0.5963204193115235,
|
| 262 |
+
"mean_token_accuracy": 0.8060598099231719,
|
| 263 |
+
"num_tokens": 2774396.0,
|
| 264 |
+
"step": 1050
|
| 265 |
+
},
|
| 266 |
+
{
|
| 267 |
+
"entropy": 0.6505853188037872,
|
| 268 |
+
"epoch": 2.6969325153374233,
|
| 269 |
+
"grad_norm": 0.6359792947769165,
|
| 270 |
+
"learning_rate": 0.00034196799101912103,
|
| 271 |
+
"loss": 0.5993065643310547,
|
| 272 |
+
"mean_token_accuracy": 0.806280387043953,
|
| 273 |
+
"num_tokens": 2912102.0,
|
| 274 |
+
"step": 1100
|
| 275 |
+
},
|
| 276 |
+
{
|
| 277 |
+
"entropy": 0.6509970253705979,
|
| 278 |
+
"epoch": 2.819631901840491,
|
| 279 |
+
"grad_norm": 0.8086174726486206,
|
| 280 |
+
"learning_rate": 0.0003373728079661707,
|
| 281 |
+
"loss": 0.6001268005371094,
|
| 282 |
+
"mean_token_accuracy": 0.806866540312767,
|
| 283 |
+
"num_tokens": 3042987.0,
|
| 284 |
+
"step": 1150
|
| 285 |
+
},
|
| 286 |
+
{
|
| 287 |
+
"entropy": 0.6404844307899475,
|
| 288 |
+
"epoch": 2.942331288343558,
|
| 289 |
+
"grad_norm": 0.7038407325744629,
|
| 290 |
+
"learning_rate": 0.0003325022010455975,
|
| 291 |
+
"loss": 0.5870526123046875,
|
| 292 |
+
"mean_token_accuracy": 0.8104586935043335,
|
| 293 |
+
"num_tokens": 3178044.0,
|
| 294 |
+
"step": 1200
|
| 295 |
+
},
|
| 296 |
+
{
|
| 297 |
+
"epoch": 3.0,
|
| 298 |
+
"eval_entropy": 0.6645996727262224,
|
| 299 |
+
"eval_mean_token_accuracy": 0.798654237134116,
|
| 300 |
+
"eval_not_syn_loss": 0.6586771011352539,
|
| 301 |
+
"eval_not_syn_runtime": 55.1088,
|
| 302 |
+
"eval_not_syn_samples_per_second": 25.314,
|
| 303 |
+
"eval_not_syn_steps_per_second": 3.176,
|
| 304 |
+
"eval_num_tokens": 3240588.0,
|
| 305 |
+
"step": 1224
|
| 306 |
+
},
|
| 307 |
+
{
|
| 308 |
+
"epoch": 3.0,
|
| 309 |
+
"eval_entropy": 0.6233607809884207,
|
| 310 |
+
"eval_mean_token_accuracy": 0.7977249867575509,
|
| 311 |
+
"eval_num_tokens": 3240588.0,
|
| 312 |
+
"eval_syn_loss": 0.6288298964500427,
|
| 313 |
+
"eval_syn_runtime": 56.6081,
|
| 314 |
+
"eval_syn_samples_per_second": 24.643,
|
| 315 |
+
"eval_syn_steps_per_second": 3.091,
|
| 316 |
+
"step": 1224
|
| 317 |
+
}
|
| 318 |
+
],
|
| 319 |
+
"logging_steps": 50,
|
| 320 |
+
"max_steps": 4080,
|
| 321 |
+
"num_input_tokens_seen": 0,
|
| 322 |
+
"num_train_epochs": 10,
|
| 323 |
+
"save_steps": 500,
|
| 324 |
+
"stateful_callbacks": {
|
| 325 |
+
"TrainerControl": {
|
| 326 |
+
"args": {
|
| 327 |
+
"should_epoch_stop": false,
|
| 328 |
+
"should_evaluate": false,
|
| 329 |
+
"should_log": false,
|
| 330 |
+
"should_save": true,
|
| 331 |
+
"should_training_stop": false
|
| 332 |
+
},
|
| 333 |
+
"attributes": {}
|
| 334 |
+
}
|
| 335 |
+
},
|
| 336 |
+
"total_flos": 6.458126379206016e+16,
|
| 337 |
+
"train_batch_size": 8,
|
| 338 |
+
"trial_name": null,
|
| 339 |
+
"trial_params": null
|
| 340 |
+
}
|
substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-1632/README.md
ADDED
|
@@ -0,0 +1,209 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
base_model: Qwen/Qwen3.5-2B-Base
|
| 3 |
+
library_name: peft
|
| 4 |
+
pipeline_tag: text-generation
|
| 5 |
+
tags:
|
| 6 |
+
- base_model:adapter:Qwen/Qwen3.5-2B-Base
|
| 7 |
+
- lora
|
| 8 |
+
- sft
|
| 9 |
+
- transformers
|
| 10 |
+
- trl
|
| 11 |
+
---
|
| 12 |
+
|
| 13 |
+
# Model Card for Model ID
|
| 14 |
+
|
| 15 |
+
<!-- Provide a quick summary of what the model is/does. -->
|
| 16 |
+
|
| 17 |
+
|
| 18 |
+
|
| 19 |
+
## Model Details
|
| 20 |
+
|
| 21 |
+
### Model Description
|
| 22 |
+
|
| 23 |
+
<!-- Provide a longer summary of what this model is. -->
|
| 24 |
+
|
| 25 |
+
|
| 26 |
+
|
| 27 |
+
- **Developed by:** [More Information Needed]
|
| 28 |
+
- **Funded by [optional]:** [More Information Needed]
|
| 29 |
+
- **Shared by [optional]:** [More Information Needed]
|
| 30 |
+
- **Model type:** [More Information Needed]
|
| 31 |
+
- **Language(s) (NLP):** [More Information Needed]
|
| 32 |
+
- **License:** [More Information Needed]
|
| 33 |
+
- **Finetuned from model [optional]:** [More Information Needed]
|
| 34 |
+
|
| 35 |
+
### Model Sources [optional]
|
| 36 |
+
|
| 37 |
+
<!-- Provide the basic links for the model. -->
|
| 38 |
+
|
| 39 |
+
- **Repository:** [More Information Needed]
|
| 40 |
+
- **Paper [optional]:** [More Information Needed]
|
| 41 |
+
- **Demo [optional]:** [More Information Needed]
|
| 42 |
+
|
| 43 |
+
## Uses
|
| 44 |
+
|
| 45 |
+
<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
|
| 46 |
+
|
| 47 |
+
### Direct Use
|
| 48 |
+
|
| 49 |
+
<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
|
| 50 |
+
|
| 51 |
+
[More Information Needed]
|
| 52 |
+
|
| 53 |
+
### Downstream Use [optional]
|
| 54 |
+
|
| 55 |
+
<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
|
| 56 |
+
|
| 57 |
+
[More Information Needed]
|
| 58 |
+
|
| 59 |
+
### Out-of-Scope Use
|
| 60 |
+
|
| 61 |
+
<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
|
| 62 |
+
|
| 63 |
+
[More Information Needed]
|
| 64 |
+
|
| 65 |
+
## Bias, Risks, and Limitations
|
| 66 |
+
|
| 67 |
+
<!-- This section is meant to convey both technical and sociotechnical limitations. -->
|
| 68 |
+
|
| 69 |
+
[More Information Needed]
|
| 70 |
+
|
| 71 |
+
### Recommendations
|
| 72 |
+
|
| 73 |
+
<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
|
| 74 |
+
|
| 75 |
+
Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
|
| 76 |
+
|
| 77 |
+
## How to Get Started with the Model
|
| 78 |
+
|
| 79 |
+
Use the code below to get started with the model.
|
| 80 |
+
|
| 81 |
+
[More Information Needed]
|
| 82 |
+
|
| 83 |
+
## Training Details
|
| 84 |
+
|
| 85 |
+
### Training Data
|
| 86 |
+
|
| 87 |
+
<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
|
| 88 |
+
|
| 89 |
+
[More Information Needed]
|
| 90 |
+
|
| 91 |
+
### Training Procedure
|
| 92 |
+
|
| 93 |
+
<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
|
| 94 |
+
|
| 95 |
+
#### Preprocessing [optional]
|
| 96 |
+
|
| 97 |
+
[More Information Needed]
|
| 98 |
+
|
| 99 |
+
|
| 100 |
+
#### Training Hyperparameters
|
| 101 |
+
|
| 102 |
+
- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
|
| 103 |
+
|
| 104 |
+
#### Speeds, Sizes, Times [optional]
|
| 105 |
+
|
| 106 |
+
<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
|
| 107 |
+
|
| 108 |
+
[More Information Needed]
|
| 109 |
+
|
| 110 |
+
## Evaluation
|
| 111 |
+
|
| 112 |
+
<!-- This section describes the evaluation protocols and provides the results. -->
|
| 113 |
+
|
| 114 |
+
### Testing Data, Factors & Metrics
|
| 115 |
+
|
| 116 |
+
#### Testing Data
|
| 117 |
+
|
| 118 |
+
<!-- This should link to a Dataset Card if possible. -->
|
| 119 |
+
|
| 120 |
+
[More Information Needed]
|
| 121 |
+
|
| 122 |
+
#### Factors
|
| 123 |
+
|
| 124 |
+
<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
|
| 125 |
+
|
| 126 |
+
[More Information Needed]
|
| 127 |
+
|
| 128 |
+
#### Metrics
|
| 129 |
+
|
| 130 |
+
<!-- These are the evaluation metrics being used, ideally with a description of why. -->
|
| 131 |
+
|
| 132 |
+
[More Information Needed]
|
| 133 |
+
|
| 134 |
+
### Results
|
| 135 |
+
|
| 136 |
+
[More Information Needed]
|
| 137 |
+
|
| 138 |
+
#### Summary
|
| 139 |
+
|
| 140 |
+
|
| 141 |
+
|
| 142 |
+
## Model Examination [optional]
|
| 143 |
+
|
| 144 |
+
<!-- Relevant interpretability work for the model goes here -->
|
| 145 |
+
|
| 146 |
+
[More Information Needed]
|
| 147 |
+
|
| 148 |
+
## Environmental Impact
|
| 149 |
+
|
| 150 |
+
<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
|
| 151 |
+
|
| 152 |
+
Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
|
| 153 |
+
|
| 154 |
+
- **Hardware Type:** [More Information Needed]
|
| 155 |
+
- **Hours used:** [More Information Needed]
|
| 156 |
+
- **Cloud Provider:** [More Information Needed]
|
| 157 |
+
- **Compute Region:** [More Information Needed]
|
| 158 |
+
- **Carbon Emitted:** [More Information Needed]
|
| 159 |
+
|
| 160 |
+
## Technical Specifications [optional]
|
| 161 |
+
|
| 162 |
+
### Model Architecture and Objective
|
| 163 |
+
|
| 164 |
+
[More Information Needed]
|
| 165 |
+
|
| 166 |
+
### Compute Infrastructure
|
| 167 |
+
|
| 168 |
+
[More Information Needed]
|
| 169 |
+
|
| 170 |
+
#### Hardware
|
| 171 |
+
|
| 172 |
+
[More Information Needed]
|
| 173 |
+
|
| 174 |
+
#### Software
|
| 175 |
+
|
| 176 |
+
[More Information Needed]
|
| 177 |
+
|
| 178 |
+
## Citation [optional]
|
| 179 |
+
|
| 180 |
+
<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
|
| 181 |
+
|
| 182 |
+
**BibTeX:**
|
| 183 |
+
|
| 184 |
+
[More Information Needed]
|
| 185 |
+
|
| 186 |
+
**APA:**
|
| 187 |
+
|
| 188 |
+
[More Information Needed]
|
| 189 |
+
|
| 190 |
+
## Glossary [optional]
|
| 191 |
+
|
| 192 |
+
<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
|
| 193 |
+
|
| 194 |
+
[More Information Needed]
|
| 195 |
+
|
| 196 |
+
## More Information [optional]
|
| 197 |
+
|
| 198 |
+
[More Information Needed]
|
| 199 |
+
|
| 200 |
+
## Model Card Authors [optional]
|
| 201 |
+
|
| 202 |
+
[More Information Needed]
|
| 203 |
+
|
| 204 |
+
## Model Card Contact
|
| 205 |
+
|
| 206 |
+
[More Information Needed]
|
| 207 |
+
### Framework versions
|
| 208 |
+
|
| 209 |
+
- PEFT 0.19.1
|
substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-1632/adapter_config.json
ADDED
|
@@ -0,0 +1,48 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"alora_invocation_tokens": null,
|
| 3 |
+
"alpha_pattern": {},
|
| 4 |
+
"arrow_config": null,
|
| 5 |
+
"auto_mapping": null,
|
| 6 |
+
"base_model_name_or_path": "Qwen/Qwen3.5-2B-Base",
|
| 7 |
+
"bias": "none",
|
| 8 |
+
"corda_config": null,
|
| 9 |
+
"ensure_weight_tying": false,
|
| 10 |
+
"eva_config": null,
|
| 11 |
+
"exclude_modules": null,
|
| 12 |
+
"fan_in_fan_out": false,
|
| 13 |
+
"inference_mode": true,
|
| 14 |
+
"init_lora_weights": true,
|
| 15 |
+
"layer_replication": null,
|
| 16 |
+
"layers_pattern": null,
|
| 17 |
+
"layers_to_transform": null,
|
| 18 |
+
"loftq_config": {},
|
| 19 |
+
"lora_alpha": 32,
|
| 20 |
+
"lora_bias": false,
|
| 21 |
+
"lora_dropout": 0.08602048083239339,
|
| 22 |
+
"lora_ga_config": null,
|
| 23 |
+
"megatron_config": null,
|
| 24 |
+
"megatron_core": "megatron.core",
|
| 25 |
+
"modules_to_save": null,
|
| 26 |
+
"peft_type": "LORA",
|
| 27 |
+
"peft_version": "0.19.1",
|
| 28 |
+
"qalora_group_size": 16,
|
| 29 |
+
"r": 16,
|
| 30 |
+
"rank_pattern": {},
|
| 31 |
+
"revision": null,
|
| 32 |
+
"target_modules": [
|
| 33 |
+
"q_proj",
|
| 34 |
+
"k_proj",
|
| 35 |
+
"gate_proj",
|
| 36 |
+
"up_proj",
|
| 37 |
+
"o_proj",
|
| 38 |
+
"down_proj",
|
| 39 |
+
"v_proj"
|
| 40 |
+
],
|
| 41 |
+
"target_parameters": null,
|
| 42 |
+
"task_type": "CAUSAL_LM",
|
| 43 |
+
"trainable_token_indices": null,
|
| 44 |
+
"use_bdlora": null,
|
| 45 |
+
"use_dora": false,
|
| 46 |
+
"use_qalora": false,
|
| 47 |
+
"use_rslora": false
|
| 48 |
+
}
|
substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/README.md
ADDED
|
@@ -0,0 +1,58 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
base_model: Qwen/Qwen3.5-2B-Base
|
| 3 |
+
library_name: transformers
|
| 4 |
+
model_name: Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2
|
| 5 |
+
tags:
|
| 6 |
+
- generated_from_trainer
|
| 7 |
+
- sft
|
| 8 |
+
- trl
|
| 9 |
+
licence: license
|
| 10 |
+
---
|
| 11 |
+
|
| 12 |
+
# Model Card for Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2
|
| 13 |
+
|
| 14 |
+
This model is a fine-tuned version of [Qwen/Qwen3.5-2B-Base](https://huggingface.co/Qwen/Qwen3.5-2B-Base).
|
| 15 |
+
It has been trained using [TRL](https://github.com/huggingface/trl).
|
| 16 |
+
|
| 17 |
+
## Quick start
|
| 18 |
+
|
| 19 |
+
```python
|
| 20 |
+
from transformers import pipeline
|
| 21 |
+
|
| 22 |
+
question = "If you had a time machine, but could only go to the past or the future once and never return, which would you choose and why?"
|
| 23 |
+
generator = pipeline("text-generation", model="None", device="cuda")
|
| 24 |
+
output = generator([{"role": "user", "content": question}], max_new_tokens=128, return_full_text=False)[0]
|
| 25 |
+
print(output["generated_text"])
|
| 26 |
+
```
|
| 27 |
+
|
| 28 |
+
## Training procedure
|
| 29 |
+
|
| 30 |
+
[<img src="https://raw.githubusercontent.com/wandb/assets/main/wandb-github-badge-28.svg" alt="Visualize in Weights & Biases" width="150" height="24"/>](https://wandb.ai/katriin-kukk/Cross_lingual_morphological_generalization/runs/ypap1lsm)
|
| 31 |
+
|
| 32 |
+
|
| 33 |
+
|
| 34 |
+
This model was trained with SFT.
|
| 35 |
+
|
| 36 |
+
### Framework versions
|
| 37 |
+
|
| 38 |
+
- TRL: 0.29.0
|
| 39 |
+
- Transformers: 5.5.4
|
| 40 |
+
- Pytorch: 2.10.0
|
| 41 |
+
- Datasets: 4.6.1
|
| 42 |
+
- Tokenizers: 0.22.2
|
| 43 |
+
|
| 44 |
+
## Citations
|
| 45 |
+
|
| 46 |
+
|
| 47 |
+
|
| 48 |
+
Cite TRL as:
|
| 49 |
+
|
| 50 |
+
```bibtex
|
| 51 |
+
@software{vonwerra2020trl,
|
| 52 |
+
title = {{TRL: Transformers Reinforcement Learning}},
|
| 53 |
+
author = {von Werra, Leandro and Belkada, Younes and Tunstall, Lewis and Beeching, Edward and Thrush, Tristan and Lambert, Nathan and Huang, Shengyi and Rasul, Kashif and Gallouédec, Quentin},
|
| 54 |
+
license = {Apache-2.0},
|
| 55 |
+
url = {https://github.com/huggingface/trl},
|
| 56 |
+
year = {2020}
|
| 57 |
+
}
|
| 58 |
+
```
|
substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-1224/README.md
ADDED
|
@@ -0,0 +1,209 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
base_model: Qwen/Qwen3.5-2B-Base
|
| 3 |
+
library_name: peft
|
| 4 |
+
pipeline_tag: text-generation
|
| 5 |
+
tags:
|
| 6 |
+
- base_model:adapter:Qwen/Qwen3.5-2B-Base
|
| 7 |
+
- lora
|
| 8 |
+
- sft
|
| 9 |
+
- transformers
|
| 10 |
+
- trl
|
| 11 |
+
---
|
| 12 |
+
|
| 13 |
+
# Model Card for Model ID
|
| 14 |
+
|
| 15 |
+
<!-- Provide a quick summary of what the model is/does. -->
|
| 16 |
+
|
| 17 |
+
|
| 18 |
+
|
| 19 |
+
## Model Details
|
| 20 |
+
|
| 21 |
+
### Model Description
|
| 22 |
+
|
| 23 |
+
<!-- Provide a longer summary of what this model is. -->
|
| 24 |
+
|
| 25 |
+
|
| 26 |
+
|
| 27 |
+
- **Developed by:** [More Information Needed]
|
| 28 |
+
- **Funded by [optional]:** [More Information Needed]
|
| 29 |
+
- **Shared by [optional]:** [More Information Needed]
|
| 30 |
+
- **Model type:** [More Information Needed]
|
| 31 |
+
- **Language(s) (NLP):** [More Information Needed]
|
| 32 |
+
- **License:** [More Information Needed]
|
| 33 |
+
- **Finetuned from model [optional]:** [More Information Needed]
|
| 34 |
+
|
| 35 |
+
### Model Sources [optional]
|
| 36 |
+
|
| 37 |
+
<!-- Provide the basic links for the model. -->
|
| 38 |
+
|
| 39 |
+
- **Repository:** [More Information Needed]
|
| 40 |
+
- **Paper [optional]:** [More Information Needed]
|
| 41 |
+
- **Demo [optional]:** [More Information Needed]
|
| 42 |
+
|
| 43 |
+
## Uses
|
| 44 |
+
|
| 45 |
+
<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
|
| 46 |
+
|
| 47 |
+
### Direct Use
|
| 48 |
+
|
| 49 |
+
<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
|
| 50 |
+
|
| 51 |
+
[More Information Needed]
|
| 52 |
+
|
| 53 |
+
### Downstream Use [optional]
|
| 54 |
+
|
| 55 |
+
<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
|
| 56 |
+
|
| 57 |
+
[More Information Needed]
|
| 58 |
+
|
| 59 |
+
### Out-of-Scope Use
|
| 60 |
+
|
| 61 |
+
<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
|
| 62 |
+
|
| 63 |
+
[More Information Needed]
|
| 64 |
+
|
| 65 |
+
## Bias, Risks, and Limitations
|
| 66 |
+
|
| 67 |
+
<!-- This section is meant to convey both technical and sociotechnical limitations. -->
|
| 68 |
+
|
| 69 |
+
[More Information Needed]
|
| 70 |
+
|
| 71 |
+
### Recommendations
|
| 72 |
+
|
| 73 |
+
<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
|
| 74 |
+
|
| 75 |
+
Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
|
| 76 |
+
|
| 77 |
+
## How to Get Started with the Model
|
| 78 |
+
|
| 79 |
+
Use the code below to get started with the model.
|
| 80 |
+
|
| 81 |
+
[More Information Needed]
|
| 82 |
+
|
| 83 |
+
## Training Details
|
| 84 |
+
|
| 85 |
+
### Training Data
|
| 86 |
+
|
| 87 |
+
<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
|
| 88 |
+
|
| 89 |
+
[More Information Needed]
|
| 90 |
+
|
| 91 |
+
### Training Procedure
|
| 92 |
+
|
| 93 |
+
<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
|
| 94 |
+
|
| 95 |
+
#### Preprocessing [optional]
|
| 96 |
+
|
| 97 |
+
[More Information Needed]
|
| 98 |
+
|
| 99 |
+
|
| 100 |
+
#### Training Hyperparameters
|
| 101 |
+
|
| 102 |
+
- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
|
| 103 |
+
|
| 104 |
+
#### Speeds, Sizes, Times [optional]
|
| 105 |
+
|
| 106 |
+
<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
|
| 107 |
+
|
| 108 |
+
[More Information Needed]
|
| 109 |
+
|
| 110 |
+
## Evaluation
|
| 111 |
+
|
| 112 |
+
<!-- This section describes the evaluation protocols and provides the results. -->
|
| 113 |
+
|
| 114 |
+
### Testing Data, Factors & Metrics
|
| 115 |
+
|
| 116 |
+
#### Testing Data
|
| 117 |
+
|
| 118 |
+
<!-- This should link to a Dataset Card if possible. -->
|
| 119 |
+
|
| 120 |
+
[More Information Needed]
|
| 121 |
+
|
| 122 |
+
#### Factors
|
| 123 |
+
|
| 124 |
+
<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
|
| 125 |
+
|
| 126 |
+
[More Information Needed]
|
| 127 |
+
|
| 128 |
+
#### Metrics
|
| 129 |
+
|
| 130 |
+
<!-- These are the evaluation metrics being used, ideally with a description of why. -->
|
| 131 |
+
|
| 132 |
+
[More Information Needed]
|
| 133 |
+
|
| 134 |
+
### Results
|
| 135 |
+
|
| 136 |
+
[More Information Needed]
|
| 137 |
+
|
| 138 |
+
#### Summary
|
| 139 |
+
|
| 140 |
+
|
| 141 |
+
|
| 142 |
+
## Model Examination [optional]
|
| 143 |
+
|
| 144 |
+
<!-- Relevant interpretability work for the model goes here -->
|
| 145 |
+
|
| 146 |
+
[More Information Needed]
|
| 147 |
+
|
| 148 |
+
## Environmental Impact
|
| 149 |
+
|
| 150 |
+
<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
|
| 151 |
+
|
| 152 |
+
Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
|
| 153 |
+
|
| 154 |
+
- **Hardware Type:** [More Information Needed]
|
| 155 |
+
- **Hours used:** [More Information Needed]
|
| 156 |
+
- **Cloud Provider:** [More Information Needed]
|
| 157 |
+
- **Compute Region:** [More Information Needed]
|
| 158 |
+
- **Carbon Emitted:** [More Information Needed]
|
| 159 |
+
|
| 160 |
+
## Technical Specifications [optional]
|
| 161 |
+
|
| 162 |
+
### Model Architecture and Objective
|
| 163 |
+
|
| 164 |
+
[More Information Needed]
|
| 165 |
+
|
| 166 |
+
### Compute Infrastructure
|
| 167 |
+
|
| 168 |
+
[More Information Needed]
|
| 169 |
+
|
| 170 |
+
#### Hardware
|
| 171 |
+
|
| 172 |
+
[More Information Needed]
|
| 173 |
+
|
| 174 |
+
#### Software
|
| 175 |
+
|
| 176 |
+
[More Information Needed]
|
| 177 |
+
|
| 178 |
+
## Citation [optional]
|
| 179 |
+
|
| 180 |
+
<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
|
| 181 |
+
|
| 182 |
+
**BibTeX:**
|
| 183 |
+
|
| 184 |
+
[More Information Needed]
|
| 185 |
+
|
| 186 |
+
**APA:**
|
| 187 |
+
|
| 188 |
+
[More Information Needed]
|
| 189 |
+
|
| 190 |
+
## Glossary [optional]
|
| 191 |
+
|
| 192 |
+
<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
|
| 193 |
+
|
| 194 |
+
[More Information Needed]
|
| 195 |
+
|
| 196 |
+
## More Information [optional]
|
| 197 |
+
|
| 198 |
+
[More Information Needed]
|
| 199 |
+
|
| 200 |
+
## Model Card Authors [optional]
|
| 201 |
+
|
| 202 |
+
[More Information Needed]
|
| 203 |
+
|
| 204 |
+
## Model Card Contact
|
| 205 |
+
|
| 206 |
+
[More Information Needed]
|
| 207 |
+
### Framework versions
|
| 208 |
+
|
| 209 |
+
- PEFT 0.19.1
|
substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-1224/adapter_config.json
ADDED
|
@@ -0,0 +1,48 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"alora_invocation_tokens": null,
|
| 3 |
+
"alpha_pattern": {},
|
| 4 |
+
"arrow_config": null,
|
| 5 |
+
"auto_mapping": null,
|
| 6 |
+
"base_model_name_or_path": "Qwen/Qwen3.5-2B-Base",
|
| 7 |
+
"bias": "none",
|
| 8 |
+
"corda_config": null,
|
| 9 |
+
"ensure_weight_tying": false,
|
| 10 |
+
"eva_config": null,
|
| 11 |
+
"exclude_modules": null,
|
| 12 |
+
"fan_in_fan_out": false,
|
| 13 |
+
"inference_mode": true,
|
| 14 |
+
"init_lora_weights": true,
|
| 15 |
+
"layer_replication": null,
|
| 16 |
+
"layers_pattern": null,
|
| 17 |
+
"layers_to_transform": null,
|
| 18 |
+
"loftq_config": {},
|
| 19 |
+
"lora_alpha": 128,
|
| 20 |
+
"lora_bias": false,
|
| 21 |
+
"lora_dropout": 0.05635041589438875,
|
| 22 |
+
"lora_ga_config": null,
|
| 23 |
+
"megatron_config": null,
|
| 24 |
+
"megatron_core": "megatron.core",
|
| 25 |
+
"modules_to_save": null,
|
| 26 |
+
"peft_type": "LORA",
|
| 27 |
+
"peft_version": "0.19.1",
|
| 28 |
+
"qalora_group_size": 16,
|
| 29 |
+
"r": 64,
|
| 30 |
+
"rank_pattern": {},
|
| 31 |
+
"revision": null,
|
| 32 |
+
"target_modules": [
|
| 33 |
+
"q_proj",
|
| 34 |
+
"k_proj",
|
| 35 |
+
"gate_proj",
|
| 36 |
+
"up_proj",
|
| 37 |
+
"o_proj",
|
| 38 |
+
"down_proj",
|
| 39 |
+
"v_proj"
|
| 40 |
+
],
|
| 41 |
+
"target_parameters": null,
|
| 42 |
+
"task_type": "CAUSAL_LM",
|
| 43 |
+
"trainable_token_indices": null,
|
| 44 |
+
"use_bdlora": null,
|
| 45 |
+
"use_dora": false,
|
| 46 |
+
"use_qalora": false,
|
| 47 |
+
"use_rslora": false
|
| 48 |
+
}
|
substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-1224/chat_template.jinja
ADDED
|
@@ -0,0 +1,154 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{%- set image_count = namespace(value=0) %}
|
| 2 |
+
{%- set video_count = namespace(value=0) %}
|
| 3 |
+
{%- macro render_content(content, do_vision_count, is_system_content=false) %}
|
| 4 |
+
{%- if content is string %}
|
| 5 |
+
{{- content }}
|
| 6 |
+
{%- elif content is iterable and content is not mapping %}
|
| 7 |
+
{%- for item in content %}
|
| 8 |
+
{%- if 'image' in item or 'image_url' in item or item.type == 'image' %}
|
| 9 |
+
{%- if is_system_content %}
|
| 10 |
+
{{- raise_exception('System message cannot contain images.') }}
|
| 11 |
+
{%- endif %}
|
| 12 |
+
{%- if do_vision_count %}
|
| 13 |
+
{%- set image_count.value = image_count.value + 1 %}
|
| 14 |
+
{%- endif %}
|
| 15 |
+
{%- if add_vision_id %}
|
| 16 |
+
{{- 'Picture ' ~ image_count.value ~ ': ' }}
|
| 17 |
+
{%- endif %}
|
| 18 |
+
{{- '<|vision_start|><|image_pad|><|vision_end|>' }}
|
| 19 |
+
{%- elif 'video' in item or item.type == 'video' %}
|
| 20 |
+
{%- if is_system_content %}
|
| 21 |
+
{{- raise_exception('System message cannot contain videos.') }}
|
| 22 |
+
{%- endif %}
|
| 23 |
+
{%- if do_vision_count %}
|
| 24 |
+
{%- set video_count.value = video_count.value + 1 %}
|
| 25 |
+
{%- endif %}
|
| 26 |
+
{%- if add_vision_id %}
|
| 27 |
+
{{- 'Video ' ~ video_count.value ~ ': ' }}
|
| 28 |
+
{%- endif %}
|
| 29 |
+
{{- '<|vision_start|><|video_pad|><|vision_end|>' }}
|
| 30 |
+
{%- elif 'text' in item %}
|
| 31 |
+
{{- item.text }}
|
| 32 |
+
{%- else %}
|
| 33 |
+
{{- raise_exception('Unexpected item type in content.') }}
|
| 34 |
+
{%- endif %}
|
| 35 |
+
{%- endfor %}
|
| 36 |
+
{%- elif content is none or content is undefined %}
|
| 37 |
+
{{- '' }}
|
| 38 |
+
{%- else %}
|
| 39 |
+
{{- raise_exception('Unexpected content type.') }}
|
| 40 |
+
{%- endif %}
|
| 41 |
+
{%- endmacro %}
|
| 42 |
+
{%- if not messages %}
|
| 43 |
+
{{- raise_exception('No messages provided.') }}
|
| 44 |
+
{%- endif %}
|
| 45 |
+
{%- if tools and tools is iterable and tools is not mapping %}
|
| 46 |
+
{{- '<|im_start|>system\n' }}
|
| 47 |
+
{{- "# Tools\n\nYou have access to the following functions:\n\n<tools>" }}
|
| 48 |
+
{%- for tool in tools %}
|
| 49 |
+
{{- "\n" }}
|
| 50 |
+
{{- tool | tojson }}
|
| 51 |
+
{%- endfor %}
|
| 52 |
+
{{- "\n</tools>" }}
|
| 53 |
+
{{- '\n\nIf you choose to call a function ONLY reply in the following format with NO suffix:\n\n<tool_call>\n<function=example_function_name>\n<parameter=example_parameter_1>\nvalue_1\n</parameter>\n<parameter=example_parameter_2>\nThis is the value for the second parameter\nthat can span\nmultiple lines\n</parameter>\n</function>\n</tool_call>\n\n<IMPORTANT>\nReminder:\n- Function calls MUST follow the specified format: an inner <function=...></function> block must be nested within <tool_call></tool_call> XML tags\n- Required parameters MUST be specified\n- You may provide optional reasoning for your function call in natural language BEFORE the function call, but NOT after\n- If there is no function call available, answer the question like normal with your current knowledge and do not tell the user about function calls\n</IMPORTANT>' }}
|
| 54 |
+
{%- if messages[0].role == 'system' %}
|
| 55 |
+
{%- set content = render_content(messages[0].content, false, true)|trim %}
|
| 56 |
+
{%- if content %}
|
| 57 |
+
{{- '\n\n' + content }}
|
| 58 |
+
{%- endif %}
|
| 59 |
+
{%- endif %}
|
| 60 |
+
{{- '<|im_end|>\n' }}
|
| 61 |
+
{%- else %}
|
| 62 |
+
{%- if messages[0].role == 'system' %}
|
| 63 |
+
{%- set content = render_content(messages[0].content, false, true)|trim %}
|
| 64 |
+
{{- '<|im_start|>system\n' + content + '<|im_end|>\n' }}
|
| 65 |
+
{%- endif %}
|
| 66 |
+
{%- endif %}
|
| 67 |
+
{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
|
| 68 |
+
{%- for message in messages[::-1] %}
|
| 69 |
+
{%- set index = (messages|length - 1) - loop.index0 %}
|
| 70 |
+
{%- if ns.multi_step_tool and message.role == "user" %}
|
| 71 |
+
{%- set content = render_content(message.content, false)|trim %}
|
| 72 |
+
{%- if not(content.startswith('<tool_response>') and content.endswith('</tool_response>')) %}
|
| 73 |
+
{%- set ns.multi_step_tool = false %}
|
| 74 |
+
{%- set ns.last_query_index = index %}
|
| 75 |
+
{%- endif %}
|
| 76 |
+
{%- endif %}
|
| 77 |
+
{%- endfor %}
|
| 78 |
+
{%- if ns.multi_step_tool %}
|
| 79 |
+
{{- raise_exception('No user query found in messages.') }}
|
| 80 |
+
{%- endif %}
|
| 81 |
+
{%- for message in messages %}
|
| 82 |
+
{%- set content = render_content(message.content, true)|trim %}
|
| 83 |
+
{%- if message.role == "system" %}
|
| 84 |
+
{%- if not loop.first %}
|
| 85 |
+
{{- raise_exception('System message must be at the beginning.') }}
|
| 86 |
+
{%- endif %}
|
| 87 |
+
{%- elif message.role == "user" %}
|
| 88 |
+
{{- '<|im_start|>' + message.role + '\n' + content + '<|im_end|>' + '\n' }}
|
| 89 |
+
{%- elif message.role == "assistant" %}
|
| 90 |
+
{%- set reasoning_content = '' %}
|
| 91 |
+
{%- if message.reasoning_content is string %}
|
| 92 |
+
{%- set reasoning_content = message.reasoning_content %}
|
| 93 |
+
{%- else %}
|
| 94 |
+
{%- if '</think>' in content %}
|
| 95 |
+
{%- set reasoning_content = content.split('</think>')[0].rstrip('\n').split('<think>')[-1].lstrip('\n') %}
|
| 96 |
+
{%- set content = content.split('</think>')[-1].lstrip('\n') %}
|
| 97 |
+
{%- endif %}
|
| 98 |
+
{%- endif %}
|
| 99 |
+
{%- set reasoning_content = reasoning_content|trim %}
|
| 100 |
+
{%- if loop.index0 > ns.last_query_index %}
|
| 101 |
+
{{- '<|im_start|>' + message.role + '\n<think>\n' + reasoning_content + '\n</think>\n\n' + content }}
|
| 102 |
+
{%- else %}
|
| 103 |
+
{{- '<|im_start|>' + message.role + '\n' + content }}
|
| 104 |
+
{%- endif %}
|
| 105 |
+
{%- if message.tool_calls and message.tool_calls is iterable and message.tool_calls is not mapping %}
|
| 106 |
+
{%- for tool_call in message.tool_calls %}
|
| 107 |
+
{%- if tool_call.function is defined %}
|
| 108 |
+
{%- set tool_call = tool_call.function %}
|
| 109 |
+
{%- endif %}
|
| 110 |
+
{%- if loop.first %}
|
| 111 |
+
{%- if content|trim %}
|
| 112 |
+
{{- '\n\n<tool_call>\n<function=' + tool_call.name + '>\n' }}
|
| 113 |
+
{%- else %}
|
| 114 |
+
{{- '<tool_call>\n<function=' + tool_call.name + '>\n' }}
|
| 115 |
+
{%- endif %}
|
| 116 |
+
{%- else %}
|
| 117 |
+
{{- '\n<tool_call>\n<function=' + tool_call.name + '>\n' }}
|
| 118 |
+
{%- endif %}
|
| 119 |
+
{%- if tool_call.arguments is defined %}
|
| 120 |
+
{%- for args_name, args_value in tool_call.arguments|items %}
|
| 121 |
+
{{- '<parameter=' + args_name + '>\n' }}
|
| 122 |
+
{%- set args_value = args_value | tojson | safe if args_value is mapping or (args_value is sequence and args_value is not string) else args_value | string %}
|
| 123 |
+
{{- args_value }}
|
| 124 |
+
{{- '\n</parameter>\n' }}
|
| 125 |
+
{%- endfor %}
|
| 126 |
+
{%- endif %}
|
| 127 |
+
{{- '</function>\n</tool_call>' }}
|
| 128 |
+
{%- endfor %}
|
| 129 |
+
{%- endif %}
|
| 130 |
+
{{- '<|im_end|>\n' }}
|
| 131 |
+
{%- elif message.role == "tool" %}
|
| 132 |
+
{%- if loop.previtem and loop.previtem.role != "tool" %}
|
| 133 |
+
{{- '<|im_start|>user' }}
|
| 134 |
+
{%- endif %}
|
| 135 |
+
{{- '\n<tool_response>\n' }}
|
| 136 |
+
{{- content }}
|
| 137 |
+
{{- '\n</tool_response>' }}
|
| 138 |
+
{%- if not loop.last and loop.nextitem.role != "tool" %}
|
| 139 |
+
{{- '<|im_end|>\n' }}
|
| 140 |
+
{%- elif loop.last %}
|
| 141 |
+
{{- '<|im_end|>\n' }}
|
| 142 |
+
{%- endif %}
|
| 143 |
+
{%- else %}
|
| 144 |
+
{{- raise_exception('Unexpected message role.') }}
|
| 145 |
+
{%- endif %}
|
| 146 |
+
{%- endfor %}
|
| 147 |
+
{%- if add_generation_prompt %}
|
| 148 |
+
{{- '<|im_start|>assistant\n' }}
|
| 149 |
+
{%- if enable_thinking is defined and enable_thinking is true %}
|
| 150 |
+
{{- '<think>\n' }}
|
| 151 |
+
{%- else %}
|
| 152 |
+
{{- '<think>\n\n</think>\n\n' }}
|
| 153 |
+
{%- endif %}
|
| 154 |
+
{%- endif %}
|
substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-1224/tokenizer_config.json
ADDED
|
@@ -0,0 +1,31 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"add_prefix_space": false,
|
| 3 |
+
"audio_bos_token": "<|audio_start|>",
|
| 4 |
+
"audio_eos_token": "<|audio_end|>",
|
| 5 |
+
"audio_token": "<|audio_pad|>",
|
| 6 |
+
"backend": "tokenizers",
|
| 7 |
+
"bos_token": null,
|
| 8 |
+
"clean_up_tokenization_spaces": false,
|
| 9 |
+
"eos_token": "<|endoftext|>",
|
| 10 |
+
"errors": "replace",
|
| 11 |
+
"image_token": "<|image_pad|>",
|
| 12 |
+
"is_local": false,
|
| 13 |
+
"model_max_length": 262144,
|
| 14 |
+
"model_specific_special_tokens": {
|
| 15 |
+
"audio_bos_token": "<|audio_start|>",
|
| 16 |
+
"audio_eos_token": "<|audio_end|>",
|
| 17 |
+
"audio_token": "<|audio_pad|>",
|
| 18 |
+
"image_token": "<|image_pad|>",
|
| 19 |
+
"video_token": "<|video_pad|>",
|
| 20 |
+
"vision_bos_token": "<|vision_start|>",
|
| 21 |
+
"vision_eos_token": "<|vision_end|>"
|
| 22 |
+
},
|
| 23 |
+
"pad_token": "<|endoftext|>",
|
| 24 |
+
"pretokenize_regex": "(?i:'s|'t|'re|'ve|'m|'ll|'d)|[^\\r\\n\\p{L}\\p{N}]?[\\p{L}\\p{M}]+|\\p{N}| ?[^\\s\\p{L}\\p{M}\\p{N}]+[\\r\\n]*|\\s*[\\r\\n]+|\\s+(?!\\S)|\\s+",
|
| 25 |
+
"split_special_tokens": false,
|
| 26 |
+
"tokenizer_class": "TokenizersBackend",
|
| 27 |
+
"unk_token": null,
|
| 28 |
+
"video_token": "<|video_pad|>",
|
| 29 |
+
"vision_bos_token": "<|vision_start|>",
|
| 30 |
+
"vision_eos_token": "<|vision_end|>"
|
| 31 |
+
}
|
substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-1224/trainer_state.json
ADDED
|
@@ -0,0 +1,340 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": null,
|
| 3 |
+
"best_metric": null,
|
| 4 |
+
"best_model_checkpoint": null,
|
| 5 |
+
"epoch": 3.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 1224,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"entropy": 1.8153233402967452,
|
| 14 |
+
"epoch": 0.12269938650306748,
|
| 15 |
+
"grad_norm": 1.8510568141937256,
|
| 16 |
+
"learning_rate": 4.8469594667113335e-05,
|
| 17 |
+
"loss": 1.7217323303222656,
|
| 18 |
+
"mean_token_accuracy": 0.6236635231971741,
|
| 19 |
+
"num_tokens": 134329.0,
|
| 20 |
+
"step": 50
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"entropy": 0.97831438601017,
|
| 24 |
+
"epoch": 0.24539877300613497,
|
| 25 |
+
"grad_norm": 1.889868974685669,
|
| 26 |
+
"learning_rate": 9.792836473559631e-05,
|
| 27 |
+
"loss": 0.9050540924072266,
|
| 28 |
+
"mean_token_accuracy": 0.7377463465929032,
|
| 29 |
+
"num_tokens": 262407.0,
|
| 30 |
+
"step": 100
|
| 31 |
+
},
|
| 32 |
+
{
|
| 33 |
+
"entropy": 0.8743131357431412,
|
| 34 |
+
"epoch": 0.36809815950920244,
|
| 35 |
+
"grad_norm": 1.1766732931137085,
|
| 36 |
+
"learning_rate": 0.0001473871348040793,
|
| 37 |
+
"loss": 0.8129763031005859,
|
| 38 |
+
"mean_token_accuracy": 0.7583417356014251,
|
| 39 |
+
"num_tokens": 401136.0,
|
| 40 |
+
"step": 150
|
| 41 |
+
},
|
| 42 |
+
{
|
| 43 |
+
"entropy": 0.8273966419696808,
|
| 44 |
+
"epoch": 0.49079754601226994,
|
| 45 |
+
"grad_norm": 1.1311014890670776,
|
| 46 |
+
"learning_rate": 0.00019684590487256229,
|
| 47 |
+
"loss": 0.7711280822753906,
|
| 48 |
+
"mean_token_accuracy": 0.7668228060007095,
|
| 49 |
+
"num_tokens": 531647.0,
|
| 50 |
+
"step": 200
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"entropy": 0.8101438617706299,
|
| 54 |
+
"epoch": 0.6134969325153374,
|
| 55 |
+
"grad_norm": 0.8905950784683228,
|
| 56 |
+
"learning_rate": 0.0002463046749410453,
|
| 57 |
+
"loss": 0.7609222412109375,
|
| 58 |
+
"mean_token_accuracy": 0.7717174577713013,
|
| 59 |
+
"num_tokens": 666127.0,
|
| 60 |
+
"step": 250
|
| 61 |
+
},
|
| 62 |
+
{
|
| 63 |
+
"entropy": 0.7885055243968964,
|
| 64 |
+
"epoch": 0.7361963190184049,
|
| 65 |
+
"grad_norm": 1.1944245100021362,
|
| 66 |
+
"learning_rate": 0.00029576344500952824,
|
| 67 |
+
"loss": 0.7477743530273437,
|
| 68 |
+
"mean_token_accuracy": 0.773262197971344,
|
| 69 |
+
"num_tokens": 799292.0,
|
| 70 |
+
"step": 300
|
| 71 |
+
},
|
| 72 |
+
{
|
| 73 |
+
"entropy": 0.7858527797460556,
|
| 74 |
+
"epoch": 0.8588957055214724,
|
| 75 |
+
"grad_norm": 1.1311068534851074,
|
| 76 |
+
"learning_rate": 0.00034522221507801124,
|
| 77 |
+
"loss": 0.7411377716064453,
|
| 78 |
+
"mean_token_accuracy": 0.7745680212974548,
|
| 79 |
+
"num_tokens": 930614.0,
|
| 80 |
+
"step": 350
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"entropy": 0.7879797518253326,
|
| 84 |
+
"epoch": 0.9815950920245399,
|
| 85 |
+
"grad_norm": 0.9370436072349548,
|
| 86 |
+
"learning_rate": 0.00039468098514649423,
|
| 87 |
+
"loss": 0.7507785034179687,
|
| 88 |
+
"mean_token_accuracy": 0.7747444450855255,
|
| 89 |
+
"num_tokens": 1061654.0,
|
| 90 |
+
"step": 400
|
| 91 |
+
},
|
| 92 |
+
{
|
| 93 |
+
"epoch": 1.0,
|
| 94 |
+
"eval_entropy": 0.8527693561145238,
|
| 95 |
+
"eval_mean_token_accuracy": 0.7730878329277039,
|
| 96 |
+
"eval_not_syn_loss": 0.781921923160553,
|
| 97 |
+
"eval_not_syn_runtime": 56.9157,
|
| 98 |
+
"eval_not_syn_samples_per_second": 24.527,
|
| 99 |
+
"eval_not_syn_steps_per_second": 3.075,
|
| 100 |
+
"eval_num_tokens": 1080196.0,
|
| 101 |
+
"step": 408
|
| 102 |
+
},
|
| 103 |
+
{
|
| 104 |
+
"epoch": 1.0,
|
| 105 |
+
"eval_entropy": 0.7952005583899362,
|
| 106 |
+
"eval_mean_token_accuracy": 0.7705784467288426,
|
| 107 |
+
"eval_num_tokens": 1080196.0,
|
| 108 |
+
"eval_syn_loss": 0.7461121082305908,
|
| 109 |
+
"eval_syn_runtime": 59.5758,
|
| 110 |
+
"eval_syn_samples_per_second": 23.432,
|
| 111 |
+
"eval_syn_steps_per_second": 2.937,
|
| 112 |
+
"step": 408
|
| 113 |
+
},
|
| 114 |
+
{
|
| 115 |
+
"entropy": 0.7564875923021875,
|
| 116 |
+
"epoch": 1.1030674846625768,
|
| 117 |
+
"grad_norm": 1.1334599256515503,
|
| 118 |
+
"learning_rate": 0.00040345942966973637,
|
| 119 |
+
"loss": 0.7208955383300781,
|
| 120 |
+
"mean_token_accuracy": 0.7788144470465304,
|
| 121 |
+
"num_tokens": 1199170.0,
|
| 122 |
+
"step": 450
|
| 123 |
+
},
|
| 124 |
+
{
|
| 125 |
+
"entropy": 0.7427570706605912,
|
| 126 |
+
"epoch": 1.2257668711656442,
|
| 127 |
+
"grad_norm": 1.0777021646499634,
|
| 128 |
+
"learning_rate": 0.00040297229629224153,
|
| 129 |
+
"loss": 0.7049168395996094,
|
| 130 |
+
"mean_token_accuracy": 0.7830300116539002,
|
| 131 |
+
"num_tokens": 1334298.0,
|
| 132 |
+
"step": 500
|
| 133 |
+
},
|
| 134 |
+
{
|
| 135 |
+
"entropy": 0.7533353108167649,
|
| 136 |
+
"epoch": 1.3484662576687116,
|
| 137 |
+
"grad_norm": 1.3439561128616333,
|
| 138 |
+
"learning_rate": 0.00040211707285041045,
|
| 139 |
+
"loss": 0.7102288818359375,
|
| 140 |
+
"mean_token_accuracy": 0.7818449640274048,
|
| 141 |
+
"num_tokens": 1463189.0,
|
| 142 |
+
"step": 550
|
| 143 |
+
},
|
| 144 |
+
{
|
| 145 |
+
"entropy": 0.7666342407464981,
|
| 146 |
+
"epoch": 1.471165644171779,
|
| 147 |
+
"grad_norm": 1.1329729557037354,
|
| 148 |
+
"learning_rate": 0.00040089532410439227,
|
| 149 |
+
"loss": 0.7270996856689453,
|
| 150 |
+
"mean_token_accuracy": 0.778404277563095,
|
| 151 |
+
"num_tokens": 1590310.0,
|
| 152 |
+
"step": 600
|
| 153 |
+
},
|
| 154 |
+
{
|
| 155 |
+
"entropy": 0.7359934949874878,
|
| 156 |
+
"epoch": 1.5938650306748468,
|
| 157 |
+
"grad_norm": 1.2227331399917603,
|
| 158 |
+
"learning_rate": 0.0003993092854276068,
|
| 159 |
+
"loss": 0.6996555328369141,
|
| 160 |
+
"mean_token_accuracy": 0.7834958267211914,
|
| 161 |
+
"num_tokens": 1721836.0,
|
| 162 |
+
"step": 650
|
| 163 |
+
},
|
| 164 |
+
{
|
| 165 |
+
"entropy": 0.7278932738304138,
|
| 166 |
+
"epoch": 1.716564417177914,
|
| 167 |
+
"grad_norm": 0.8753179311752319,
|
| 168 |
+
"learning_rate": 0.0003973618587167924,
|
| 169 |
+
"loss": 0.694002456665039,
|
| 170 |
+
"mean_token_accuracy": 0.7849677371978759,
|
| 171 |
+
"num_tokens": 1853507.0,
|
| 172 |
+
"step": 700
|
| 173 |
+
},
|
| 174 |
+
{
|
| 175 |
+
"entropy": 0.7190143716335297,
|
| 176 |
+
"epoch": 1.8392638036809816,
|
| 177 |
+
"grad_norm": 0.8524742126464844,
|
| 178 |
+
"learning_rate": 0.0003950566070825483,
|
| 179 |
+
"loss": 0.6890143585205079,
|
| 180 |
+
"mean_token_accuracy": 0.7865350896120071,
|
| 181 |
+
"num_tokens": 1988113.0,
|
| 182 |
+
"step": 750
|
| 183 |
+
},
|
| 184 |
+
{
|
| 185 |
+
"entropy": 0.7072037667036056,
|
| 186 |
+
"epoch": 1.961963190184049,
|
| 187 |
+
"grad_norm": 0.877348005771637,
|
| 188 |
+
"learning_rate": 0.00039239774833008714,
|
| 189 |
+
"loss": 0.6795162200927735,
|
| 190 |
+
"mean_token_accuracy": 0.7885233855247498,
|
| 191 |
+
"num_tokens": 2121314.0,
|
| 192 |
+
"step": 800
|
| 193 |
+
},
|
| 194 |
+
{
|
| 195 |
+
"epoch": 2.0,
|
| 196 |
+
"eval_entropy": 0.818787784235818,
|
| 197 |
+
"eval_mean_token_accuracy": 0.7573730857031686,
|
| 198 |
+
"eval_not_syn_loss": 0.7486392855644226,
|
| 199 |
+
"eval_not_syn_runtime": 56.7426,
|
| 200 |
+
"eval_not_syn_samples_per_second": 24.602,
|
| 201 |
+
"eval_not_syn_steps_per_second": 3.084,
|
| 202 |
+
"eval_num_tokens": 2160392.0,
|
| 203 |
+
"step": 816
|
| 204 |
+
},
|
| 205 |
+
{
|
| 206 |
+
"epoch": 2.0,
|
| 207 |
+
"eval_entropy": 0.7632522354807173,
|
| 208 |
+
"eval_mean_token_accuracy": 0.8101163625717163,
|
| 209 |
+
"eval_num_tokens": 2160392.0,
|
| 210 |
+
"eval_syn_loss": 0.6829464435577393,
|
| 211 |
+
"eval_syn_runtime": 58.5255,
|
| 212 |
+
"eval_syn_samples_per_second": 23.853,
|
| 213 |
+
"eval_syn_steps_per_second": 2.99,
|
| 214 |
+
"step": 816
|
| 215 |
+
},
|
| 216 |
+
{
|
| 217 |
+
"entropy": 0.6619202801675508,
|
| 218 |
+
"epoch": 2.083435582822086,
|
| 219 |
+
"grad_norm": 0.8124546408653259,
|
| 220 |
+
"learning_rate": 0.00038939014724212456,
|
| 221 |
+
"loss": 0.6311377716064454,
|
| 222 |
+
"mean_token_accuracy": 0.7991988044796567,
|
| 223 |
+
"num_tokens": 2253179.0,
|
| 224 |
+
"step": 850
|
| 225 |
+
},
|
| 226 |
+
{
|
| 227 |
+
"entropy": 0.6668815296888352,
|
| 228 |
+
"epoch": 2.2061349693251535,
|
| 229 |
+
"grad_norm": 0.8693709969520569,
|
| 230 |
+
"learning_rate": 0.0003860393066780265,
|
| 231 |
+
"loss": 0.6279015731811524,
|
| 232 |
+
"mean_token_accuracy": 0.7983763587474823,
|
| 233 |
+
"num_tokens": 2379965.0,
|
| 234 |
+
"step": 900
|
| 235 |
+
},
|
| 236 |
+
{
|
| 237 |
+
"entropy": 0.6557431703805924,
|
| 238 |
+
"epoch": 2.3288343558282207,
|
| 239 |
+
"grad_norm": 0.8323768377304077,
|
| 240 |
+
"learning_rate": 0.0003823513575054986,
|
| 241 |
+
"loss": 0.6214478683471679,
|
| 242 |
+
"mean_token_accuracy": 0.8012632429599762,
|
| 243 |
+
"num_tokens": 2515593.0,
|
| 244 |
+
"step": 950
|
| 245 |
+
},
|
| 246 |
+
{
|
| 247 |
+
"entropy": 0.6633540654182434,
|
| 248 |
+
"epoch": 2.4515337423312884,
|
| 249 |
+
"grad_norm": 0.8328827023506165,
|
| 250 |
+
"learning_rate": 0.0003783330473832399,
|
| 251 |
+
"loss": 0.6310849761962891,
|
| 252 |
+
"mean_token_accuracy": 0.7995503985881806,
|
| 253 |
+
"num_tokens": 2642790.0,
|
| 254 |
+
"step": 1000
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"entropy": 0.6542247200012207,
|
| 258 |
+
"epoch": 2.574233128834356,
|
| 259 |
+
"grad_norm": 0.8204748630523682,
|
| 260 |
+
"learning_rate": 0.000373991728415085,
|
| 261 |
+
"loss": 0.6195787048339844,
|
| 262 |
+
"mean_token_accuracy": 0.8006457507610321,
|
| 263 |
+
"num_tokens": 2774396.0,
|
| 264 |
+
"step": 1050
|
| 265 |
+
},
|
| 266 |
+
{
|
| 267 |
+
"entropy": 0.6526653742790223,
|
| 268 |
+
"epoch": 2.6969325153374233,
|
| 269 |
+
"grad_norm": 1.065635323524475,
|
| 270 |
+
"learning_rate": 0.0003693353436982218,
|
| 271 |
+
"loss": 0.6206568145751953,
|
| 272 |
+
"mean_token_accuracy": 0.8002620726823807,
|
| 273 |
+
"num_tokens": 2912102.0,
|
| 274 |
+
"step": 1100
|
| 275 |
+
},
|
| 276 |
+
{
|
| 277 |
+
"entropy": 0.654406590461731,
|
| 278 |
+
"epoch": 2.819631901840491,
|
| 279 |
+
"grad_norm": 0.8653210401535034,
|
| 280 |
+
"learning_rate": 0.00036437241279009834,
|
| 281 |
+
"loss": 0.6259954452514649,
|
| 282 |
+
"mean_token_accuracy": 0.8008036535978317,
|
| 283 |
+
"num_tokens": 3042987.0,
|
| 284 |
+
"step": 1150
|
| 285 |
+
},
|
| 286 |
+
{
|
| 287 |
+
"entropy": 0.6310120761394501,
|
| 288 |
+
"epoch": 2.942331288343558,
|
| 289 |
+
"grad_norm": 0.7713245749473572,
|
| 290 |
+
"learning_rate": 0.0003591120161206093,
|
| 291 |
+
"loss": 0.6115898132324219,
|
| 292 |
+
"mean_token_accuracy": 0.8043569129705429,
|
| 293 |
+
"num_tokens": 3178044.0,
|
| 294 |
+
"step": 1200
|
| 295 |
+
},
|
| 296 |
+
{
|
| 297 |
+
"epoch": 3.0,
|
| 298 |
+
"eval_entropy": 0.6591688336644854,
|
| 299 |
+
"eval_mean_token_accuracy": 0.7992405581474304,
|
| 300 |
+
"eval_not_syn_loss": 0.6962469220161438,
|
| 301 |
+
"eval_not_syn_runtime": 56.161,
|
| 302 |
+
"eval_not_syn_samples_per_second": 24.857,
|
| 303 |
+
"eval_not_syn_steps_per_second": 3.116,
|
| 304 |
+
"eval_num_tokens": 3240588.0,
|
| 305 |
+
"step": 1224
|
| 306 |
+
},
|
| 307 |
+
{
|
| 308 |
+
"epoch": 3.0,
|
| 309 |
+
"eval_entropy": 0.6191406740461077,
|
| 310 |
+
"eval_mean_token_accuracy": 0.7810143651281084,
|
| 311 |
+
"eval_num_tokens": 3240588.0,
|
| 312 |
+
"eval_syn_loss": 0.6663697361946106,
|
| 313 |
+
"eval_syn_runtime": 58.4102,
|
| 314 |
+
"eval_syn_samples_per_second": 23.9,
|
| 315 |
+
"eval_syn_steps_per_second": 2.996,
|
| 316 |
+
"step": 1224
|
| 317 |
+
}
|
| 318 |
+
],
|
| 319 |
+
"logging_steps": 50,
|
| 320 |
+
"max_steps": 4080,
|
| 321 |
+
"num_input_tokens_seen": 0,
|
| 322 |
+
"num_train_epochs": 10,
|
| 323 |
+
"save_steps": 500,
|
| 324 |
+
"stateful_callbacks": {
|
| 325 |
+
"TrainerControl": {
|
| 326 |
+
"args": {
|
| 327 |
+
"should_epoch_stop": false,
|
| 328 |
+
"should_evaluate": false,
|
| 329 |
+
"should_log": false,
|
| 330 |
+
"should_save": true,
|
| 331 |
+
"should_training_stop": false
|
| 332 |
+
},
|
| 333 |
+
"attributes": {}
|
| 334 |
+
}
|
| 335 |
+
},
|
| 336 |
+
"total_flos": 6.610858430973312e+16,
|
| 337 |
+
"train_batch_size": 8,
|
| 338 |
+
"trial_name": null,
|
| 339 |
+
"trial_params": null
|
| 340 |
+
}
|
substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-1632/adapter_config.json
ADDED
|
@@ -0,0 +1,48 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"alora_invocation_tokens": null,
|
| 3 |
+
"alpha_pattern": {},
|
| 4 |
+
"arrow_config": null,
|
| 5 |
+
"auto_mapping": null,
|
| 6 |
+
"base_model_name_or_path": "Qwen/Qwen3.5-2B-Base",
|
| 7 |
+
"bias": "none",
|
| 8 |
+
"corda_config": null,
|
| 9 |
+
"ensure_weight_tying": false,
|
| 10 |
+
"eva_config": null,
|
| 11 |
+
"exclude_modules": null,
|
| 12 |
+
"fan_in_fan_out": false,
|
| 13 |
+
"inference_mode": true,
|
| 14 |
+
"init_lora_weights": true,
|
| 15 |
+
"layer_replication": null,
|
| 16 |
+
"layers_pattern": null,
|
| 17 |
+
"layers_to_transform": null,
|
| 18 |
+
"loftq_config": {},
|
| 19 |
+
"lora_alpha": 128,
|
| 20 |
+
"lora_bias": false,
|
| 21 |
+
"lora_dropout": 0.05635041589438875,
|
| 22 |
+
"lora_ga_config": null,
|
| 23 |
+
"megatron_config": null,
|
| 24 |
+
"megatron_core": "megatron.core",
|
| 25 |
+
"modules_to_save": null,
|
| 26 |
+
"peft_type": "LORA",
|
| 27 |
+
"peft_version": "0.19.1",
|
| 28 |
+
"qalora_group_size": 16,
|
| 29 |
+
"r": 64,
|
| 30 |
+
"rank_pattern": {},
|
| 31 |
+
"revision": null,
|
| 32 |
+
"target_modules": [
|
| 33 |
+
"q_proj",
|
| 34 |
+
"k_proj",
|
| 35 |
+
"gate_proj",
|
| 36 |
+
"up_proj",
|
| 37 |
+
"o_proj",
|
| 38 |
+
"down_proj",
|
| 39 |
+
"v_proj"
|
| 40 |
+
],
|
| 41 |
+
"target_parameters": null,
|
| 42 |
+
"task_type": "CAUSAL_LM",
|
| 43 |
+
"trainable_token_indices": null,
|
| 44 |
+
"use_bdlora": null,
|
| 45 |
+
"use_dora": false,
|
| 46 |
+
"use_qalora": false,
|
| 47 |
+
"use_rslora": false
|
| 48 |
+
}
|
substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-1632/chat_template.jinja
ADDED
|
@@ -0,0 +1,154 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{%- set image_count = namespace(value=0) %}
|
| 2 |
+
{%- set video_count = namespace(value=0) %}
|
| 3 |
+
{%- macro render_content(content, do_vision_count, is_system_content=false) %}
|
| 4 |
+
{%- if content is string %}
|
| 5 |
+
{{- content }}
|
| 6 |
+
{%- elif content is iterable and content is not mapping %}
|
| 7 |
+
{%- for item in content %}
|
| 8 |
+
{%- if 'image' in item or 'image_url' in item or item.type == 'image' %}
|
| 9 |
+
{%- if is_system_content %}
|
| 10 |
+
{{- raise_exception('System message cannot contain images.') }}
|
| 11 |
+
{%- endif %}
|
| 12 |
+
{%- if do_vision_count %}
|
| 13 |
+
{%- set image_count.value = image_count.value + 1 %}
|
| 14 |
+
{%- endif %}
|
| 15 |
+
{%- if add_vision_id %}
|
| 16 |
+
{{- 'Picture ' ~ image_count.value ~ ': ' }}
|
| 17 |
+
{%- endif %}
|
| 18 |
+
{{- '<|vision_start|><|image_pad|><|vision_end|>' }}
|
| 19 |
+
{%- elif 'video' in item or item.type == 'video' %}
|
| 20 |
+
{%- if is_system_content %}
|
| 21 |
+
{{- raise_exception('System message cannot contain videos.') }}
|
| 22 |
+
{%- endif %}
|
| 23 |
+
{%- if do_vision_count %}
|
| 24 |
+
{%- set video_count.value = video_count.value + 1 %}
|
| 25 |
+
{%- endif %}
|
| 26 |
+
{%- if add_vision_id %}
|
| 27 |
+
{{- 'Video ' ~ video_count.value ~ ': ' }}
|
| 28 |
+
{%- endif %}
|
| 29 |
+
{{- '<|vision_start|><|video_pad|><|vision_end|>' }}
|
| 30 |
+
{%- elif 'text' in item %}
|
| 31 |
+
{{- item.text }}
|
| 32 |
+
{%- else %}
|
| 33 |
+
{{- raise_exception('Unexpected item type in content.') }}
|
| 34 |
+
{%- endif %}
|
| 35 |
+
{%- endfor %}
|
| 36 |
+
{%- elif content is none or content is undefined %}
|
| 37 |
+
{{- '' }}
|
| 38 |
+
{%- else %}
|
| 39 |
+
{{- raise_exception('Unexpected content type.') }}
|
| 40 |
+
{%- endif %}
|
| 41 |
+
{%- endmacro %}
|
| 42 |
+
{%- if not messages %}
|
| 43 |
+
{{- raise_exception('No messages provided.') }}
|
| 44 |
+
{%- endif %}
|
| 45 |
+
{%- if tools and tools is iterable and tools is not mapping %}
|
| 46 |
+
{{- '<|im_start|>system\n' }}
|
| 47 |
+
{{- "# Tools\n\nYou have access to the following functions:\n\n<tools>" }}
|
| 48 |
+
{%- for tool in tools %}
|
| 49 |
+
{{- "\n" }}
|
| 50 |
+
{{- tool | tojson }}
|
| 51 |
+
{%- endfor %}
|
| 52 |
+
{{- "\n</tools>" }}
|
| 53 |
+
{{- '\n\nIf you choose to call a function ONLY reply in the following format with NO suffix:\n\n<tool_call>\n<function=example_function_name>\n<parameter=example_parameter_1>\nvalue_1\n</parameter>\n<parameter=example_parameter_2>\nThis is the value for the second parameter\nthat can span\nmultiple lines\n</parameter>\n</function>\n</tool_call>\n\n<IMPORTANT>\nReminder:\n- Function calls MUST follow the specified format: an inner <function=...></function> block must be nested within <tool_call></tool_call> XML tags\n- Required parameters MUST be specified\n- You may provide optional reasoning for your function call in natural language BEFORE the function call, but NOT after\n- If there is no function call available, answer the question like normal with your current knowledge and do not tell the user about function calls\n</IMPORTANT>' }}
|
| 54 |
+
{%- if messages[0].role == 'system' %}
|
| 55 |
+
{%- set content = render_content(messages[0].content, false, true)|trim %}
|
| 56 |
+
{%- if content %}
|
| 57 |
+
{{- '\n\n' + content }}
|
| 58 |
+
{%- endif %}
|
| 59 |
+
{%- endif %}
|
| 60 |
+
{{- '<|im_end|>\n' }}
|
| 61 |
+
{%- else %}
|
| 62 |
+
{%- if messages[0].role == 'system' %}
|
| 63 |
+
{%- set content = render_content(messages[0].content, false, true)|trim %}
|
| 64 |
+
{{- '<|im_start|>system\n' + content + '<|im_end|>\n' }}
|
| 65 |
+
{%- endif %}
|
| 66 |
+
{%- endif %}
|
| 67 |
+
{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
|
| 68 |
+
{%- for message in messages[::-1] %}
|
| 69 |
+
{%- set index = (messages|length - 1) - loop.index0 %}
|
| 70 |
+
{%- if ns.multi_step_tool and message.role == "user" %}
|
| 71 |
+
{%- set content = render_content(message.content, false)|trim %}
|
| 72 |
+
{%- if not(content.startswith('<tool_response>') and content.endswith('</tool_response>')) %}
|
| 73 |
+
{%- set ns.multi_step_tool = false %}
|
| 74 |
+
{%- set ns.last_query_index = index %}
|
| 75 |
+
{%- endif %}
|
| 76 |
+
{%- endif %}
|
| 77 |
+
{%- endfor %}
|
| 78 |
+
{%- if ns.multi_step_tool %}
|
| 79 |
+
{{- raise_exception('No user query found in messages.') }}
|
| 80 |
+
{%- endif %}
|
| 81 |
+
{%- for message in messages %}
|
| 82 |
+
{%- set content = render_content(message.content, true)|trim %}
|
| 83 |
+
{%- if message.role == "system" %}
|
| 84 |
+
{%- if not loop.first %}
|
| 85 |
+
{{- raise_exception('System message must be at the beginning.') }}
|
| 86 |
+
{%- endif %}
|
| 87 |
+
{%- elif message.role == "user" %}
|
| 88 |
+
{{- '<|im_start|>' + message.role + '\n' + content + '<|im_end|>' + '\n' }}
|
| 89 |
+
{%- elif message.role == "assistant" %}
|
| 90 |
+
{%- set reasoning_content = '' %}
|
| 91 |
+
{%- if message.reasoning_content is string %}
|
| 92 |
+
{%- set reasoning_content = message.reasoning_content %}
|
| 93 |
+
{%- else %}
|
| 94 |
+
{%- if '</think>' in content %}
|
| 95 |
+
{%- set reasoning_content = content.split('</think>')[0].rstrip('\n').split('<think>')[-1].lstrip('\n') %}
|
| 96 |
+
{%- set content = content.split('</think>')[-1].lstrip('\n') %}
|
| 97 |
+
{%- endif %}
|
| 98 |
+
{%- endif %}
|
| 99 |
+
{%- set reasoning_content = reasoning_content|trim %}
|
| 100 |
+
{%- if loop.index0 > ns.last_query_index %}
|
| 101 |
+
{{- '<|im_start|>' + message.role + '\n<think>\n' + reasoning_content + '\n</think>\n\n' + content }}
|
| 102 |
+
{%- else %}
|
| 103 |
+
{{- '<|im_start|>' + message.role + '\n' + content }}
|
| 104 |
+
{%- endif %}
|
| 105 |
+
{%- if message.tool_calls and message.tool_calls is iterable and message.tool_calls is not mapping %}
|
| 106 |
+
{%- for tool_call in message.tool_calls %}
|
| 107 |
+
{%- if tool_call.function is defined %}
|
| 108 |
+
{%- set tool_call = tool_call.function %}
|
| 109 |
+
{%- endif %}
|
| 110 |
+
{%- if loop.first %}
|
| 111 |
+
{%- if content|trim %}
|
| 112 |
+
{{- '\n\n<tool_call>\n<function=' + tool_call.name + '>\n' }}
|
| 113 |
+
{%- else %}
|
| 114 |
+
{{- '<tool_call>\n<function=' + tool_call.name + '>\n' }}
|
| 115 |
+
{%- endif %}
|
| 116 |
+
{%- else %}
|
| 117 |
+
{{- '\n<tool_call>\n<function=' + tool_call.name + '>\n' }}
|
| 118 |
+
{%- endif %}
|
| 119 |
+
{%- if tool_call.arguments is defined %}
|
| 120 |
+
{%- for args_name, args_value in tool_call.arguments|items %}
|
| 121 |
+
{{- '<parameter=' + args_name + '>\n' }}
|
| 122 |
+
{%- set args_value = args_value | tojson | safe if args_value is mapping or (args_value is sequence and args_value is not string) else args_value | string %}
|
| 123 |
+
{{- args_value }}
|
| 124 |
+
{{- '\n</parameter>\n' }}
|
| 125 |
+
{%- endfor %}
|
| 126 |
+
{%- endif %}
|
| 127 |
+
{{- '</function>\n</tool_call>' }}
|
| 128 |
+
{%- endfor %}
|
| 129 |
+
{%- endif %}
|
| 130 |
+
{{- '<|im_end|>\n' }}
|
| 131 |
+
{%- elif message.role == "tool" %}
|
| 132 |
+
{%- if loop.previtem and loop.previtem.role != "tool" %}
|
| 133 |
+
{{- '<|im_start|>user' }}
|
| 134 |
+
{%- endif %}
|
| 135 |
+
{{- '\n<tool_response>\n' }}
|
| 136 |
+
{{- content }}
|
| 137 |
+
{{- '\n</tool_response>' }}
|
| 138 |
+
{%- if not loop.last and loop.nextitem.role != "tool" %}
|
| 139 |
+
{{- '<|im_end|>\n' }}
|
| 140 |
+
{%- elif loop.last %}
|
| 141 |
+
{{- '<|im_end|>\n' }}
|
| 142 |
+
{%- endif %}
|
| 143 |
+
{%- else %}
|
| 144 |
+
{{- raise_exception('Unexpected message role.') }}
|
| 145 |
+
{%- endif %}
|
| 146 |
+
{%- endfor %}
|
| 147 |
+
{%- if add_generation_prompt %}
|
| 148 |
+
{{- '<|im_start|>assistant\n' }}
|
| 149 |
+
{%- if enable_thinking is defined and enable_thinking is true %}
|
| 150 |
+
{{- '<think>\n' }}
|
| 151 |
+
{%- else %}
|
| 152 |
+
{{- '<think>\n\n</think>\n\n' }}
|
| 153 |
+
{%- endif %}
|
| 154 |
+
{%- endif %}
|
substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-1632/tokenizer_config.json
ADDED
|
@@ -0,0 +1,31 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"add_prefix_space": false,
|
| 3 |
+
"audio_bos_token": "<|audio_start|>",
|
| 4 |
+
"audio_eos_token": "<|audio_end|>",
|
| 5 |
+
"audio_token": "<|audio_pad|>",
|
| 6 |
+
"backend": "tokenizers",
|
| 7 |
+
"bos_token": null,
|
| 8 |
+
"clean_up_tokenization_spaces": false,
|
| 9 |
+
"eos_token": "<|endoftext|>",
|
| 10 |
+
"errors": "replace",
|
| 11 |
+
"image_token": "<|image_pad|>",
|
| 12 |
+
"is_local": false,
|
| 13 |
+
"model_max_length": 262144,
|
| 14 |
+
"model_specific_special_tokens": {
|
| 15 |
+
"audio_bos_token": "<|audio_start|>",
|
| 16 |
+
"audio_eos_token": "<|audio_end|>",
|
| 17 |
+
"audio_token": "<|audio_pad|>",
|
| 18 |
+
"image_token": "<|image_pad|>",
|
| 19 |
+
"video_token": "<|video_pad|>",
|
| 20 |
+
"vision_bos_token": "<|vision_start|>",
|
| 21 |
+
"vision_eos_token": "<|vision_end|>"
|
| 22 |
+
},
|
| 23 |
+
"pad_token": "<|endoftext|>",
|
| 24 |
+
"pretokenize_regex": "(?i:'s|'t|'re|'ve|'m|'ll|'d)|[^\\r\\n\\p{L}\\p{N}]?[\\p{L}\\p{M}]+|\\p{N}| ?[^\\s\\p{L}\\p{M}\\p{N}]+[\\r\\n]*|\\s*[\\r\\n]+|\\s+(?!\\S)|\\s+",
|
| 25 |
+
"split_special_tokens": false,
|
| 26 |
+
"tokenizer_class": "TokenizersBackend",
|
| 27 |
+
"unk_token": null,
|
| 28 |
+
"video_token": "<|video_pad|>",
|
| 29 |
+
"vision_bos_token": "<|vision_start|>",
|
| 30 |
+
"vision_eos_token": "<|vision_end|>"
|
| 31 |
+
}
|
substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-1632/trainer_state.json
ADDED
|
@@ -0,0 +1,442 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": null,
|
| 3 |
+
"best_metric": null,
|
| 4 |
+
"best_model_checkpoint": null,
|
| 5 |
+
"epoch": 4.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 1632,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"entropy": 1.8153233402967452,
|
| 14 |
+
"epoch": 0.12269938650306748,
|
| 15 |
+
"grad_norm": 1.8510568141937256,
|
| 16 |
+
"learning_rate": 4.8469594667113335e-05,
|
| 17 |
+
"loss": 1.7217323303222656,
|
| 18 |
+
"mean_token_accuracy": 0.6236635231971741,
|
| 19 |
+
"num_tokens": 134329.0,
|
| 20 |
+
"step": 50
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"entropy": 0.97831438601017,
|
| 24 |
+
"epoch": 0.24539877300613497,
|
| 25 |
+
"grad_norm": 1.889868974685669,
|
| 26 |
+
"learning_rate": 9.792836473559631e-05,
|
| 27 |
+
"loss": 0.9050540924072266,
|
| 28 |
+
"mean_token_accuracy": 0.7377463465929032,
|
| 29 |
+
"num_tokens": 262407.0,
|
| 30 |
+
"step": 100
|
| 31 |
+
},
|
| 32 |
+
{
|
| 33 |
+
"entropy": 0.8743131357431412,
|
| 34 |
+
"epoch": 0.36809815950920244,
|
| 35 |
+
"grad_norm": 1.1766732931137085,
|
| 36 |
+
"learning_rate": 0.0001473871348040793,
|
| 37 |
+
"loss": 0.8129763031005859,
|
| 38 |
+
"mean_token_accuracy": 0.7583417356014251,
|
| 39 |
+
"num_tokens": 401136.0,
|
| 40 |
+
"step": 150
|
| 41 |
+
},
|
| 42 |
+
{
|
| 43 |
+
"entropy": 0.8273966419696808,
|
| 44 |
+
"epoch": 0.49079754601226994,
|
| 45 |
+
"grad_norm": 1.1311014890670776,
|
| 46 |
+
"learning_rate": 0.00019684590487256229,
|
| 47 |
+
"loss": 0.7711280822753906,
|
| 48 |
+
"mean_token_accuracy": 0.7668228060007095,
|
| 49 |
+
"num_tokens": 531647.0,
|
| 50 |
+
"step": 200
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"entropy": 0.8101438617706299,
|
| 54 |
+
"epoch": 0.6134969325153374,
|
| 55 |
+
"grad_norm": 0.8905950784683228,
|
| 56 |
+
"learning_rate": 0.0002463046749410453,
|
| 57 |
+
"loss": 0.7609222412109375,
|
| 58 |
+
"mean_token_accuracy": 0.7717174577713013,
|
| 59 |
+
"num_tokens": 666127.0,
|
| 60 |
+
"step": 250
|
| 61 |
+
},
|
| 62 |
+
{
|
| 63 |
+
"entropy": 0.7885055243968964,
|
| 64 |
+
"epoch": 0.7361963190184049,
|
| 65 |
+
"grad_norm": 1.1944245100021362,
|
| 66 |
+
"learning_rate": 0.00029576344500952824,
|
| 67 |
+
"loss": 0.7477743530273437,
|
| 68 |
+
"mean_token_accuracy": 0.773262197971344,
|
| 69 |
+
"num_tokens": 799292.0,
|
| 70 |
+
"step": 300
|
| 71 |
+
},
|
| 72 |
+
{
|
| 73 |
+
"entropy": 0.7858527797460556,
|
| 74 |
+
"epoch": 0.8588957055214724,
|
| 75 |
+
"grad_norm": 1.1311068534851074,
|
| 76 |
+
"learning_rate": 0.00034522221507801124,
|
| 77 |
+
"loss": 0.7411377716064453,
|
| 78 |
+
"mean_token_accuracy": 0.7745680212974548,
|
| 79 |
+
"num_tokens": 930614.0,
|
| 80 |
+
"step": 350
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"entropy": 0.7879797518253326,
|
| 84 |
+
"epoch": 0.9815950920245399,
|
| 85 |
+
"grad_norm": 0.9370436072349548,
|
| 86 |
+
"learning_rate": 0.00039468098514649423,
|
| 87 |
+
"loss": 0.7507785034179687,
|
| 88 |
+
"mean_token_accuracy": 0.7747444450855255,
|
| 89 |
+
"num_tokens": 1061654.0,
|
| 90 |
+
"step": 400
|
| 91 |
+
},
|
| 92 |
+
{
|
| 93 |
+
"epoch": 1.0,
|
| 94 |
+
"eval_entropy": 0.8527693561145238,
|
| 95 |
+
"eval_mean_token_accuracy": 0.7730878329277039,
|
| 96 |
+
"eval_not_syn_loss": 0.781921923160553,
|
| 97 |
+
"eval_not_syn_runtime": 56.9157,
|
| 98 |
+
"eval_not_syn_samples_per_second": 24.527,
|
| 99 |
+
"eval_not_syn_steps_per_second": 3.075,
|
| 100 |
+
"eval_num_tokens": 1080196.0,
|
| 101 |
+
"step": 408
|
| 102 |
+
},
|
| 103 |
+
{
|
| 104 |
+
"epoch": 1.0,
|
| 105 |
+
"eval_entropy": 0.7952005583899362,
|
| 106 |
+
"eval_mean_token_accuracy": 0.7705784467288426,
|
| 107 |
+
"eval_num_tokens": 1080196.0,
|
| 108 |
+
"eval_syn_loss": 0.7461121082305908,
|
| 109 |
+
"eval_syn_runtime": 59.5758,
|
| 110 |
+
"eval_syn_samples_per_second": 23.432,
|
| 111 |
+
"eval_syn_steps_per_second": 2.937,
|
| 112 |
+
"step": 408
|
| 113 |
+
},
|
| 114 |
+
{
|
| 115 |
+
"entropy": 0.7564875923021875,
|
| 116 |
+
"epoch": 1.1030674846625768,
|
| 117 |
+
"grad_norm": 1.1334599256515503,
|
| 118 |
+
"learning_rate": 0.00040345942966973637,
|
| 119 |
+
"loss": 0.7208955383300781,
|
| 120 |
+
"mean_token_accuracy": 0.7788144470465304,
|
| 121 |
+
"num_tokens": 1199170.0,
|
| 122 |
+
"step": 450
|
| 123 |
+
},
|
| 124 |
+
{
|
| 125 |
+
"entropy": 0.7427570706605912,
|
| 126 |
+
"epoch": 1.2257668711656442,
|
| 127 |
+
"grad_norm": 1.0777021646499634,
|
| 128 |
+
"learning_rate": 0.00040297229629224153,
|
| 129 |
+
"loss": 0.7049168395996094,
|
| 130 |
+
"mean_token_accuracy": 0.7830300116539002,
|
| 131 |
+
"num_tokens": 1334298.0,
|
| 132 |
+
"step": 500
|
| 133 |
+
},
|
| 134 |
+
{
|
| 135 |
+
"entropy": 0.7533353108167649,
|
| 136 |
+
"epoch": 1.3484662576687116,
|
| 137 |
+
"grad_norm": 1.3439561128616333,
|
| 138 |
+
"learning_rate": 0.00040211707285041045,
|
| 139 |
+
"loss": 0.7102288818359375,
|
| 140 |
+
"mean_token_accuracy": 0.7818449640274048,
|
| 141 |
+
"num_tokens": 1463189.0,
|
| 142 |
+
"step": 550
|
| 143 |
+
},
|
| 144 |
+
{
|
| 145 |
+
"entropy": 0.7666342407464981,
|
| 146 |
+
"epoch": 1.471165644171779,
|
| 147 |
+
"grad_norm": 1.1329729557037354,
|
| 148 |
+
"learning_rate": 0.00040089532410439227,
|
| 149 |
+
"loss": 0.7270996856689453,
|
| 150 |
+
"mean_token_accuracy": 0.778404277563095,
|
| 151 |
+
"num_tokens": 1590310.0,
|
| 152 |
+
"step": 600
|
| 153 |
+
},
|
| 154 |
+
{
|
| 155 |
+
"entropy": 0.7359934949874878,
|
| 156 |
+
"epoch": 1.5938650306748468,
|
| 157 |
+
"grad_norm": 1.2227331399917603,
|
| 158 |
+
"learning_rate": 0.0003993092854276068,
|
| 159 |
+
"loss": 0.6996555328369141,
|
| 160 |
+
"mean_token_accuracy": 0.7834958267211914,
|
| 161 |
+
"num_tokens": 1721836.0,
|
| 162 |
+
"step": 650
|
| 163 |
+
},
|
| 164 |
+
{
|
| 165 |
+
"entropy": 0.7278932738304138,
|
| 166 |
+
"epoch": 1.716564417177914,
|
| 167 |
+
"grad_norm": 0.8753179311752319,
|
| 168 |
+
"learning_rate": 0.0003973618587167924,
|
| 169 |
+
"loss": 0.694002456665039,
|
| 170 |
+
"mean_token_accuracy": 0.7849677371978759,
|
| 171 |
+
"num_tokens": 1853507.0,
|
| 172 |
+
"step": 700
|
| 173 |
+
},
|
| 174 |
+
{
|
| 175 |
+
"entropy": 0.7190143716335297,
|
| 176 |
+
"epoch": 1.8392638036809816,
|
| 177 |
+
"grad_norm": 0.8524742126464844,
|
| 178 |
+
"learning_rate": 0.0003950566070825483,
|
| 179 |
+
"loss": 0.6890143585205079,
|
| 180 |
+
"mean_token_accuracy": 0.7865350896120071,
|
| 181 |
+
"num_tokens": 1988113.0,
|
| 182 |
+
"step": 750
|
| 183 |
+
},
|
| 184 |
+
{
|
| 185 |
+
"entropy": 0.7072037667036056,
|
| 186 |
+
"epoch": 1.961963190184049,
|
| 187 |
+
"grad_norm": 0.877348005771637,
|
| 188 |
+
"learning_rate": 0.00039239774833008714,
|
| 189 |
+
"loss": 0.6795162200927735,
|
| 190 |
+
"mean_token_accuracy": 0.7885233855247498,
|
| 191 |
+
"num_tokens": 2121314.0,
|
| 192 |
+
"step": 800
|
| 193 |
+
},
|
| 194 |
+
{
|
| 195 |
+
"epoch": 2.0,
|
| 196 |
+
"eval_entropy": 0.818787784235818,
|
| 197 |
+
"eval_mean_token_accuracy": 0.7573730857031686,
|
| 198 |
+
"eval_not_syn_loss": 0.7486392855644226,
|
| 199 |
+
"eval_not_syn_runtime": 56.7426,
|
| 200 |
+
"eval_not_syn_samples_per_second": 24.602,
|
| 201 |
+
"eval_not_syn_steps_per_second": 3.084,
|
| 202 |
+
"eval_num_tokens": 2160392.0,
|
| 203 |
+
"step": 816
|
| 204 |
+
},
|
| 205 |
+
{
|
| 206 |
+
"epoch": 2.0,
|
| 207 |
+
"eval_entropy": 0.7632522354807173,
|
| 208 |
+
"eval_mean_token_accuracy": 0.8101163625717163,
|
| 209 |
+
"eval_num_tokens": 2160392.0,
|
| 210 |
+
"eval_syn_loss": 0.6829464435577393,
|
| 211 |
+
"eval_syn_runtime": 58.5255,
|
| 212 |
+
"eval_syn_samples_per_second": 23.853,
|
| 213 |
+
"eval_syn_steps_per_second": 2.99,
|
| 214 |
+
"step": 816
|
| 215 |
+
},
|
| 216 |
+
{
|
| 217 |
+
"entropy": 0.6619202801675508,
|
| 218 |
+
"epoch": 2.083435582822086,
|
| 219 |
+
"grad_norm": 0.8124546408653259,
|
| 220 |
+
"learning_rate": 0.00038939014724212456,
|
| 221 |
+
"loss": 0.6311377716064454,
|
| 222 |
+
"mean_token_accuracy": 0.7991988044796567,
|
| 223 |
+
"num_tokens": 2253179.0,
|
| 224 |
+
"step": 850
|
| 225 |
+
},
|
| 226 |
+
{
|
| 227 |
+
"entropy": 0.6668815296888352,
|
| 228 |
+
"epoch": 2.2061349693251535,
|
| 229 |
+
"grad_norm": 0.8693709969520569,
|
| 230 |
+
"learning_rate": 0.0003860393066780265,
|
| 231 |
+
"loss": 0.6279015731811524,
|
| 232 |
+
"mean_token_accuracy": 0.7983763587474823,
|
| 233 |
+
"num_tokens": 2379965.0,
|
| 234 |
+
"step": 900
|
| 235 |
+
},
|
| 236 |
+
{
|
| 237 |
+
"entropy": 0.6557431703805924,
|
| 238 |
+
"epoch": 2.3288343558282207,
|
| 239 |
+
"grad_norm": 0.8323768377304077,
|
| 240 |
+
"learning_rate": 0.0003823513575054986,
|
| 241 |
+
"loss": 0.6214478683471679,
|
| 242 |
+
"mean_token_accuracy": 0.8012632429599762,
|
| 243 |
+
"num_tokens": 2515593.0,
|
| 244 |
+
"step": 950
|
| 245 |
+
},
|
| 246 |
+
{
|
| 247 |
+
"entropy": 0.6633540654182434,
|
| 248 |
+
"epoch": 2.4515337423312884,
|
| 249 |
+
"grad_norm": 0.8328827023506165,
|
| 250 |
+
"learning_rate": 0.0003783330473832399,
|
| 251 |
+
"loss": 0.6310849761962891,
|
| 252 |
+
"mean_token_accuracy": 0.7995503985881806,
|
| 253 |
+
"num_tokens": 2642790.0,
|
| 254 |
+
"step": 1000
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"entropy": 0.6542247200012207,
|
| 258 |
+
"epoch": 2.574233128834356,
|
| 259 |
+
"grad_norm": 0.8204748630523682,
|
| 260 |
+
"learning_rate": 0.000373991728415085,
|
| 261 |
+
"loss": 0.6195787048339844,
|
| 262 |
+
"mean_token_accuracy": 0.8006457507610321,
|
| 263 |
+
"num_tokens": 2774396.0,
|
| 264 |
+
"step": 1050
|
| 265 |
+
},
|
| 266 |
+
{
|
| 267 |
+
"entropy": 0.6526653742790223,
|
| 268 |
+
"epoch": 2.6969325153374233,
|
| 269 |
+
"grad_norm": 1.065635323524475,
|
| 270 |
+
"learning_rate": 0.0003693353436982218,
|
| 271 |
+
"loss": 0.6206568145751953,
|
| 272 |
+
"mean_token_accuracy": 0.8002620726823807,
|
| 273 |
+
"num_tokens": 2912102.0,
|
| 274 |
+
"step": 1100
|
| 275 |
+
},
|
| 276 |
+
{
|
| 277 |
+
"entropy": 0.654406590461731,
|
| 278 |
+
"epoch": 2.819631901840491,
|
| 279 |
+
"grad_norm": 0.8653210401535034,
|
| 280 |
+
"learning_rate": 0.00036437241279009834,
|
| 281 |
+
"loss": 0.6259954452514649,
|
| 282 |
+
"mean_token_accuracy": 0.8008036535978317,
|
| 283 |
+
"num_tokens": 3042987.0,
|
| 284 |
+
"step": 1150
|
| 285 |
+
},
|
| 286 |
+
{
|
| 287 |
+
"entropy": 0.6310120761394501,
|
| 288 |
+
"epoch": 2.942331288343558,
|
| 289 |
+
"grad_norm": 0.7713245749473572,
|
| 290 |
+
"learning_rate": 0.0003591120161206093,
|
| 291 |
+
"loss": 0.6115898132324219,
|
| 292 |
+
"mean_token_accuracy": 0.8043569129705429,
|
| 293 |
+
"num_tokens": 3178044.0,
|
| 294 |
+
"step": 1200
|
| 295 |
+
},
|
| 296 |
+
{
|
| 297 |
+
"epoch": 3.0,
|
| 298 |
+
"eval_entropy": 0.6591688336644854,
|
| 299 |
+
"eval_mean_token_accuracy": 0.7992405581474304,
|
| 300 |
+
"eval_not_syn_loss": 0.6962469220161438,
|
| 301 |
+
"eval_not_syn_runtime": 56.161,
|
| 302 |
+
"eval_not_syn_samples_per_second": 24.857,
|
| 303 |
+
"eval_not_syn_steps_per_second": 3.116,
|
| 304 |
+
"eval_num_tokens": 3240588.0,
|
| 305 |
+
"step": 1224
|
| 306 |
+
},
|
| 307 |
+
{
|
| 308 |
+
"epoch": 3.0,
|
| 309 |
+
"eval_entropy": 0.6191406740461077,
|
| 310 |
+
"eval_mean_token_accuracy": 0.7810143651281084,
|
| 311 |
+
"eval_num_tokens": 3240588.0,
|
| 312 |
+
"eval_syn_loss": 0.6663697361946106,
|
| 313 |
+
"eval_syn_runtime": 58.4102,
|
| 314 |
+
"eval_syn_samples_per_second": 23.9,
|
| 315 |
+
"eval_syn_steps_per_second": 2.996,
|
| 316 |
+
"step": 1224
|
| 317 |
+
},
|
| 318 |
+
{
|
| 319 |
+
"entropy": 0.5853322166385073,
|
| 320 |
+
"epoch": 3.063803680981595,
|
| 321 |
+
"grad_norm": 0.7313649654388428,
|
| 322 |
+
"learning_rate": 0.00035356377837808084,
|
| 323 |
+
"loss": 0.5616408538818359,
|
| 324 |
+
"mean_token_accuracy": 0.8144418639366073,
|
| 325 |
+
"num_tokens": 3312797.0,
|
| 326 |
+
"step": 1250
|
| 327 |
+
},
|
| 328 |
+
{
|
| 329 |
+
"entropy": 0.5654504173994064,
|
| 330 |
+
"epoch": 3.1865030674846624,
|
| 331 |
+
"grad_norm": 0.713893711566925,
|
| 332 |
+
"learning_rate": 0.0003477378508994529,
|
| 333 |
+
"loss": 0.5305458068847656,
|
| 334 |
+
"mean_token_accuracy": 0.8218590825796127,
|
| 335 |
+
"num_tokens": 3441629.0,
|
| 336 |
+
"step": 1300
|
| 337 |
+
},
|
| 338 |
+
{
|
| 339 |
+
"entropy": 0.5721602493524551,
|
| 340 |
+
"epoch": 3.30920245398773,
|
| 341 |
+
"grad_norm": 0.8487221002578735,
|
| 342 |
+
"learning_rate": 0.00034164489309687927,
|
| 343 |
+
"loss": 0.5437272262573242,
|
| 344 |
+
"mean_token_accuracy": 0.8169907212257386,
|
| 345 |
+
"num_tokens": 3576093.0,
|
| 346 |
+
"step": 1350
|
| 347 |
+
},
|
| 348 |
+
{
|
| 349 |
+
"entropy": 0.5797226822376251,
|
| 350 |
+
"epoch": 3.4319018404907977,
|
| 351 |
+
"grad_norm": 0.9307994842529297,
|
| 352 |
+
"learning_rate": 0.0003352960529547267,
|
| 353 |
+
"loss": 0.5405152130126953,
|
| 354 |
+
"mean_token_accuracy": 0.818563597202301,
|
| 355 |
+
"num_tokens": 3705912.0,
|
| 356 |
+
"step": 1400
|
| 357 |
+
},
|
| 358 |
+
{
|
| 359 |
+
"entropy": 0.5822264245152473,
|
| 360 |
+
"epoch": 3.554601226993865,
|
| 361 |
+
"grad_norm": 0.71127849817276,
|
| 362 |
+
"learning_rate": 0.00032870294663265757,
|
| 363 |
+
"loss": 0.5559784698486329,
|
| 364 |
+
"mean_token_accuracy": 0.8159428876638413,
|
| 365 |
+
"num_tokens": 3832376.0,
|
| 366 |
+
"step": 1450
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"entropy": 0.5702577340602875,
|
| 370 |
+
"epoch": 3.6773006134969326,
|
| 371 |
+
"grad_norm": 0.8090758919715881,
|
| 372 |
+
"learning_rate": 0.0003218776372121157,
|
| 373 |
+
"loss": 0.5466982650756836,
|
| 374 |
+
"mean_token_accuracy": 0.8186432421207428,
|
| 375 |
+
"num_tokens": 3967273.0,
|
| 376 |
+
"step": 1500
|
| 377 |
+
},
|
| 378 |
+
{
|
| 379 |
+
"entropy": 0.5708746567368508,
|
| 380 |
+
"epoch": 3.8,
|
| 381 |
+
"grad_norm": 0.7777291536331177,
|
| 382 |
+
"learning_rate": 0.000314832612625101,
|
| 383 |
+
"loss": 0.5412663650512696,
|
| 384 |
+
"mean_token_accuracy": 0.8195344746112824,
|
| 385 |
+
"num_tokens": 4102434.0,
|
| 386 |
+
"step": 1550
|
| 387 |
+
},
|
| 388 |
+
{
|
| 389 |
+
"entropy": 0.578739655315876,
|
| 390 |
+
"epoch": 3.9226993865030675,
|
| 391 |
+
"grad_norm": 0.8103565573692322,
|
| 392 |
+
"learning_rate": 0.0003075807628056167,
|
| 393 |
+
"loss": 0.5605202865600586,
|
| 394 |
+
"mean_token_accuracy": 0.815134603381157,
|
| 395 |
+
"num_tokens": 4234037.0,
|
| 396 |
+
"step": 1600
|
| 397 |
+
},
|
| 398 |
+
{
|
| 399 |
+
"epoch": 4.0,
|
| 400 |
+
"eval_entropy": 0.6183825404303415,
|
| 401 |
+
"eval_mean_token_accuracy": 0.7838981601170131,
|
| 402 |
+
"eval_not_syn_loss": 0.6814553737640381,
|
| 403 |
+
"eval_not_syn_runtime": 56.4858,
|
| 404 |
+
"eval_not_syn_samples_per_second": 24.714,
|
| 405 |
+
"eval_not_syn_steps_per_second": 3.098,
|
| 406 |
+
"eval_num_tokens": 4320784.0,
|
| 407 |
+
"step": 1632
|
| 408 |
+
},
|
| 409 |
+
{
|
| 410 |
+
"epoch": 4.0,
|
| 411 |
+
"eval_entropy": 0.5816301565510886,
|
| 412 |
+
"eval_mean_token_accuracy": 0.8187636685371399,
|
| 413 |
+
"eval_num_tokens": 4320784.0,
|
| 414 |
+
"eval_syn_loss": 0.6366032361984253,
|
| 415 |
+
"eval_syn_runtime": 58.1317,
|
| 416 |
+
"eval_syn_samples_per_second": 24.014,
|
| 417 |
+
"eval_syn_steps_per_second": 3.01,
|
| 418 |
+
"step": 1632
|
| 419 |
+
}
|
| 420 |
+
],
|
| 421 |
+
"logging_steps": 50,
|
| 422 |
+
"max_steps": 4080,
|
| 423 |
+
"num_input_tokens_seen": 0,
|
| 424 |
+
"num_train_epochs": 10,
|
| 425 |
+
"save_steps": 500,
|
| 426 |
+
"stateful_callbacks": {
|
| 427 |
+
"TrainerControl": {
|
| 428 |
+
"args": {
|
| 429 |
+
"should_epoch_stop": false,
|
| 430 |
+
"should_evaluate": false,
|
| 431 |
+
"should_log": false,
|
| 432 |
+
"should_save": true,
|
| 433 |
+
"should_training_stop": false
|
| 434 |
+
},
|
| 435 |
+
"attributes": {}
|
| 436 |
+
}
|
| 437 |
+
},
|
| 438 |
+
"total_flos": 8.819362572820186e+16,
|
| 439 |
+
"train_batch_size": 8,
|
| 440 |
+
"trial_name": null,
|
| 441 |
+
"trial_params": null
|
| 442 |
+
}
|
substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-2040/README.md
ADDED
|
@@ -0,0 +1,209 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
base_model: Qwen/Qwen3.5-2B-Base
|
| 3 |
+
library_name: peft
|
| 4 |
+
pipeline_tag: text-generation
|
| 5 |
+
tags:
|
| 6 |
+
- base_model:adapter:Qwen/Qwen3.5-2B-Base
|
| 7 |
+
- lora
|
| 8 |
+
- sft
|
| 9 |
+
- transformers
|
| 10 |
+
- trl
|
| 11 |
+
---
|
| 12 |
+
|
| 13 |
+
# Model Card for Model ID
|
| 14 |
+
|
| 15 |
+
<!-- Provide a quick summary of what the model is/does. -->
|
| 16 |
+
|
| 17 |
+
|
| 18 |
+
|
| 19 |
+
## Model Details
|
| 20 |
+
|
| 21 |
+
### Model Description
|
| 22 |
+
|
| 23 |
+
<!-- Provide a longer summary of what this model is. -->
|
| 24 |
+
|
| 25 |
+
|
| 26 |
+
|
| 27 |
+
- **Developed by:** [More Information Needed]
|
| 28 |
+
- **Funded by [optional]:** [More Information Needed]
|
| 29 |
+
- **Shared by [optional]:** [More Information Needed]
|
| 30 |
+
- **Model type:** [More Information Needed]
|
| 31 |
+
- **Language(s) (NLP):** [More Information Needed]
|
| 32 |
+
- **License:** [More Information Needed]
|
| 33 |
+
- **Finetuned from model [optional]:** [More Information Needed]
|
| 34 |
+
|
| 35 |
+
### Model Sources [optional]
|
| 36 |
+
|
| 37 |
+
<!-- Provide the basic links for the model. -->
|
| 38 |
+
|
| 39 |
+
- **Repository:** [More Information Needed]
|
| 40 |
+
- **Paper [optional]:** [More Information Needed]
|
| 41 |
+
- **Demo [optional]:** [More Information Needed]
|
| 42 |
+
|
| 43 |
+
## Uses
|
| 44 |
+
|
| 45 |
+
<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
|
| 46 |
+
|
| 47 |
+
### Direct Use
|
| 48 |
+
|
| 49 |
+
<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
|
| 50 |
+
|
| 51 |
+
[More Information Needed]
|
| 52 |
+
|
| 53 |
+
### Downstream Use [optional]
|
| 54 |
+
|
| 55 |
+
<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
|
| 56 |
+
|
| 57 |
+
[More Information Needed]
|
| 58 |
+
|
| 59 |
+
### Out-of-Scope Use
|
| 60 |
+
|
| 61 |
+
<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
|
| 62 |
+
|
| 63 |
+
[More Information Needed]
|
| 64 |
+
|
| 65 |
+
## Bias, Risks, and Limitations
|
| 66 |
+
|
| 67 |
+
<!-- This section is meant to convey both technical and sociotechnical limitations. -->
|
| 68 |
+
|
| 69 |
+
[More Information Needed]
|
| 70 |
+
|
| 71 |
+
### Recommendations
|
| 72 |
+
|
| 73 |
+
<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
|
| 74 |
+
|
| 75 |
+
Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
|
| 76 |
+
|
| 77 |
+
## How to Get Started with the Model
|
| 78 |
+
|
| 79 |
+
Use the code below to get started with the model.
|
| 80 |
+
|
| 81 |
+
[More Information Needed]
|
| 82 |
+
|
| 83 |
+
## Training Details
|
| 84 |
+
|
| 85 |
+
### Training Data
|
| 86 |
+
|
| 87 |
+
<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
|
| 88 |
+
|
| 89 |
+
[More Information Needed]
|
| 90 |
+
|
| 91 |
+
### Training Procedure
|
| 92 |
+
|
| 93 |
+
<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
|
| 94 |
+
|
| 95 |
+
#### Preprocessing [optional]
|
| 96 |
+
|
| 97 |
+
[More Information Needed]
|
| 98 |
+
|
| 99 |
+
|
| 100 |
+
#### Training Hyperparameters
|
| 101 |
+
|
| 102 |
+
- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
|
| 103 |
+
|
| 104 |
+
#### Speeds, Sizes, Times [optional]
|
| 105 |
+
|
| 106 |
+
<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
|
| 107 |
+
|
| 108 |
+
[More Information Needed]
|
| 109 |
+
|
| 110 |
+
## Evaluation
|
| 111 |
+
|
| 112 |
+
<!-- This section describes the evaluation protocols and provides the results. -->
|
| 113 |
+
|
| 114 |
+
### Testing Data, Factors & Metrics
|
| 115 |
+
|
| 116 |
+
#### Testing Data
|
| 117 |
+
|
| 118 |
+
<!-- This should link to a Dataset Card if possible. -->
|
| 119 |
+
|
| 120 |
+
[More Information Needed]
|
| 121 |
+
|
| 122 |
+
#### Factors
|
| 123 |
+
|
| 124 |
+
<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
|
| 125 |
+
|
| 126 |
+
[More Information Needed]
|
| 127 |
+
|
| 128 |
+
#### Metrics
|
| 129 |
+
|
| 130 |
+
<!-- These are the evaluation metrics being used, ideally with a description of why. -->
|
| 131 |
+
|
| 132 |
+
[More Information Needed]
|
| 133 |
+
|
| 134 |
+
### Results
|
| 135 |
+
|
| 136 |
+
[More Information Needed]
|
| 137 |
+
|
| 138 |
+
#### Summary
|
| 139 |
+
|
| 140 |
+
|
| 141 |
+
|
| 142 |
+
## Model Examination [optional]
|
| 143 |
+
|
| 144 |
+
<!-- Relevant interpretability work for the model goes here -->
|
| 145 |
+
|
| 146 |
+
[More Information Needed]
|
| 147 |
+
|
| 148 |
+
## Environmental Impact
|
| 149 |
+
|
| 150 |
+
<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
|
| 151 |
+
|
| 152 |
+
Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
|
| 153 |
+
|
| 154 |
+
- **Hardware Type:** [More Information Needed]
|
| 155 |
+
- **Hours used:** [More Information Needed]
|
| 156 |
+
- **Cloud Provider:** [More Information Needed]
|
| 157 |
+
- **Compute Region:** [More Information Needed]
|
| 158 |
+
- **Carbon Emitted:** [More Information Needed]
|
| 159 |
+
|
| 160 |
+
## Technical Specifications [optional]
|
| 161 |
+
|
| 162 |
+
### Model Architecture and Objective
|
| 163 |
+
|
| 164 |
+
[More Information Needed]
|
| 165 |
+
|
| 166 |
+
### Compute Infrastructure
|
| 167 |
+
|
| 168 |
+
[More Information Needed]
|
| 169 |
+
|
| 170 |
+
#### Hardware
|
| 171 |
+
|
| 172 |
+
[More Information Needed]
|
| 173 |
+
|
| 174 |
+
#### Software
|
| 175 |
+
|
| 176 |
+
[More Information Needed]
|
| 177 |
+
|
| 178 |
+
## Citation [optional]
|
| 179 |
+
|
| 180 |
+
<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
|
| 181 |
+
|
| 182 |
+
**BibTeX:**
|
| 183 |
+
|
| 184 |
+
[More Information Needed]
|
| 185 |
+
|
| 186 |
+
**APA:**
|
| 187 |
+
|
| 188 |
+
[More Information Needed]
|
| 189 |
+
|
| 190 |
+
## Glossary [optional]
|
| 191 |
+
|
| 192 |
+
<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
|
| 193 |
+
|
| 194 |
+
[More Information Needed]
|
| 195 |
+
|
| 196 |
+
## More Information [optional]
|
| 197 |
+
|
| 198 |
+
[More Information Needed]
|
| 199 |
+
|
| 200 |
+
## Model Card Authors [optional]
|
| 201 |
+
|
| 202 |
+
[More Information Needed]
|
| 203 |
+
|
| 204 |
+
## Model Card Contact
|
| 205 |
+
|
| 206 |
+
[More Information Needed]
|
| 207 |
+
### Framework versions
|
| 208 |
+
|
| 209 |
+
- PEFT 0.19.1
|
substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-2040/adapter_config.json
ADDED
|
@@ -0,0 +1,48 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"alora_invocation_tokens": null,
|
| 3 |
+
"alpha_pattern": {},
|
| 4 |
+
"arrow_config": null,
|
| 5 |
+
"auto_mapping": null,
|
| 6 |
+
"base_model_name_or_path": "Qwen/Qwen3.5-2B-Base",
|
| 7 |
+
"bias": "none",
|
| 8 |
+
"corda_config": null,
|
| 9 |
+
"ensure_weight_tying": false,
|
| 10 |
+
"eva_config": null,
|
| 11 |
+
"exclude_modules": null,
|
| 12 |
+
"fan_in_fan_out": false,
|
| 13 |
+
"inference_mode": true,
|
| 14 |
+
"init_lora_weights": true,
|
| 15 |
+
"layer_replication": null,
|
| 16 |
+
"layers_pattern": null,
|
| 17 |
+
"layers_to_transform": null,
|
| 18 |
+
"loftq_config": {},
|
| 19 |
+
"lora_alpha": 128,
|
| 20 |
+
"lora_bias": false,
|
| 21 |
+
"lora_dropout": 0.05635041589438875,
|
| 22 |
+
"lora_ga_config": null,
|
| 23 |
+
"megatron_config": null,
|
| 24 |
+
"megatron_core": "megatron.core",
|
| 25 |
+
"modules_to_save": null,
|
| 26 |
+
"peft_type": "LORA",
|
| 27 |
+
"peft_version": "0.19.1",
|
| 28 |
+
"qalora_group_size": 16,
|
| 29 |
+
"r": 64,
|
| 30 |
+
"rank_pattern": {},
|
| 31 |
+
"revision": null,
|
| 32 |
+
"target_modules": [
|
| 33 |
+
"q_proj",
|
| 34 |
+
"k_proj",
|
| 35 |
+
"gate_proj",
|
| 36 |
+
"up_proj",
|
| 37 |
+
"o_proj",
|
| 38 |
+
"down_proj",
|
| 39 |
+
"v_proj"
|
| 40 |
+
],
|
| 41 |
+
"target_parameters": null,
|
| 42 |
+
"task_type": "CAUSAL_LM",
|
| 43 |
+
"trainable_token_indices": null,
|
| 44 |
+
"use_bdlora": null,
|
| 45 |
+
"use_dora": false,
|
| 46 |
+
"use_qalora": false,
|
| 47 |
+
"use_rslora": false
|
| 48 |
+
}
|
substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-2040/chat_template.jinja
ADDED
|
@@ -0,0 +1,154 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{%- set image_count = namespace(value=0) %}
|
| 2 |
+
{%- set video_count = namespace(value=0) %}
|
| 3 |
+
{%- macro render_content(content, do_vision_count, is_system_content=false) %}
|
| 4 |
+
{%- if content is string %}
|
| 5 |
+
{{- content }}
|
| 6 |
+
{%- elif content is iterable and content is not mapping %}
|
| 7 |
+
{%- for item in content %}
|
| 8 |
+
{%- if 'image' in item or 'image_url' in item or item.type == 'image' %}
|
| 9 |
+
{%- if is_system_content %}
|
| 10 |
+
{{- raise_exception('System message cannot contain images.') }}
|
| 11 |
+
{%- endif %}
|
| 12 |
+
{%- if do_vision_count %}
|
| 13 |
+
{%- set image_count.value = image_count.value + 1 %}
|
| 14 |
+
{%- endif %}
|
| 15 |
+
{%- if add_vision_id %}
|
| 16 |
+
{{- 'Picture ' ~ image_count.value ~ ': ' }}
|
| 17 |
+
{%- endif %}
|
| 18 |
+
{{- '<|vision_start|><|image_pad|><|vision_end|>' }}
|
| 19 |
+
{%- elif 'video' in item or item.type == 'video' %}
|
| 20 |
+
{%- if is_system_content %}
|
| 21 |
+
{{- raise_exception('System message cannot contain videos.') }}
|
| 22 |
+
{%- endif %}
|
| 23 |
+
{%- if do_vision_count %}
|
| 24 |
+
{%- set video_count.value = video_count.value + 1 %}
|
| 25 |
+
{%- endif %}
|
| 26 |
+
{%- if add_vision_id %}
|
| 27 |
+
{{- 'Video ' ~ video_count.value ~ ': ' }}
|
| 28 |
+
{%- endif %}
|
| 29 |
+
{{- '<|vision_start|><|video_pad|><|vision_end|>' }}
|
| 30 |
+
{%- elif 'text' in item %}
|
| 31 |
+
{{- item.text }}
|
| 32 |
+
{%- else %}
|
| 33 |
+
{{- raise_exception('Unexpected item type in content.') }}
|
| 34 |
+
{%- endif %}
|
| 35 |
+
{%- endfor %}
|
| 36 |
+
{%- elif content is none or content is undefined %}
|
| 37 |
+
{{- '' }}
|
| 38 |
+
{%- else %}
|
| 39 |
+
{{- raise_exception('Unexpected content type.') }}
|
| 40 |
+
{%- endif %}
|
| 41 |
+
{%- endmacro %}
|
| 42 |
+
{%- if not messages %}
|
| 43 |
+
{{- raise_exception('No messages provided.') }}
|
| 44 |
+
{%- endif %}
|
| 45 |
+
{%- if tools and tools is iterable and tools is not mapping %}
|
| 46 |
+
{{- '<|im_start|>system\n' }}
|
| 47 |
+
{{- "# Tools\n\nYou have access to the following functions:\n\n<tools>" }}
|
| 48 |
+
{%- for tool in tools %}
|
| 49 |
+
{{- "\n" }}
|
| 50 |
+
{{- tool | tojson }}
|
| 51 |
+
{%- endfor %}
|
| 52 |
+
{{- "\n</tools>" }}
|
| 53 |
+
{{- '\n\nIf you choose to call a function ONLY reply in the following format with NO suffix:\n\n<tool_call>\n<function=example_function_name>\n<parameter=example_parameter_1>\nvalue_1\n</parameter>\n<parameter=example_parameter_2>\nThis is the value for the second parameter\nthat can span\nmultiple lines\n</parameter>\n</function>\n</tool_call>\n\n<IMPORTANT>\nReminder:\n- Function calls MUST follow the specified format: an inner <function=...></function> block must be nested within <tool_call></tool_call> XML tags\n- Required parameters MUST be specified\n- You may provide optional reasoning for your function call in natural language BEFORE the function call, but NOT after\n- If there is no function call available, answer the question like normal with your current knowledge and do not tell the user about function calls\n</IMPORTANT>' }}
|
| 54 |
+
{%- if messages[0].role == 'system' %}
|
| 55 |
+
{%- set content = render_content(messages[0].content, false, true)|trim %}
|
| 56 |
+
{%- if content %}
|
| 57 |
+
{{- '\n\n' + content }}
|
| 58 |
+
{%- endif %}
|
| 59 |
+
{%- endif %}
|
| 60 |
+
{{- '<|im_end|>\n' }}
|
| 61 |
+
{%- else %}
|
| 62 |
+
{%- if messages[0].role == 'system' %}
|
| 63 |
+
{%- set content = render_content(messages[0].content, false, true)|trim %}
|
| 64 |
+
{{- '<|im_start|>system\n' + content + '<|im_end|>\n' }}
|
| 65 |
+
{%- endif %}
|
| 66 |
+
{%- endif %}
|
| 67 |
+
{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
|
| 68 |
+
{%- for message in messages[::-1] %}
|
| 69 |
+
{%- set index = (messages|length - 1) - loop.index0 %}
|
| 70 |
+
{%- if ns.multi_step_tool and message.role == "user" %}
|
| 71 |
+
{%- set content = render_content(message.content, false)|trim %}
|
| 72 |
+
{%- if not(content.startswith('<tool_response>') and content.endswith('</tool_response>')) %}
|
| 73 |
+
{%- set ns.multi_step_tool = false %}
|
| 74 |
+
{%- set ns.last_query_index = index %}
|
| 75 |
+
{%- endif %}
|
| 76 |
+
{%- endif %}
|
| 77 |
+
{%- endfor %}
|
| 78 |
+
{%- if ns.multi_step_tool %}
|
| 79 |
+
{{- raise_exception('No user query found in messages.') }}
|
| 80 |
+
{%- endif %}
|
| 81 |
+
{%- for message in messages %}
|
| 82 |
+
{%- set content = render_content(message.content, true)|trim %}
|
| 83 |
+
{%- if message.role == "system" %}
|
| 84 |
+
{%- if not loop.first %}
|
| 85 |
+
{{- raise_exception('System message must be at the beginning.') }}
|
| 86 |
+
{%- endif %}
|
| 87 |
+
{%- elif message.role == "user" %}
|
| 88 |
+
{{- '<|im_start|>' + message.role + '\n' + content + '<|im_end|>' + '\n' }}
|
| 89 |
+
{%- elif message.role == "assistant" %}
|
| 90 |
+
{%- set reasoning_content = '' %}
|
| 91 |
+
{%- if message.reasoning_content is string %}
|
| 92 |
+
{%- set reasoning_content = message.reasoning_content %}
|
| 93 |
+
{%- else %}
|
| 94 |
+
{%- if '</think>' in content %}
|
| 95 |
+
{%- set reasoning_content = content.split('</think>')[0].rstrip('\n').split('<think>')[-1].lstrip('\n') %}
|
| 96 |
+
{%- set content = content.split('</think>')[-1].lstrip('\n') %}
|
| 97 |
+
{%- endif %}
|
| 98 |
+
{%- endif %}
|
| 99 |
+
{%- set reasoning_content = reasoning_content|trim %}
|
| 100 |
+
{%- if loop.index0 > ns.last_query_index %}
|
| 101 |
+
{{- '<|im_start|>' + message.role + '\n<think>\n' + reasoning_content + '\n</think>\n\n' + content }}
|
| 102 |
+
{%- else %}
|
| 103 |
+
{{- '<|im_start|>' + message.role + '\n' + content }}
|
| 104 |
+
{%- endif %}
|
| 105 |
+
{%- if message.tool_calls and message.tool_calls is iterable and message.tool_calls is not mapping %}
|
| 106 |
+
{%- for tool_call in message.tool_calls %}
|
| 107 |
+
{%- if tool_call.function is defined %}
|
| 108 |
+
{%- set tool_call = tool_call.function %}
|
| 109 |
+
{%- endif %}
|
| 110 |
+
{%- if loop.first %}
|
| 111 |
+
{%- if content|trim %}
|
| 112 |
+
{{- '\n\n<tool_call>\n<function=' + tool_call.name + '>\n' }}
|
| 113 |
+
{%- else %}
|
| 114 |
+
{{- '<tool_call>\n<function=' + tool_call.name + '>\n' }}
|
| 115 |
+
{%- endif %}
|
| 116 |
+
{%- else %}
|
| 117 |
+
{{- '\n<tool_call>\n<function=' + tool_call.name + '>\n' }}
|
| 118 |
+
{%- endif %}
|
| 119 |
+
{%- if tool_call.arguments is defined %}
|
| 120 |
+
{%- for args_name, args_value in tool_call.arguments|items %}
|
| 121 |
+
{{- '<parameter=' + args_name + '>\n' }}
|
| 122 |
+
{%- set args_value = args_value | tojson | safe if args_value is mapping or (args_value is sequence and args_value is not string) else args_value | string %}
|
| 123 |
+
{{- args_value }}
|
| 124 |
+
{{- '\n</parameter>\n' }}
|
| 125 |
+
{%- endfor %}
|
| 126 |
+
{%- endif %}
|
| 127 |
+
{{- '</function>\n</tool_call>' }}
|
| 128 |
+
{%- endfor %}
|
| 129 |
+
{%- endif %}
|
| 130 |
+
{{- '<|im_end|>\n' }}
|
| 131 |
+
{%- elif message.role == "tool" %}
|
| 132 |
+
{%- if loop.previtem and loop.previtem.role != "tool" %}
|
| 133 |
+
{{- '<|im_start|>user' }}
|
| 134 |
+
{%- endif %}
|
| 135 |
+
{{- '\n<tool_response>\n' }}
|
| 136 |
+
{{- content }}
|
| 137 |
+
{{- '\n</tool_response>' }}
|
| 138 |
+
{%- if not loop.last and loop.nextitem.role != "tool" %}
|
| 139 |
+
{{- '<|im_end|>\n' }}
|
| 140 |
+
{%- elif loop.last %}
|
| 141 |
+
{{- '<|im_end|>\n' }}
|
| 142 |
+
{%- endif %}
|
| 143 |
+
{%- else %}
|
| 144 |
+
{{- raise_exception('Unexpected message role.') }}
|
| 145 |
+
{%- endif %}
|
| 146 |
+
{%- endfor %}
|
| 147 |
+
{%- if add_generation_prompt %}
|
| 148 |
+
{{- '<|im_start|>assistant\n' }}
|
| 149 |
+
{%- if enable_thinking is defined and enable_thinking is true %}
|
| 150 |
+
{{- '<think>\n' }}
|
| 151 |
+
{%- else %}
|
| 152 |
+
{{- '<think>\n\n</think>\n\n' }}
|
| 153 |
+
{%- endif %}
|
| 154 |
+
{%- endif %}
|
substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-2040/tokenizer_config.json
ADDED
|
@@ -0,0 +1,31 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"add_prefix_space": false,
|
| 3 |
+
"audio_bos_token": "<|audio_start|>",
|
| 4 |
+
"audio_eos_token": "<|audio_end|>",
|
| 5 |
+
"audio_token": "<|audio_pad|>",
|
| 6 |
+
"backend": "tokenizers",
|
| 7 |
+
"bos_token": null,
|
| 8 |
+
"clean_up_tokenization_spaces": false,
|
| 9 |
+
"eos_token": "<|endoftext|>",
|
| 10 |
+
"errors": "replace",
|
| 11 |
+
"image_token": "<|image_pad|>",
|
| 12 |
+
"is_local": false,
|
| 13 |
+
"model_max_length": 262144,
|
| 14 |
+
"model_specific_special_tokens": {
|
| 15 |
+
"audio_bos_token": "<|audio_start|>",
|
| 16 |
+
"audio_eos_token": "<|audio_end|>",
|
| 17 |
+
"audio_token": "<|audio_pad|>",
|
| 18 |
+
"image_token": "<|image_pad|>",
|
| 19 |
+
"video_token": "<|video_pad|>",
|
| 20 |
+
"vision_bos_token": "<|vision_start|>",
|
| 21 |
+
"vision_eos_token": "<|vision_end|>"
|
| 22 |
+
},
|
| 23 |
+
"pad_token": "<|endoftext|>",
|
| 24 |
+
"pretokenize_regex": "(?i:'s|'t|'re|'ve|'m|'ll|'d)|[^\\r\\n\\p{L}\\p{N}]?[\\p{L}\\p{M}]+|\\p{N}| ?[^\\s\\p{L}\\p{M}\\p{N}]+[\\r\\n]*|\\s*[\\r\\n]+|\\s+(?!\\S)|\\s+",
|
| 25 |
+
"split_special_tokens": false,
|
| 26 |
+
"tokenizer_class": "TokenizersBackend",
|
| 27 |
+
"unk_token": null,
|
| 28 |
+
"video_token": "<|video_pad|>",
|
| 29 |
+
"vision_bos_token": "<|vision_start|>",
|
| 30 |
+
"vision_eos_token": "<|vision_end|>"
|
| 31 |
+
}
|
substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-2040/trainer_state.json
ADDED
|
@@ -0,0 +1,544 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": null,
|
| 3 |
+
"best_metric": null,
|
| 4 |
+
"best_model_checkpoint": null,
|
| 5 |
+
"epoch": 5.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 2040,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"entropy": 1.8153233402967452,
|
| 14 |
+
"epoch": 0.12269938650306748,
|
| 15 |
+
"grad_norm": 1.8510568141937256,
|
| 16 |
+
"learning_rate": 4.8469594667113335e-05,
|
| 17 |
+
"loss": 1.7217323303222656,
|
| 18 |
+
"mean_token_accuracy": 0.6236635231971741,
|
| 19 |
+
"num_tokens": 134329.0,
|
| 20 |
+
"step": 50
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"entropy": 0.97831438601017,
|
| 24 |
+
"epoch": 0.24539877300613497,
|
| 25 |
+
"grad_norm": 1.889868974685669,
|
| 26 |
+
"learning_rate": 9.792836473559631e-05,
|
| 27 |
+
"loss": 0.9050540924072266,
|
| 28 |
+
"mean_token_accuracy": 0.7377463465929032,
|
| 29 |
+
"num_tokens": 262407.0,
|
| 30 |
+
"step": 100
|
| 31 |
+
},
|
| 32 |
+
{
|
| 33 |
+
"entropy": 0.8743131357431412,
|
| 34 |
+
"epoch": 0.36809815950920244,
|
| 35 |
+
"grad_norm": 1.1766732931137085,
|
| 36 |
+
"learning_rate": 0.0001473871348040793,
|
| 37 |
+
"loss": 0.8129763031005859,
|
| 38 |
+
"mean_token_accuracy": 0.7583417356014251,
|
| 39 |
+
"num_tokens": 401136.0,
|
| 40 |
+
"step": 150
|
| 41 |
+
},
|
| 42 |
+
{
|
| 43 |
+
"entropy": 0.8273966419696808,
|
| 44 |
+
"epoch": 0.49079754601226994,
|
| 45 |
+
"grad_norm": 1.1311014890670776,
|
| 46 |
+
"learning_rate": 0.00019684590487256229,
|
| 47 |
+
"loss": 0.7711280822753906,
|
| 48 |
+
"mean_token_accuracy": 0.7668228060007095,
|
| 49 |
+
"num_tokens": 531647.0,
|
| 50 |
+
"step": 200
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"entropy": 0.8101438617706299,
|
| 54 |
+
"epoch": 0.6134969325153374,
|
| 55 |
+
"grad_norm": 0.8905950784683228,
|
| 56 |
+
"learning_rate": 0.0002463046749410453,
|
| 57 |
+
"loss": 0.7609222412109375,
|
| 58 |
+
"mean_token_accuracy": 0.7717174577713013,
|
| 59 |
+
"num_tokens": 666127.0,
|
| 60 |
+
"step": 250
|
| 61 |
+
},
|
| 62 |
+
{
|
| 63 |
+
"entropy": 0.7885055243968964,
|
| 64 |
+
"epoch": 0.7361963190184049,
|
| 65 |
+
"grad_norm": 1.1944245100021362,
|
| 66 |
+
"learning_rate": 0.00029576344500952824,
|
| 67 |
+
"loss": 0.7477743530273437,
|
| 68 |
+
"mean_token_accuracy": 0.773262197971344,
|
| 69 |
+
"num_tokens": 799292.0,
|
| 70 |
+
"step": 300
|
| 71 |
+
},
|
| 72 |
+
{
|
| 73 |
+
"entropy": 0.7858527797460556,
|
| 74 |
+
"epoch": 0.8588957055214724,
|
| 75 |
+
"grad_norm": 1.1311068534851074,
|
| 76 |
+
"learning_rate": 0.00034522221507801124,
|
| 77 |
+
"loss": 0.7411377716064453,
|
| 78 |
+
"mean_token_accuracy": 0.7745680212974548,
|
| 79 |
+
"num_tokens": 930614.0,
|
| 80 |
+
"step": 350
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"entropy": 0.7879797518253326,
|
| 84 |
+
"epoch": 0.9815950920245399,
|
| 85 |
+
"grad_norm": 0.9370436072349548,
|
| 86 |
+
"learning_rate": 0.00039468098514649423,
|
| 87 |
+
"loss": 0.7507785034179687,
|
| 88 |
+
"mean_token_accuracy": 0.7747444450855255,
|
| 89 |
+
"num_tokens": 1061654.0,
|
| 90 |
+
"step": 400
|
| 91 |
+
},
|
| 92 |
+
{
|
| 93 |
+
"epoch": 1.0,
|
| 94 |
+
"eval_entropy": 0.8527693561145238,
|
| 95 |
+
"eval_mean_token_accuracy": 0.7730878329277039,
|
| 96 |
+
"eval_not_syn_loss": 0.781921923160553,
|
| 97 |
+
"eval_not_syn_runtime": 56.9157,
|
| 98 |
+
"eval_not_syn_samples_per_second": 24.527,
|
| 99 |
+
"eval_not_syn_steps_per_second": 3.075,
|
| 100 |
+
"eval_num_tokens": 1080196.0,
|
| 101 |
+
"step": 408
|
| 102 |
+
},
|
| 103 |
+
{
|
| 104 |
+
"epoch": 1.0,
|
| 105 |
+
"eval_entropy": 0.7952005583899362,
|
| 106 |
+
"eval_mean_token_accuracy": 0.7705784467288426,
|
| 107 |
+
"eval_num_tokens": 1080196.0,
|
| 108 |
+
"eval_syn_loss": 0.7461121082305908,
|
| 109 |
+
"eval_syn_runtime": 59.5758,
|
| 110 |
+
"eval_syn_samples_per_second": 23.432,
|
| 111 |
+
"eval_syn_steps_per_second": 2.937,
|
| 112 |
+
"step": 408
|
| 113 |
+
},
|
| 114 |
+
{
|
| 115 |
+
"entropy": 0.7564875923021875,
|
| 116 |
+
"epoch": 1.1030674846625768,
|
| 117 |
+
"grad_norm": 1.1334599256515503,
|
| 118 |
+
"learning_rate": 0.00040345942966973637,
|
| 119 |
+
"loss": 0.7208955383300781,
|
| 120 |
+
"mean_token_accuracy": 0.7788144470465304,
|
| 121 |
+
"num_tokens": 1199170.0,
|
| 122 |
+
"step": 450
|
| 123 |
+
},
|
| 124 |
+
{
|
| 125 |
+
"entropy": 0.7427570706605912,
|
| 126 |
+
"epoch": 1.2257668711656442,
|
| 127 |
+
"grad_norm": 1.0777021646499634,
|
| 128 |
+
"learning_rate": 0.00040297229629224153,
|
| 129 |
+
"loss": 0.7049168395996094,
|
| 130 |
+
"mean_token_accuracy": 0.7830300116539002,
|
| 131 |
+
"num_tokens": 1334298.0,
|
| 132 |
+
"step": 500
|
| 133 |
+
},
|
| 134 |
+
{
|
| 135 |
+
"entropy": 0.7533353108167649,
|
| 136 |
+
"epoch": 1.3484662576687116,
|
| 137 |
+
"grad_norm": 1.3439561128616333,
|
| 138 |
+
"learning_rate": 0.00040211707285041045,
|
| 139 |
+
"loss": 0.7102288818359375,
|
| 140 |
+
"mean_token_accuracy": 0.7818449640274048,
|
| 141 |
+
"num_tokens": 1463189.0,
|
| 142 |
+
"step": 550
|
| 143 |
+
},
|
| 144 |
+
{
|
| 145 |
+
"entropy": 0.7666342407464981,
|
| 146 |
+
"epoch": 1.471165644171779,
|
| 147 |
+
"grad_norm": 1.1329729557037354,
|
| 148 |
+
"learning_rate": 0.00040089532410439227,
|
| 149 |
+
"loss": 0.7270996856689453,
|
| 150 |
+
"mean_token_accuracy": 0.778404277563095,
|
| 151 |
+
"num_tokens": 1590310.0,
|
| 152 |
+
"step": 600
|
| 153 |
+
},
|
| 154 |
+
{
|
| 155 |
+
"entropy": 0.7359934949874878,
|
| 156 |
+
"epoch": 1.5938650306748468,
|
| 157 |
+
"grad_norm": 1.2227331399917603,
|
| 158 |
+
"learning_rate": 0.0003993092854276068,
|
| 159 |
+
"loss": 0.6996555328369141,
|
| 160 |
+
"mean_token_accuracy": 0.7834958267211914,
|
| 161 |
+
"num_tokens": 1721836.0,
|
| 162 |
+
"step": 650
|
| 163 |
+
},
|
| 164 |
+
{
|
| 165 |
+
"entropy": 0.7278932738304138,
|
| 166 |
+
"epoch": 1.716564417177914,
|
| 167 |
+
"grad_norm": 0.8753179311752319,
|
| 168 |
+
"learning_rate": 0.0003973618587167924,
|
| 169 |
+
"loss": 0.694002456665039,
|
| 170 |
+
"mean_token_accuracy": 0.7849677371978759,
|
| 171 |
+
"num_tokens": 1853507.0,
|
| 172 |
+
"step": 700
|
| 173 |
+
},
|
| 174 |
+
{
|
| 175 |
+
"entropy": 0.7190143716335297,
|
| 176 |
+
"epoch": 1.8392638036809816,
|
| 177 |
+
"grad_norm": 0.8524742126464844,
|
| 178 |
+
"learning_rate": 0.0003950566070825483,
|
| 179 |
+
"loss": 0.6890143585205079,
|
| 180 |
+
"mean_token_accuracy": 0.7865350896120071,
|
| 181 |
+
"num_tokens": 1988113.0,
|
| 182 |
+
"step": 750
|
| 183 |
+
},
|
| 184 |
+
{
|
| 185 |
+
"entropy": 0.7072037667036056,
|
| 186 |
+
"epoch": 1.961963190184049,
|
| 187 |
+
"grad_norm": 0.877348005771637,
|
| 188 |
+
"learning_rate": 0.00039239774833008714,
|
| 189 |
+
"loss": 0.6795162200927735,
|
| 190 |
+
"mean_token_accuracy": 0.7885233855247498,
|
| 191 |
+
"num_tokens": 2121314.0,
|
| 192 |
+
"step": 800
|
| 193 |
+
},
|
| 194 |
+
{
|
| 195 |
+
"epoch": 2.0,
|
| 196 |
+
"eval_entropy": 0.818787784235818,
|
| 197 |
+
"eval_mean_token_accuracy": 0.7573730857031686,
|
| 198 |
+
"eval_not_syn_loss": 0.7486392855644226,
|
| 199 |
+
"eval_not_syn_runtime": 56.7426,
|
| 200 |
+
"eval_not_syn_samples_per_second": 24.602,
|
| 201 |
+
"eval_not_syn_steps_per_second": 3.084,
|
| 202 |
+
"eval_num_tokens": 2160392.0,
|
| 203 |
+
"step": 816
|
| 204 |
+
},
|
| 205 |
+
{
|
| 206 |
+
"epoch": 2.0,
|
| 207 |
+
"eval_entropy": 0.7632522354807173,
|
| 208 |
+
"eval_mean_token_accuracy": 0.8101163625717163,
|
| 209 |
+
"eval_num_tokens": 2160392.0,
|
| 210 |
+
"eval_syn_loss": 0.6829464435577393,
|
| 211 |
+
"eval_syn_runtime": 58.5255,
|
| 212 |
+
"eval_syn_samples_per_second": 23.853,
|
| 213 |
+
"eval_syn_steps_per_second": 2.99,
|
| 214 |
+
"step": 816
|
| 215 |
+
},
|
| 216 |
+
{
|
| 217 |
+
"entropy": 0.6619202801675508,
|
| 218 |
+
"epoch": 2.083435582822086,
|
| 219 |
+
"grad_norm": 0.8124546408653259,
|
| 220 |
+
"learning_rate": 0.00038939014724212456,
|
| 221 |
+
"loss": 0.6311377716064454,
|
| 222 |
+
"mean_token_accuracy": 0.7991988044796567,
|
| 223 |
+
"num_tokens": 2253179.0,
|
| 224 |
+
"step": 850
|
| 225 |
+
},
|
| 226 |
+
{
|
| 227 |
+
"entropy": 0.6668815296888352,
|
| 228 |
+
"epoch": 2.2061349693251535,
|
| 229 |
+
"grad_norm": 0.8693709969520569,
|
| 230 |
+
"learning_rate": 0.0003860393066780265,
|
| 231 |
+
"loss": 0.6279015731811524,
|
| 232 |
+
"mean_token_accuracy": 0.7983763587474823,
|
| 233 |
+
"num_tokens": 2379965.0,
|
| 234 |
+
"step": 900
|
| 235 |
+
},
|
| 236 |
+
{
|
| 237 |
+
"entropy": 0.6557431703805924,
|
| 238 |
+
"epoch": 2.3288343558282207,
|
| 239 |
+
"grad_norm": 0.8323768377304077,
|
| 240 |
+
"learning_rate": 0.0003823513575054986,
|
| 241 |
+
"loss": 0.6214478683471679,
|
| 242 |
+
"mean_token_accuracy": 0.8012632429599762,
|
| 243 |
+
"num_tokens": 2515593.0,
|
| 244 |
+
"step": 950
|
| 245 |
+
},
|
| 246 |
+
{
|
| 247 |
+
"entropy": 0.6633540654182434,
|
| 248 |
+
"epoch": 2.4515337423312884,
|
| 249 |
+
"grad_norm": 0.8328827023506165,
|
| 250 |
+
"learning_rate": 0.0003783330473832399,
|
| 251 |
+
"loss": 0.6310849761962891,
|
| 252 |
+
"mean_token_accuracy": 0.7995503985881806,
|
| 253 |
+
"num_tokens": 2642790.0,
|
| 254 |
+
"step": 1000
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"entropy": 0.6542247200012207,
|
| 258 |
+
"epoch": 2.574233128834356,
|
| 259 |
+
"grad_norm": 0.8204748630523682,
|
| 260 |
+
"learning_rate": 0.000373991728415085,
|
| 261 |
+
"loss": 0.6195787048339844,
|
| 262 |
+
"mean_token_accuracy": 0.8006457507610321,
|
| 263 |
+
"num_tokens": 2774396.0,
|
| 264 |
+
"step": 1050
|
| 265 |
+
},
|
| 266 |
+
{
|
| 267 |
+
"entropy": 0.6526653742790223,
|
| 268 |
+
"epoch": 2.6969325153374233,
|
| 269 |
+
"grad_norm": 1.065635323524475,
|
| 270 |
+
"learning_rate": 0.0003693353436982218,
|
| 271 |
+
"loss": 0.6206568145751953,
|
| 272 |
+
"mean_token_accuracy": 0.8002620726823807,
|
| 273 |
+
"num_tokens": 2912102.0,
|
| 274 |
+
"step": 1100
|
| 275 |
+
},
|
| 276 |
+
{
|
| 277 |
+
"entropy": 0.654406590461731,
|
| 278 |
+
"epoch": 2.819631901840491,
|
| 279 |
+
"grad_norm": 0.8653210401535034,
|
| 280 |
+
"learning_rate": 0.00036437241279009834,
|
| 281 |
+
"loss": 0.6259954452514649,
|
| 282 |
+
"mean_token_accuracy": 0.8008036535978317,
|
| 283 |
+
"num_tokens": 3042987.0,
|
| 284 |
+
"step": 1150
|
| 285 |
+
},
|
| 286 |
+
{
|
| 287 |
+
"entropy": 0.6310120761394501,
|
| 288 |
+
"epoch": 2.942331288343558,
|
| 289 |
+
"grad_norm": 0.7713245749473572,
|
| 290 |
+
"learning_rate": 0.0003591120161206093,
|
| 291 |
+
"loss": 0.6115898132324219,
|
| 292 |
+
"mean_token_accuracy": 0.8043569129705429,
|
| 293 |
+
"num_tokens": 3178044.0,
|
| 294 |
+
"step": 1200
|
| 295 |
+
},
|
| 296 |
+
{
|
| 297 |
+
"epoch": 3.0,
|
| 298 |
+
"eval_entropy": 0.6591688336644854,
|
| 299 |
+
"eval_mean_token_accuracy": 0.7992405581474304,
|
| 300 |
+
"eval_not_syn_loss": 0.6962469220161438,
|
| 301 |
+
"eval_not_syn_runtime": 56.161,
|
| 302 |
+
"eval_not_syn_samples_per_second": 24.857,
|
| 303 |
+
"eval_not_syn_steps_per_second": 3.116,
|
| 304 |
+
"eval_num_tokens": 3240588.0,
|
| 305 |
+
"step": 1224
|
| 306 |
+
},
|
| 307 |
+
{
|
| 308 |
+
"epoch": 3.0,
|
| 309 |
+
"eval_entropy": 0.6191406740461077,
|
| 310 |
+
"eval_mean_token_accuracy": 0.7810143651281084,
|
| 311 |
+
"eval_num_tokens": 3240588.0,
|
| 312 |
+
"eval_syn_loss": 0.6663697361946106,
|
| 313 |
+
"eval_syn_runtime": 58.4102,
|
| 314 |
+
"eval_syn_samples_per_second": 23.9,
|
| 315 |
+
"eval_syn_steps_per_second": 2.996,
|
| 316 |
+
"step": 1224
|
| 317 |
+
},
|
| 318 |
+
{
|
| 319 |
+
"entropy": 0.5853322166385073,
|
| 320 |
+
"epoch": 3.063803680981595,
|
| 321 |
+
"grad_norm": 0.7313649654388428,
|
| 322 |
+
"learning_rate": 0.00035356377837808084,
|
| 323 |
+
"loss": 0.5616408538818359,
|
| 324 |
+
"mean_token_accuracy": 0.8144418639366073,
|
| 325 |
+
"num_tokens": 3312797.0,
|
| 326 |
+
"step": 1250
|
| 327 |
+
},
|
| 328 |
+
{
|
| 329 |
+
"entropy": 0.5654504173994064,
|
| 330 |
+
"epoch": 3.1865030674846624,
|
| 331 |
+
"grad_norm": 0.713893711566925,
|
| 332 |
+
"learning_rate": 0.0003477378508994529,
|
| 333 |
+
"loss": 0.5305458068847656,
|
| 334 |
+
"mean_token_accuracy": 0.8218590825796127,
|
| 335 |
+
"num_tokens": 3441629.0,
|
| 336 |
+
"step": 1300
|
| 337 |
+
},
|
| 338 |
+
{
|
| 339 |
+
"entropy": 0.5721602493524551,
|
| 340 |
+
"epoch": 3.30920245398773,
|
| 341 |
+
"grad_norm": 0.8487221002578735,
|
| 342 |
+
"learning_rate": 0.00034164489309687927,
|
| 343 |
+
"loss": 0.5437272262573242,
|
| 344 |
+
"mean_token_accuracy": 0.8169907212257386,
|
| 345 |
+
"num_tokens": 3576093.0,
|
| 346 |
+
"step": 1350
|
| 347 |
+
},
|
| 348 |
+
{
|
| 349 |
+
"entropy": 0.5797226822376251,
|
| 350 |
+
"epoch": 3.4319018404907977,
|
| 351 |
+
"grad_norm": 0.9307994842529297,
|
| 352 |
+
"learning_rate": 0.0003352960529547267,
|
| 353 |
+
"loss": 0.5405152130126953,
|
| 354 |
+
"mean_token_accuracy": 0.818563597202301,
|
| 355 |
+
"num_tokens": 3705912.0,
|
| 356 |
+
"step": 1400
|
| 357 |
+
},
|
| 358 |
+
{
|
| 359 |
+
"entropy": 0.5822264245152473,
|
| 360 |
+
"epoch": 3.554601226993865,
|
| 361 |
+
"grad_norm": 0.71127849817276,
|
| 362 |
+
"learning_rate": 0.00032870294663265757,
|
| 363 |
+
"loss": 0.5559784698486329,
|
| 364 |
+
"mean_token_accuracy": 0.8159428876638413,
|
| 365 |
+
"num_tokens": 3832376.0,
|
| 366 |
+
"step": 1450
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"entropy": 0.5702577340602875,
|
| 370 |
+
"epoch": 3.6773006134969326,
|
| 371 |
+
"grad_norm": 0.8090758919715881,
|
| 372 |
+
"learning_rate": 0.0003218776372121157,
|
| 373 |
+
"loss": 0.5466982650756836,
|
| 374 |
+
"mean_token_accuracy": 0.8186432421207428,
|
| 375 |
+
"num_tokens": 3967273.0,
|
| 376 |
+
"step": 1500
|
| 377 |
+
},
|
| 378 |
+
{
|
| 379 |
+
"entropy": 0.5708746567368508,
|
| 380 |
+
"epoch": 3.8,
|
| 381 |
+
"grad_norm": 0.7777291536331177,
|
| 382 |
+
"learning_rate": 0.000314832612625101,
|
| 383 |
+
"loss": 0.5412663650512696,
|
| 384 |
+
"mean_token_accuracy": 0.8195344746112824,
|
| 385 |
+
"num_tokens": 4102434.0,
|
| 386 |
+
"step": 1550
|
| 387 |
+
},
|
| 388 |
+
{
|
| 389 |
+
"entropy": 0.578739655315876,
|
| 390 |
+
"epoch": 3.9226993865030675,
|
| 391 |
+
"grad_norm": 0.8103565573692322,
|
| 392 |
+
"learning_rate": 0.0003075807628056167,
|
| 393 |
+
"loss": 0.5605202865600586,
|
| 394 |
+
"mean_token_accuracy": 0.815134603381157,
|
| 395 |
+
"num_tokens": 4234037.0,
|
| 396 |
+
"step": 1600
|
| 397 |
+
},
|
| 398 |
+
{
|
| 399 |
+
"epoch": 4.0,
|
| 400 |
+
"eval_entropy": 0.6183825404303415,
|
| 401 |
+
"eval_mean_token_accuracy": 0.7838981601170131,
|
| 402 |
+
"eval_not_syn_loss": 0.6814553737640381,
|
| 403 |
+
"eval_not_syn_runtime": 56.4858,
|
| 404 |
+
"eval_not_syn_samples_per_second": 24.714,
|
| 405 |
+
"eval_not_syn_steps_per_second": 3.098,
|
| 406 |
+
"eval_num_tokens": 4320784.0,
|
| 407 |
+
"step": 1632
|
| 408 |
+
},
|
| 409 |
+
{
|
| 410 |
+
"epoch": 4.0,
|
| 411 |
+
"eval_entropy": 0.5816301565510886,
|
| 412 |
+
"eval_mean_token_accuracy": 0.8187636685371399,
|
| 413 |
+
"eval_num_tokens": 4320784.0,
|
| 414 |
+
"eval_syn_loss": 0.6366032361984253,
|
| 415 |
+
"eval_syn_runtime": 58.1317,
|
| 416 |
+
"eval_syn_samples_per_second": 24.014,
|
| 417 |
+
"eval_syn_steps_per_second": 3.01,
|
| 418 |
+
"step": 1632
|
| 419 |
+
},
|
| 420 |
+
{
|
| 421 |
+
"entropy": 0.5312421954039371,
|
| 422 |
+
"epoch": 4.044171779141104,
|
| 423 |
+
"grad_norm": 0.6530773043632507,
|
| 424 |
+
"learning_rate": 0.00030013535610559226,
|
| 425 |
+
"loss": 0.5023544311523438,
|
| 426 |
+
"mean_token_accuracy": 0.8297024351177793,
|
| 427 |
+
"num_tokens": 4368268.0,
|
| 428 |
+
"step": 1650
|
| 429 |
+
},
|
| 430 |
+
{
|
| 431 |
+
"entropy": 0.4853435277938843,
|
| 432 |
+
"epoch": 4.166871165644172,
|
| 433 |
+
"grad_norm": 1.0383695363998413,
|
| 434 |
+
"learning_rate": 0.00029251001501843485,
|
| 435 |
+
"loss": 0.45016471862792967,
|
| 436 |
+
"mean_token_accuracy": 0.8418685424327851,
|
| 437 |
+
"num_tokens": 4494525.0,
|
| 438 |
+
"step": 1700
|
| 439 |
+
},
|
| 440 |
+
{
|
| 441 |
+
"entropy": 0.48153585255146025,
|
| 442 |
+
"epoch": 4.289570552147239,
|
| 443 |
+
"grad_norm": 0.7215332388877869,
|
| 444 |
+
"learning_rate": 0.00028471869125462477,
|
| 445 |
+
"loss": 0.4460752487182617,
|
| 446 |
+
"mean_token_accuracy": 0.8421206694841384,
|
| 447 |
+
"num_tokens": 4631985.0,
|
| 448 |
+
"step": 1750
|
| 449 |
+
},
|
| 450 |
+
{
|
| 451 |
+
"entropy": 0.49278041124343874,
|
| 452 |
+
"epoch": 4.412269938650307,
|
| 453 |
+
"grad_norm": 0.7771472334861755,
|
| 454 |
+
"learning_rate": 0.0002767756402149588,
|
| 455 |
+
"loss": 0.45794551849365234,
|
| 456 |
+
"mean_token_accuracy": 0.8404830145835877,
|
| 457 |
+
"num_tokens": 4766323.0,
|
| 458 |
+
"step": 1800
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"entropy": 0.5082326257228851,
|
| 462 |
+
"epoch": 4.534969325153375,
|
| 463 |
+
"grad_norm": 0.9560676217079163,
|
| 464 |
+
"learning_rate": 0.00026869539490814704,
|
| 465 |
+
"loss": 0.4632451629638672,
|
| 466 |
+
"mean_token_accuracy": 0.8386451864242553,
|
| 467 |
+
"num_tokens": 4895606.0,
|
| 468 |
+
"step": 1850
|
| 469 |
+
},
|
| 470 |
+
{
|
| 471 |
+
"entropy": 0.5081156292557716,
|
| 472 |
+
"epoch": 4.6576687116564415,
|
| 473 |
+
"grad_norm": 0.7719491720199585,
|
| 474 |
+
"learning_rate": 0.0002604927393604828,
|
| 475 |
+
"loss": 0.4630893325805664,
|
| 476 |
+
"mean_token_accuracy": 0.839306333065033,
|
| 477 |
+
"num_tokens": 5025912.0,
|
| 478 |
+
"step": 1900
|
| 479 |
+
},
|
| 480 |
+
{
|
| 481 |
+
"entropy": 0.5121548187732696,
|
| 482 |
+
"epoch": 4.780368098159509,
|
| 483 |
+
"grad_norm": 0.8171904683113098,
|
| 484 |
+
"learning_rate": 0.00025218268156623985,
|
| 485 |
+
"loss": 0.46541053771972657,
|
| 486 |
+
"mean_token_accuracy": 0.8369353520870209,
|
| 487 |
+
"num_tokens": 5162972.0,
|
| 488 |
+
"step": 1950
|
| 489 |
+
},
|
| 490 |
+
{
|
| 491 |
+
"entropy": 0.5085560208559037,
|
| 492 |
+
"epoch": 4.903067484662577,
|
| 493 |
+
"grad_norm": 0.5919133424758911,
|
| 494 |
+
"learning_rate": 0.00024378042602828492,
|
| 495 |
+
"loss": 0.46405506134033203,
|
| 496 |
+
"mean_token_accuracy": 0.8379521882534027,
|
| 497 |
+
"num_tokens": 5297935.0,
|
| 498 |
+
"step": 2000
|
| 499 |
+
},
|
| 500 |
+
{
|
| 501 |
+
"epoch": 5.0,
|
| 502 |
+
"eval_entropy": 0.5431888212476458,
|
| 503 |
+
"eval_mean_token_accuracy": 0.7749751152311053,
|
| 504 |
+
"eval_not_syn_loss": 0.7035888433456421,
|
| 505 |
+
"eval_not_syn_runtime": 56.5207,
|
| 506 |
+
"eval_not_syn_samples_per_second": 24.699,
|
| 507 |
+
"eval_not_syn_steps_per_second": 3.096,
|
| 508 |
+
"eval_num_tokens": 5400980.0,
|
| 509 |
+
"step": 2040
|
| 510 |
+
},
|
| 511 |
+
{
|
| 512 |
+
"epoch": 5.0,
|
| 513 |
+
"eval_entropy": 0.5098190077713558,
|
| 514 |
+
"eval_mean_token_accuracy": 0.832349077633449,
|
| 515 |
+
"eval_num_tokens": 5400980.0,
|
| 516 |
+
"eval_syn_loss": 0.6227898597717285,
|
| 517 |
+
"eval_syn_runtime": 57.6978,
|
| 518 |
+
"eval_syn_samples_per_second": 24.195,
|
| 519 |
+
"eval_syn_steps_per_second": 3.033,
|
| 520 |
+
"step": 2040
|
| 521 |
+
}
|
| 522 |
+
],
|
| 523 |
+
"logging_steps": 50,
|
| 524 |
+
"max_steps": 4080,
|
| 525 |
+
"num_input_tokens_seen": 0,
|
| 526 |
+
"num_train_epochs": 10,
|
| 527 |
+
"save_steps": 500,
|
| 528 |
+
"stateful_callbacks": {
|
| 529 |
+
"TrainerControl": {
|
| 530 |
+
"args": {
|
| 531 |
+
"should_epoch_stop": false,
|
| 532 |
+
"should_evaluate": false,
|
| 533 |
+
"should_log": false,
|
| 534 |
+
"should_save": true,
|
| 535 |
+
"should_training_stop": false
|
| 536 |
+
},
|
| 537 |
+
"attributes": {}
|
| 538 |
+
}
|
| 539 |
+
},
|
| 540 |
+
"total_flos": 1.1014210793408448e+17,
|
| 541 |
+
"train_batch_size": 8,
|
| 542 |
+
"trial_name": null,
|
| 543 |
+
"trial_params": null
|
| 544 |
+
}
|
substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-2448/README.md
ADDED
|
@@ -0,0 +1,209 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
base_model: Qwen/Qwen3.5-2B-Base
|
| 3 |
+
library_name: peft
|
| 4 |
+
pipeline_tag: text-generation
|
| 5 |
+
tags:
|
| 6 |
+
- base_model:adapter:Qwen/Qwen3.5-2B-Base
|
| 7 |
+
- lora
|
| 8 |
+
- sft
|
| 9 |
+
- transformers
|
| 10 |
+
- trl
|
| 11 |
+
---
|
| 12 |
+
|
| 13 |
+
# Model Card for Model ID
|
| 14 |
+
|
| 15 |
+
<!-- Provide a quick summary of what the model is/does. -->
|
| 16 |
+
|
| 17 |
+
|
| 18 |
+
|
| 19 |
+
## Model Details
|
| 20 |
+
|
| 21 |
+
### Model Description
|
| 22 |
+
|
| 23 |
+
<!-- Provide a longer summary of what this model is. -->
|
| 24 |
+
|
| 25 |
+
|
| 26 |
+
|
| 27 |
+
- **Developed by:** [More Information Needed]
|
| 28 |
+
- **Funded by [optional]:** [More Information Needed]
|
| 29 |
+
- **Shared by [optional]:** [More Information Needed]
|
| 30 |
+
- **Model type:** [More Information Needed]
|
| 31 |
+
- **Language(s) (NLP):** [More Information Needed]
|
| 32 |
+
- **License:** [More Information Needed]
|
| 33 |
+
- **Finetuned from model [optional]:** [More Information Needed]
|
| 34 |
+
|
| 35 |
+
### Model Sources [optional]
|
| 36 |
+
|
| 37 |
+
<!-- Provide the basic links for the model. -->
|
| 38 |
+
|
| 39 |
+
- **Repository:** [More Information Needed]
|
| 40 |
+
- **Paper [optional]:** [More Information Needed]
|
| 41 |
+
- **Demo [optional]:** [More Information Needed]
|
| 42 |
+
|
| 43 |
+
## Uses
|
| 44 |
+
|
| 45 |
+
<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
|
| 46 |
+
|
| 47 |
+
### Direct Use
|
| 48 |
+
|
| 49 |
+
<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
|
| 50 |
+
|
| 51 |
+
[More Information Needed]
|
| 52 |
+
|
| 53 |
+
### Downstream Use [optional]
|
| 54 |
+
|
| 55 |
+
<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
|
| 56 |
+
|
| 57 |
+
[More Information Needed]
|
| 58 |
+
|
| 59 |
+
### Out-of-Scope Use
|
| 60 |
+
|
| 61 |
+
<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
|
| 62 |
+
|
| 63 |
+
[More Information Needed]
|
| 64 |
+
|
| 65 |
+
## Bias, Risks, and Limitations
|
| 66 |
+
|
| 67 |
+
<!-- This section is meant to convey both technical and sociotechnical limitations. -->
|
| 68 |
+
|
| 69 |
+
[More Information Needed]
|
| 70 |
+
|
| 71 |
+
### Recommendations
|
| 72 |
+
|
| 73 |
+
<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
|
| 74 |
+
|
| 75 |
+
Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
|
| 76 |
+
|
| 77 |
+
## How to Get Started with the Model
|
| 78 |
+
|
| 79 |
+
Use the code below to get started with the model.
|
| 80 |
+
|
| 81 |
+
[More Information Needed]
|
| 82 |
+
|
| 83 |
+
## Training Details
|
| 84 |
+
|
| 85 |
+
### Training Data
|
| 86 |
+
|
| 87 |
+
<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
|
| 88 |
+
|
| 89 |
+
[More Information Needed]
|
| 90 |
+
|
| 91 |
+
### Training Procedure
|
| 92 |
+
|
| 93 |
+
<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
|
| 94 |
+
|
| 95 |
+
#### Preprocessing [optional]
|
| 96 |
+
|
| 97 |
+
[More Information Needed]
|
| 98 |
+
|
| 99 |
+
|
| 100 |
+
#### Training Hyperparameters
|
| 101 |
+
|
| 102 |
+
- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
|
| 103 |
+
|
| 104 |
+
#### Speeds, Sizes, Times [optional]
|
| 105 |
+
|
| 106 |
+
<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
|
| 107 |
+
|
| 108 |
+
[More Information Needed]
|
| 109 |
+
|
| 110 |
+
## Evaluation
|
| 111 |
+
|
| 112 |
+
<!-- This section describes the evaluation protocols and provides the results. -->
|
| 113 |
+
|
| 114 |
+
### Testing Data, Factors & Metrics
|
| 115 |
+
|
| 116 |
+
#### Testing Data
|
| 117 |
+
|
| 118 |
+
<!-- This should link to a Dataset Card if possible. -->
|
| 119 |
+
|
| 120 |
+
[More Information Needed]
|
| 121 |
+
|
| 122 |
+
#### Factors
|
| 123 |
+
|
| 124 |
+
<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
|
| 125 |
+
|
| 126 |
+
[More Information Needed]
|
| 127 |
+
|
| 128 |
+
#### Metrics
|
| 129 |
+
|
| 130 |
+
<!-- These are the evaluation metrics being used, ideally with a description of why. -->
|
| 131 |
+
|
| 132 |
+
[More Information Needed]
|
| 133 |
+
|
| 134 |
+
### Results
|
| 135 |
+
|
| 136 |
+
[More Information Needed]
|
| 137 |
+
|
| 138 |
+
#### Summary
|
| 139 |
+
|
| 140 |
+
|
| 141 |
+
|
| 142 |
+
## Model Examination [optional]
|
| 143 |
+
|
| 144 |
+
<!-- Relevant interpretability work for the model goes here -->
|
| 145 |
+
|
| 146 |
+
[More Information Needed]
|
| 147 |
+
|
| 148 |
+
## Environmental Impact
|
| 149 |
+
|
| 150 |
+
<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
|
| 151 |
+
|
| 152 |
+
Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
|
| 153 |
+
|
| 154 |
+
- **Hardware Type:** [More Information Needed]
|
| 155 |
+
- **Hours used:** [More Information Needed]
|
| 156 |
+
- **Cloud Provider:** [More Information Needed]
|
| 157 |
+
- **Compute Region:** [More Information Needed]
|
| 158 |
+
- **Carbon Emitted:** [More Information Needed]
|
| 159 |
+
|
| 160 |
+
## Technical Specifications [optional]
|
| 161 |
+
|
| 162 |
+
### Model Architecture and Objective
|
| 163 |
+
|
| 164 |
+
[More Information Needed]
|
| 165 |
+
|
| 166 |
+
### Compute Infrastructure
|
| 167 |
+
|
| 168 |
+
[More Information Needed]
|
| 169 |
+
|
| 170 |
+
#### Hardware
|
| 171 |
+
|
| 172 |
+
[More Information Needed]
|
| 173 |
+
|
| 174 |
+
#### Software
|
| 175 |
+
|
| 176 |
+
[More Information Needed]
|
| 177 |
+
|
| 178 |
+
## Citation [optional]
|
| 179 |
+
|
| 180 |
+
<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
|
| 181 |
+
|
| 182 |
+
**BibTeX:**
|
| 183 |
+
|
| 184 |
+
[More Information Needed]
|
| 185 |
+
|
| 186 |
+
**APA:**
|
| 187 |
+
|
| 188 |
+
[More Information Needed]
|
| 189 |
+
|
| 190 |
+
## Glossary [optional]
|
| 191 |
+
|
| 192 |
+
<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
|
| 193 |
+
|
| 194 |
+
[More Information Needed]
|
| 195 |
+
|
| 196 |
+
## More Information [optional]
|
| 197 |
+
|
| 198 |
+
[More Information Needed]
|
| 199 |
+
|
| 200 |
+
## Model Card Authors [optional]
|
| 201 |
+
|
| 202 |
+
[More Information Needed]
|
| 203 |
+
|
| 204 |
+
## Model Card Contact
|
| 205 |
+
|
| 206 |
+
[More Information Needed]
|
| 207 |
+
### Framework versions
|
| 208 |
+
|
| 209 |
+
- PEFT 0.19.1
|
substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-2448/adapter_config.json
ADDED
|
@@ -0,0 +1,48 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"alora_invocation_tokens": null,
|
| 3 |
+
"alpha_pattern": {},
|
| 4 |
+
"arrow_config": null,
|
| 5 |
+
"auto_mapping": null,
|
| 6 |
+
"base_model_name_or_path": "Qwen/Qwen3.5-2B-Base",
|
| 7 |
+
"bias": "none",
|
| 8 |
+
"corda_config": null,
|
| 9 |
+
"ensure_weight_tying": false,
|
| 10 |
+
"eva_config": null,
|
| 11 |
+
"exclude_modules": null,
|
| 12 |
+
"fan_in_fan_out": false,
|
| 13 |
+
"inference_mode": true,
|
| 14 |
+
"init_lora_weights": true,
|
| 15 |
+
"layer_replication": null,
|
| 16 |
+
"layers_pattern": null,
|
| 17 |
+
"layers_to_transform": null,
|
| 18 |
+
"loftq_config": {},
|
| 19 |
+
"lora_alpha": 128,
|
| 20 |
+
"lora_bias": false,
|
| 21 |
+
"lora_dropout": 0.05635041589438875,
|
| 22 |
+
"lora_ga_config": null,
|
| 23 |
+
"megatron_config": null,
|
| 24 |
+
"megatron_core": "megatron.core",
|
| 25 |
+
"modules_to_save": null,
|
| 26 |
+
"peft_type": "LORA",
|
| 27 |
+
"peft_version": "0.19.1",
|
| 28 |
+
"qalora_group_size": 16,
|
| 29 |
+
"r": 64,
|
| 30 |
+
"rank_pattern": {},
|
| 31 |
+
"revision": null,
|
| 32 |
+
"target_modules": [
|
| 33 |
+
"q_proj",
|
| 34 |
+
"k_proj",
|
| 35 |
+
"gate_proj",
|
| 36 |
+
"up_proj",
|
| 37 |
+
"o_proj",
|
| 38 |
+
"down_proj",
|
| 39 |
+
"v_proj"
|
| 40 |
+
],
|
| 41 |
+
"target_parameters": null,
|
| 42 |
+
"task_type": "CAUSAL_LM",
|
| 43 |
+
"trainable_token_indices": null,
|
| 44 |
+
"use_bdlora": null,
|
| 45 |
+
"use_dora": false,
|
| 46 |
+
"use_qalora": false,
|
| 47 |
+
"use_rslora": false
|
| 48 |
+
}
|
substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-2448/chat_template.jinja
ADDED
|
@@ -0,0 +1,154 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{%- set image_count = namespace(value=0) %}
|
| 2 |
+
{%- set video_count = namespace(value=0) %}
|
| 3 |
+
{%- macro render_content(content, do_vision_count, is_system_content=false) %}
|
| 4 |
+
{%- if content is string %}
|
| 5 |
+
{{- content }}
|
| 6 |
+
{%- elif content is iterable and content is not mapping %}
|
| 7 |
+
{%- for item in content %}
|
| 8 |
+
{%- if 'image' in item or 'image_url' in item or item.type == 'image' %}
|
| 9 |
+
{%- if is_system_content %}
|
| 10 |
+
{{- raise_exception('System message cannot contain images.') }}
|
| 11 |
+
{%- endif %}
|
| 12 |
+
{%- if do_vision_count %}
|
| 13 |
+
{%- set image_count.value = image_count.value + 1 %}
|
| 14 |
+
{%- endif %}
|
| 15 |
+
{%- if add_vision_id %}
|
| 16 |
+
{{- 'Picture ' ~ image_count.value ~ ': ' }}
|
| 17 |
+
{%- endif %}
|
| 18 |
+
{{- '<|vision_start|><|image_pad|><|vision_end|>' }}
|
| 19 |
+
{%- elif 'video' in item or item.type == 'video' %}
|
| 20 |
+
{%- if is_system_content %}
|
| 21 |
+
{{- raise_exception('System message cannot contain videos.') }}
|
| 22 |
+
{%- endif %}
|
| 23 |
+
{%- if do_vision_count %}
|
| 24 |
+
{%- set video_count.value = video_count.value + 1 %}
|
| 25 |
+
{%- endif %}
|
| 26 |
+
{%- if add_vision_id %}
|
| 27 |
+
{{- 'Video ' ~ video_count.value ~ ': ' }}
|
| 28 |
+
{%- endif %}
|
| 29 |
+
{{- '<|vision_start|><|video_pad|><|vision_end|>' }}
|
| 30 |
+
{%- elif 'text' in item %}
|
| 31 |
+
{{- item.text }}
|
| 32 |
+
{%- else %}
|
| 33 |
+
{{- raise_exception('Unexpected item type in content.') }}
|
| 34 |
+
{%- endif %}
|
| 35 |
+
{%- endfor %}
|
| 36 |
+
{%- elif content is none or content is undefined %}
|
| 37 |
+
{{- '' }}
|
| 38 |
+
{%- else %}
|
| 39 |
+
{{- raise_exception('Unexpected content type.') }}
|
| 40 |
+
{%- endif %}
|
| 41 |
+
{%- endmacro %}
|
| 42 |
+
{%- if not messages %}
|
| 43 |
+
{{- raise_exception('No messages provided.') }}
|
| 44 |
+
{%- endif %}
|
| 45 |
+
{%- if tools and tools is iterable and tools is not mapping %}
|
| 46 |
+
{{- '<|im_start|>system\n' }}
|
| 47 |
+
{{- "# Tools\n\nYou have access to the following functions:\n\n<tools>" }}
|
| 48 |
+
{%- for tool in tools %}
|
| 49 |
+
{{- "\n" }}
|
| 50 |
+
{{- tool | tojson }}
|
| 51 |
+
{%- endfor %}
|
| 52 |
+
{{- "\n</tools>" }}
|
| 53 |
+
{{- '\n\nIf you choose to call a function ONLY reply in the following format with NO suffix:\n\n<tool_call>\n<function=example_function_name>\n<parameter=example_parameter_1>\nvalue_1\n</parameter>\n<parameter=example_parameter_2>\nThis is the value for the second parameter\nthat can span\nmultiple lines\n</parameter>\n</function>\n</tool_call>\n\n<IMPORTANT>\nReminder:\n- Function calls MUST follow the specified format: an inner <function=...></function> block must be nested within <tool_call></tool_call> XML tags\n- Required parameters MUST be specified\n- You may provide optional reasoning for your function call in natural language BEFORE the function call, but NOT after\n- If there is no function call available, answer the question like normal with your current knowledge and do not tell the user about function calls\n</IMPORTANT>' }}
|
| 54 |
+
{%- if messages[0].role == 'system' %}
|
| 55 |
+
{%- set content = render_content(messages[0].content, false, true)|trim %}
|
| 56 |
+
{%- if content %}
|
| 57 |
+
{{- '\n\n' + content }}
|
| 58 |
+
{%- endif %}
|
| 59 |
+
{%- endif %}
|
| 60 |
+
{{- '<|im_end|>\n' }}
|
| 61 |
+
{%- else %}
|
| 62 |
+
{%- if messages[0].role == 'system' %}
|
| 63 |
+
{%- set content = render_content(messages[0].content, false, true)|trim %}
|
| 64 |
+
{{- '<|im_start|>system\n' + content + '<|im_end|>\n' }}
|
| 65 |
+
{%- endif %}
|
| 66 |
+
{%- endif %}
|
| 67 |
+
{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
|
| 68 |
+
{%- for message in messages[::-1] %}
|
| 69 |
+
{%- set index = (messages|length - 1) - loop.index0 %}
|
| 70 |
+
{%- if ns.multi_step_tool and message.role == "user" %}
|
| 71 |
+
{%- set content = render_content(message.content, false)|trim %}
|
| 72 |
+
{%- if not(content.startswith('<tool_response>') and content.endswith('</tool_response>')) %}
|
| 73 |
+
{%- set ns.multi_step_tool = false %}
|
| 74 |
+
{%- set ns.last_query_index = index %}
|
| 75 |
+
{%- endif %}
|
| 76 |
+
{%- endif %}
|
| 77 |
+
{%- endfor %}
|
| 78 |
+
{%- if ns.multi_step_tool %}
|
| 79 |
+
{{- raise_exception('No user query found in messages.') }}
|
| 80 |
+
{%- endif %}
|
| 81 |
+
{%- for message in messages %}
|
| 82 |
+
{%- set content = render_content(message.content, true)|trim %}
|
| 83 |
+
{%- if message.role == "system" %}
|
| 84 |
+
{%- if not loop.first %}
|
| 85 |
+
{{- raise_exception('System message must be at the beginning.') }}
|
| 86 |
+
{%- endif %}
|
| 87 |
+
{%- elif message.role == "user" %}
|
| 88 |
+
{{- '<|im_start|>' + message.role + '\n' + content + '<|im_end|>' + '\n' }}
|
| 89 |
+
{%- elif message.role == "assistant" %}
|
| 90 |
+
{%- set reasoning_content = '' %}
|
| 91 |
+
{%- if message.reasoning_content is string %}
|
| 92 |
+
{%- set reasoning_content = message.reasoning_content %}
|
| 93 |
+
{%- else %}
|
| 94 |
+
{%- if '</think>' in content %}
|
| 95 |
+
{%- set reasoning_content = content.split('</think>')[0].rstrip('\n').split('<think>')[-1].lstrip('\n') %}
|
| 96 |
+
{%- set content = content.split('</think>')[-1].lstrip('\n') %}
|
| 97 |
+
{%- endif %}
|
| 98 |
+
{%- endif %}
|
| 99 |
+
{%- set reasoning_content = reasoning_content|trim %}
|
| 100 |
+
{%- if loop.index0 > ns.last_query_index %}
|
| 101 |
+
{{- '<|im_start|>' + message.role + '\n<think>\n' + reasoning_content + '\n</think>\n\n' + content }}
|
| 102 |
+
{%- else %}
|
| 103 |
+
{{- '<|im_start|>' + message.role + '\n' + content }}
|
| 104 |
+
{%- endif %}
|
| 105 |
+
{%- if message.tool_calls and message.tool_calls is iterable and message.tool_calls is not mapping %}
|
| 106 |
+
{%- for tool_call in message.tool_calls %}
|
| 107 |
+
{%- if tool_call.function is defined %}
|
| 108 |
+
{%- set tool_call = tool_call.function %}
|
| 109 |
+
{%- endif %}
|
| 110 |
+
{%- if loop.first %}
|
| 111 |
+
{%- if content|trim %}
|
| 112 |
+
{{- '\n\n<tool_call>\n<function=' + tool_call.name + '>\n' }}
|
| 113 |
+
{%- else %}
|
| 114 |
+
{{- '<tool_call>\n<function=' + tool_call.name + '>\n' }}
|
| 115 |
+
{%- endif %}
|
| 116 |
+
{%- else %}
|
| 117 |
+
{{- '\n<tool_call>\n<function=' + tool_call.name + '>\n' }}
|
| 118 |
+
{%- endif %}
|
| 119 |
+
{%- if tool_call.arguments is defined %}
|
| 120 |
+
{%- for args_name, args_value in tool_call.arguments|items %}
|
| 121 |
+
{{- '<parameter=' + args_name + '>\n' }}
|
| 122 |
+
{%- set args_value = args_value | tojson | safe if args_value is mapping or (args_value is sequence and args_value is not string) else args_value | string %}
|
| 123 |
+
{{- args_value }}
|
| 124 |
+
{{- '\n</parameter>\n' }}
|
| 125 |
+
{%- endfor %}
|
| 126 |
+
{%- endif %}
|
| 127 |
+
{{- '</function>\n</tool_call>' }}
|
| 128 |
+
{%- endfor %}
|
| 129 |
+
{%- endif %}
|
| 130 |
+
{{- '<|im_end|>\n' }}
|
| 131 |
+
{%- elif message.role == "tool" %}
|
| 132 |
+
{%- if loop.previtem and loop.previtem.role != "tool" %}
|
| 133 |
+
{{- '<|im_start|>user' }}
|
| 134 |
+
{%- endif %}
|
| 135 |
+
{{- '\n<tool_response>\n' }}
|
| 136 |
+
{{- content }}
|
| 137 |
+
{{- '\n</tool_response>' }}
|
| 138 |
+
{%- if not loop.last and loop.nextitem.role != "tool" %}
|
| 139 |
+
{{- '<|im_end|>\n' }}
|
| 140 |
+
{%- elif loop.last %}
|
| 141 |
+
{{- '<|im_end|>\n' }}
|
| 142 |
+
{%- endif %}
|
| 143 |
+
{%- else %}
|
| 144 |
+
{{- raise_exception('Unexpected message role.') }}
|
| 145 |
+
{%- endif %}
|
| 146 |
+
{%- endfor %}
|
| 147 |
+
{%- if add_generation_prompt %}
|
| 148 |
+
{{- '<|im_start|>assistant\n' }}
|
| 149 |
+
{%- if enable_thinking is defined and enable_thinking is true %}
|
| 150 |
+
{{- '<think>\n' }}
|
| 151 |
+
{%- else %}
|
| 152 |
+
{{- '<think>\n\n</think>\n\n' }}
|
| 153 |
+
{%- endif %}
|
| 154 |
+
{%- endif %}
|
substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-2448/tokenizer_config.json
ADDED
|
@@ -0,0 +1,31 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"add_prefix_space": false,
|
| 3 |
+
"audio_bos_token": "<|audio_start|>",
|
| 4 |
+
"audio_eos_token": "<|audio_end|>",
|
| 5 |
+
"audio_token": "<|audio_pad|>",
|
| 6 |
+
"backend": "tokenizers",
|
| 7 |
+
"bos_token": null,
|
| 8 |
+
"clean_up_tokenization_spaces": false,
|
| 9 |
+
"eos_token": "<|endoftext|>",
|
| 10 |
+
"errors": "replace",
|
| 11 |
+
"image_token": "<|image_pad|>",
|
| 12 |
+
"is_local": false,
|
| 13 |
+
"model_max_length": 262144,
|
| 14 |
+
"model_specific_special_tokens": {
|
| 15 |
+
"audio_bos_token": "<|audio_start|>",
|
| 16 |
+
"audio_eos_token": "<|audio_end|>",
|
| 17 |
+
"audio_token": "<|audio_pad|>",
|
| 18 |
+
"image_token": "<|image_pad|>",
|
| 19 |
+
"video_token": "<|video_pad|>",
|
| 20 |
+
"vision_bos_token": "<|vision_start|>",
|
| 21 |
+
"vision_eos_token": "<|vision_end|>"
|
| 22 |
+
},
|
| 23 |
+
"pad_token": "<|endoftext|>",
|
| 24 |
+
"pretokenize_regex": "(?i:'s|'t|'re|'ve|'m|'ll|'d)|[^\\r\\n\\p{L}\\p{N}]?[\\p{L}\\p{M}]+|\\p{N}| ?[^\\s\\p{L}\\p{M}\\p{N}]+[\\r\\n]*|\\s*[\\r\\n]+|\\s+(?!\\S)|\\s+",
|
| 25 |
+
"split_special_tokens": false,
|
| 26 |
+
"tokenizer_class": "TokenizersBackend",
|
| 27 |
+
"unk_token": null,
|
| 28 |
+
"video_token": "<|video_pad|>",
|
| 29 |
+
"vision_bos_token": "<|vision_start|>",
|
| 30 |
+
"vision_eos_token": "<|vision_end|>"
|
| 31 |
+
}
|
substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-2448/trainer_state.json
ADDED
|
@@ -0,0 +1,646 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": null,
|
| 3 |
+
"best_metric": null,
|
| 4 |
+
"best_model_checkpoint": null,
|
| 5 |
+
"epoch": 6.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 2448,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"entropy": 1.8153233402967452,
|
| 14 |
+
"epoch": 0.12269938650306748,
|
| 15 |
+
"grad_norm": 1.8510568141937256,
|
| 16 |
+
"learning_rate": 4.8469594667113335e-05,
|
| 17 |
+
"loss": 1.7217323303222656,
|
| 18 |
+
"mean_token_accuracy": 0.6236635231971741,
|
| 19 |
+
"num_tokens": 134329.0,
|
| 20 |
+
"step": 50
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"entropy": 0.97831438601017,
|
| 24 |
+
"epoch": 0.24539877300613497,
|
| 25 |
+
"grad_norm": 1.889868974685669,
|
| 26 |
+
"learning_rate": 9.792836473559631e-05,
|
| 27 |
+
"loss": 0.9050540924072266,
|
| 28 |
+
"mean_token_accuracy": 0.7377463465929032,
|
| 29 |
+
"num_tokens": 262407.0,
|
| 30 |
+
"step": 100
|
| 31 |
+
},
|
| 32 |
+
{
|
| 33 |
+
"entropy": 0.8743131357431412,
|
| 34 |
+
"epoch": 0.36809815950920244,
|
| 35 |
+
"grad_norm": 1.1766732931137085,
|
| 36 |
+
"learning_rate": 0.0001473871348040793,
|
| 37 |
+
"loss": 0.8129763031005859,
|
| 38 |
+
"mean_token_accuracy": 0.7583417356014251,
|
| 39 |
+
"num_tokens": 401136.0,
|
| 40 |
+
"step": 150
|
| 41 |
+
},
|
| 42 |
+
{
|
| 43 |
+
"entropy": 0.8273966419696808,
|
| 44 |
+
"epoch": 0.49079754601226994,
|
| 45 |
+
"grad_norm": 1.1311014890670776,
|
| 46 |
+
"learning_rate": 0.00019684590487256229,
|
| 47 |
+
"loss": 0.7711280822753906,
|
| 48 |
+
"mean_token_accuracy": 0.7668228060007095,
|
| 49 |
+
"num_tokens": 531647.0,
|
| 50 |
+
"step": 200
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"entropy": 0.8101438617706299,
|
| 54 |
+
"epoch": 0.6134969325153374,
|
| 55 |
+
"grad_norm": 0.8905950784683228,
|
| 56 |
+
"learning_rate": 0.0002463046749410453,
|
| 57 |
+
"loss": 0.7609222412109375,
|
| 58 |
+
"mean_token_accuracy": 0.7717174577713013,
|
| 59 |
+
"num_tokens": 666127.0,
|
| 60 |
+
"step": 250
|
| 61 |
+
},
|
| 62 |
+
{
|
| 63 |
+
"entropy": 0.7885055243968964,
|
| 64 |
+
"epoch": 0.7361963190184049,
|
| 65 |
+
"grad_norm": 1.1944245100021362,
|
| 66 |
+
"learning_rate": 0.00029576344500952824,
|
| 67 |
+
"loss": 0.7477743530273437,
|
| 68 |
+
"mean_token_accuracy": 0.773262197971344,
|
| 69 |
+
"num_tokens": 799292.0,
|
| 70 |
+
"step": 300
|
| 71 |
+
},
|
| 72 |
+
{
|
| 73 |
+
"entropy": 0.7858527797460556,
|
| 74 |
+
"epoch": 0.8588957055214724,
|
| 75 |
+
"grad_norm": 1.1311068534851074,
|
| 76 |
+
"learning_rate": 0.00034522221507801124,
|
| 77 |
+
"loss": 0.7411377716064453,
|
| 78 |
+
"mean_token_accuracy": 0.7745680212974548,
|
| 79 |
+
"num_tokens": 930614.0,
|
| 80 |
+
"step": 350
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"entropy": 0.7879797518253326,
|
| 84 |
+
"epoch": 0.9815950920245399,
|
| 85 |
+
"grad_norm": 0.9370436072349548,
|
| 86 |
+
"learning_rate": 0.00039468098514649423,
|
| 87 |
+
"loss": 0.7507785034179687,
|
| 88 |
+
"mean_token_accuracy": 0.7747444450855255,
|
| 89 |
+
"num_tokens": 1061654.0,
|
| 90 |
+
"step": 400
|
| 91 |
+
},
|
| 92 |
+
{
|
| 93 |
+
"epoch": 1.0,
|
| 94 |
+
"eval_entropy": 0.8527693561145238,
|
| 95 |
+
"eval_mean_token_accuracy": 0.7730878329277039,
|
| 96 |
+
"eval_not_syn_loss": 0.781921923160553,
|
| 97 |
+
"eval_not_syn_runtime": 56.9157,
|
| 98 |
+
"eval_not_syn_samples_per_second": 24.527,
|
| 99 |
+
"eval_not_syn_steps_per_second": 3.075,
|
| 100 |
+
"eval_num_tokens": 1080196.0,
|
| 101 |
+
"step": 408
|
| 102 |
+
},
|
| 103 |
+
{
|
| 104 |
+
"epoch": 1.0,
|
| 105 |
+
"eval_entropy": 0.7952005583899362,
|
| 106 |
+
"eval_mean_token_accuracy": 0.7705784467288426,
|
| 107 |
+
"eval_num_tokens": 1080196.0,
|
| 108 |
+
"eval_syn_loss": 0.7461121082305908,
|
| 109 |
+
"eval_syn_runtime": 59.5758,
|
| 110 |
+
"eval_syn_samples_per_second": 23.432,
|
| 111 |
+
"eval_syn_steps_per_second": 2.937,
|
| 112 |
+
"step": 408
|
| 113 |
+
},
|
| 114 |
+
{
|
| 115 |
+
"entropy": 0.7564875923021875,
|
| 116 |
+
"epoch": 1.1030674846625768,
|
| 117 |
+
"grad_norm": 1.1334599256515503,
|
| 118 |
+
"learning_rate": 0.00040345942966973637,
|
| 119 |
+
"loss": 0.7208955383300781,
|
| 120 |
+
"mean_token_accuracy": 0.7788144470465304,
|
| 121 |
+
"num_tokens": 1199170.0,
|
| 122 |
+
"step": 450
|
| 123 |
+
},
|
| 124 |
+
{
|
| 125 |
+
"entropy": 0.7427570706605912,
|
| 126 |
+
"epoch": 1.2257668711656442,
|
| 127 |
+
"grad_norm": 1.0777021646499634,
|
| 128 |
+
"learning_rate": 0.00040297229629224153,
|
| 129 |
+
"loss": 0.7049168395996094,
|
| 130 |
+
"mean_token_accuracy": 0.7830300116539002,
|
| 131 |
+
"num_tokens": 1334298.0,
|
| 132 |
+
"step": 500
|
| 133 |
+
},
|
| 134 |
+
{
|
| 135 |
+
"entropy": 0.7533353108167649,
|
| 136 |
+
"epoch": 1.3484662576687116,
|
| 137 |
+
"grad_norm": 1.3439561128616333,
|
| 138 |
+
"learning_rate": 0.00040211707285041045,
|
| 139 |
+
"loss": 0.7102288818359375,
|
| 140 |
+
"mean_token_accuracy": 0.7818449640274048,
|
| 141 |
+
"num_tokens": 1463189.0,
|
| 142 |
+
"step": 550
|
| 143 |
+
},
|
| 144 |
+
{
|
| 145 |
+
"entropy": 0.7666342407464981,
|
| 146 |
+
"epoch": 1.471165644171779,
|
| 147 |
+
"grad_norm": 1.1329729557037354,
|
| 148 |
+
"learning_rate": 0.00040089532410439227,
|
| 149 |
+
"loss": 0.7270996856689453,
|
| 150 |
+
"mean_token_accuracy": 0.778404277563095,
|
| 151 |
+
"num_tokens": 1590310.0,
|
| 152 |
+
"step": 600
|
| 153 |
+
},
|
| 154 |
+
{
|
| 155 |
+
"entropy": 0.7359934949874878,
|
| 156 |
+
"epoch": 1.5938650306748468,
|
| 157 |
+
"grad_norm": 1.2227331399917603,
|
| 158 |
+
"learning_rate": 0.0003993092854276068,
|
| 159 |
+
"loss": 0.6996555328369141,
|
| 160 |
+
"mean_token_accuracy": 0.7834958267211914,
|
| 161 |
+
"num_tokens": 1721836.0,
|
| 162 |
+
"step": 650
|
| 163 |
+
},
|
| 164 |
+
{
|
| 165 |
+
"entropy": 0.7278932738304138,
|
| 166 |
+
"epoch": 1.716564417177914,
|
| 167 |
+
"grad_norm": 0.8753179311752319,
|
| 168 |
+
"learning_rate": 0.0003973618587167924,
|
| 169 |
+
"loss": 0.694002456665039,
|
| 170 |
+
"mean_token_accuracy": 0.7849677371978759,
|
| 171 |
+
"num_tokens": 1853507.0,
|
| 172 |
+
"step": 700
|
| 173 |
+
},
|
| 174 |
+
{
|
| 175 |
+
"entropy": 0.7190143716335297,
|
| 176 |
+
"epoch": 1.8392638036809816,
|
| 177 |
+
"grad_norm": 0.8524742126464844,
|
| 178 |
+
"learning_rate": 0.0003950566070825483,
|
| 179 |
+
"loss": 0.6890143585205079,
|
| 180 |
+
"mean_token_accuracy": 0.7865350896120071,
|
| 181 |
+
"num_tokens": 1988113.0,
|
| 182 |
+
"step": 750
|
| 183 |
+
},
|
| 184 |
+
{
|
| 185 |
+
"entropy": 0.7072037667036056,
|
| 186 |
+
"epoch": 1.961963190184049,
|
| 187 |
+
"grad_norm": 0.877348005771637,
|
| 188 |
+
"learning_rate": 0.00039239774833008714,
|
| 189 |
+
"loss": 0.6795162200927735,
|
| 190 |
+
"mean_token_accuracy": 0.7885233855247498,
|
| 191 |
+
"num_tokens": 2121314.0,
|
| 192 |
+
"step": 800
|
| 193 |
+
},
|
| 194 |
+
{
|
| 195 |
+
"epoch": 2.0,
|
| 196 |
+
"eval_entropy": 0.818787784235818,
|
| 197 |
+
"eval_mean_token_accuracy": 0.7573730857031686,
|
| 198 |
+
"eval_not_syn_loss": 0.7486392855644226,
|
| 199 |
+
"eval_not_syn_runtime": 56.7426,
|
| 200 |
+
"eval_not_syn_samples_per_second": 24.602,
|
| 201 |
+
"eval_not_syn_steps_per_second": 3.084,
|
| 202 |
+
"eval_num_tokens": 2160392.0,
|
| 203 |
+
"step": 816
|
| 204 |
+
},
|
| 205 |
+
{
|
| 206 |
+
"epoch": 2.0,
|
| 207 |
+
"eval_entropy": 0.7632522354807173,
|
| 208 |
+
"eval_mean_token_accuracy": 0.8101163625717163,
|
| 209 |
+
"eval_num_tokens": 2160392.0,
|
| 210 |
+
"eval_syn_loss": 0.6829464435577393,
|
| 211 |
+
"eval_syn_runtime": 58.5255,
|
| 212 |
+
"eval_syn_samples_per_second": 23.853,
|
| 213 |
+
"eval_syn_steps_per_second": 2.99,
|
| 214 |
+
"step": 816
|
| 215 |
+
},
|
| 216 |
+
{
|
| 217 |
+
"entropy": 0.6619202801675508,
|
| 218 |
+
"epoch": 2.083435582822086,
|
| 219 |
+
"grad_norm": 0.8124546408653259,
|
| 220 |
+
"learning_rate": 0.00038939014724212456,
|
| 221 |
+
"loss": 0.6311377716064454,
|
| 222 |
+
"mean_token_accuracy": 0.7991988044796567,
|
| 223 |
+
"num_tokens": 2253179.0,
|
| 224 |
+
"step": 850
|
| 225 |
+
},
|
| 226 |
+
{
|
| 227 |
+
"entropy": 0.6668815296888352,
|
| 228 |
+
"epoch": 2.2061349693251535,
|
| 229 |
+
"grad_norm": 0.8693709969520569,
|
| 230 |
+
"learning_rate": 0.0003860393066780265,
|
| 231 |
+
"loss": 0.6279015731811524,
|
| 232 |
+
"mean_token_accuracy": 0.7983763587474823,
|
| 233 |
+
"num_tokens": 2379965.0,
|
| 234 |
+
"step": 900
|
| 235 |
+
},
|
| 236 |
+
{
|
| 237 |
+
"entropy": 0.6557431703805924,
|
| 238 |
+
"epoch": 2.3288343558282207,
|
| 239 |
+
"grad_norm": 0.8323768377304077,
|
| 240 |
+
"learning_rate": 0.0003823513575054986,
|
| 241 |
+
"loss": 0.6214478683471679,
|
| 242 |
+
"mean_token_accuracy": 0.8012632429599762,
|
| 243 |
+
"num_tokens": 2515593.0,
|
| 244 |
+
"step": 950
|
| 245 |
+
},
|
| 246 |
+
{
|
| 247 |
+
"entropy": 0.6633540654182434,
|
| 248 |
+
"epoch": 2.4515337423312884,
|
| 249 |
+
"grad_norm": 0.8328827023506165,
|
| 250 |
+
"learning_rate": 0.0003783330473832399,
|
| 251 |
+
"loss": 0.6310849761962891,
|
| 252 |
+
"mean_token_accuracy": 0.7995503985881806,
|
| 253 |
+
"num_tokens": 2642790.0,
|
| 254 |
+
"step": 1000
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"entropy": 0.6542247200012207,
|
| 258 |
+
"epoch": 2.574233128834356,
|
| 259 |
+
"grad_norm": 0.8204748630523682,
|
| 260 |
+
"learning_rate": 0.000373991728415085,
|
| 261 |
+
"loss": 0.6195787048339844,
|
| 262 |
+
"mean_token_accuracy": 0.8006457507610321,
|
| 263 |
+
"num_tokens": 2774396.0,
|
| 264 |
+
"step": 1050
|
| 265 |
+
},
|
| 266 |
+
{
|
| 267 |
+
"entropy": 0.6526653742790223,
|
| 268 |
+
"epoch": 2.6969325153374233,
|
| 269 |
+
"grad_norm": 1.065635323524475,
|
| 270 |
+
"learning_rate": 0.0003693353436982218,
|
| 271 |
+
"loss": 0.6206568145751953,
|
| 272 |
+
"mean_token_accuracy": 0.8002620726823807,
|
| 273 |
+
"num_tokens": 2912102.0,
|
| 274 |
+
"step": 1100
|
| 275 |
+
},
|
| 276 |
+
{
|
| 277 |
+
"entropy": 0.654406590461731,
|
| 278 |
+
"epoch": 2.819631901840491,
|
| 279 |
+
"grad_norm": 0.8653210401535034,
|
| 280 |
+
"learning_rate": 0.00036437241279009834,
|
| 281 |
+
"loss": 0.6259954452514649,
|
| 282 |
+
"mean_token_accuracy": 0.8008036535978317,
|
| 283 |
+
"num_tokens": 3042987.0,
|
| 284 |
+
"step": 1150
|
| 285 |
+
},
|
| 286 |
+
{
|
| 287 |
+
"entropy": 0.6310120761394501,
|
| 288 |
+
"epoch": 2.942331288343558,
|
| 289 |
+
"grad_norm": 0.7713245749473572,
|
| 290 |
+
"learning_rate": 0.0003591120161206093,
|
| 291 |
+
"loss": 0.6115898132324219,
|
| 292 |
+
"mean_token_accuracy": 0.8043569129705429,
|
| 293 |
+
"num_tokens": 3178044.0,
|
| 294 |
+
"step": 1200
|
| 295 |
+
},
|
| 296 |
+
{
|
| 297 |
+
"epoch": 3.0,
|
| 298 |
+
"eval_entropy": 0.6591688336644854,
|
| 299 |
+
"eval_mean_token_accuracy": 0.7992405581474304,
|
| 300 |
+
"eval_not_syn_loss": 0.6962469220161438,
|
| 301 |
+
"eval_not_syn_runtime": 56.161,
|
| 302 |
+
"eval_not_syn_samples_per_second": 24.857,
|
| 303 |
+
"eval_not_syn_steps_per_second": 3.116,
|
| 304 |
+
"eval_num_tokens": 3240588.0,
|
| 305 |
+
"step": 1224
|
| 306 |
+
},
|
| 307 |
+
{
|
| 308 |
+
"epoch": 3.0,
|
| 309 |
+
"eval_entropy": 0.6191406740461077,
|
| 310 |
+
"eval_mean_token_accuracy": 0.7810143651281084,
|
| 311 |
+
"eval_num_tokens": 3240588.0,
|
| 312 |
+
"eval_syn_loss": 0.6663697361946106,
|
| 313 |
+
"eval_syn_runtime": 58.4102,
|
| 314 |
+
"eval_syn_samples_per_second": 23.9,
|
| 315 |
+
"eval_syn_steps_per_second": 2.996,
|
| 316 |
+
"step": 1224
|
| 317 |
+
},
|
| 318 |
+
{
|
| 319 |
+
"entropy": 0.5853322166385073,
|
| 320 |
+
"epoch": 3.063803680981595,
|
| 321 |
+
"grad_norm": 0.7313649654388428,
|
| 322 |
+
"learning_rate": 0.00035356377837808084,
|
| 323 |
+
"loss": 0.5616408538818359,
|
| 324 |
+
"mean_token_accuracy": 0.8144418639366073,
|
| 325 |
+
"num_tokens": 3312797.0,
|
| 326 |
+
"step": 1250
|
| 327 |
+
},
|
| 328 |
+
{
|
| 329 |
+
"entropy": 0.5654504173994064,
|
| 330 |
+
"epoch": 3.1865030674846624,
|
| 331 |
+
"grad_norm": 0.713893711566925,
|
| 332 |
+
"learning_rate": 0.0003477378508994529,
|
| 333 |
+
"loss": 0.5305458068847656,
|
| 334 |
+
"mean_token_accuracy": 0.8218590825796127,
|
| 335 |
+
"num_tokens": 3441629.0,
|
| 336 |
+
"step": 1300
|
| 337 |
+
},
|
| 338 |
+
{
|
| 339 |
+
"entropy": 0.5721602493524551,
|
| 340 |
+
"epoch": 3.30920245398773,
|
| 341 |
+
"grad_norm": 0.8487221002578735,
|
| 342 |
+
"learning_rate": 0.00034164489309687927,
|
| 343 |
+
"loss": 0.5437272262573242,
|
| 344 |
+
"mean_token_accuracy": 0.8169907212257386,
|
| 345 |
+
"num_tokens": 3576093.0,
|
| 346 |
+
"step": 1350
|
| 347 |
+
},
|
| 348 |
+
{
|
| 349 |
+
"entropy": 0.5797226822376251,
|
| 350 |
+
"epoch": 3.4319018404907977,
|
| 351 |
+
"grad_norm": 0.9307994842529297,
|
| 352 |
+
"learning_rate": 0.0003352960529547267,
|
| 353 |
+
"loss": 0.5405152130126953,
|
| 354 |
+
"mean_token_accuracy": 0.818563597202301,
|
| 355 |
+
"num_tokens": 3705912.0,
|
| 356 |
+
"step": 1400
|
| 357 |
+
},
|
| 358 |
+
{
|
| 359 |
+
"entropy": 0.5822264245152473,
|
| 360 |
+
"epoch": 3.554601226993865,
|
| 361 |
+
"grad_norm": 0.71127849817276,
|
| 362 |
+
"learning_rate": 0.00032870294663265757,
|
| 363 |
+
"loss": 0.5559784698486329,
|
| 364 |
+
"mean_token_accuracy": 0.8159428876638413,
|
| 365 |
+
"num_tokens": 3832376.0,
|
| 366 |
+
"step": 1450
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"entropy": 0.5702577340602875,
|
| 370 |
+
"epoch": 3.6773006134969326,
|
| 371 |
+
"grad_norm": 0.8090758919715881,
|
| 372 |
+
"learning_rate": 0.0003218776372121157,
|
| 373 |
+
"loss": 0.5466982650756836,
|
| 374 |
+
"mean_token_accuracy": 0.8186432421207428,
|
| 375 |
+
"num_tokens": 3967273.0,
|
| 376 |
+
"step": 1500
|
| 377 |
+
},
|
| 378 |
+
{
|
| 379 |
+
"entropy": 0.5708746567368508,
|
| 380 |
+
"epoch": 3.8,
|
| 381 |
+
"grad_norm": 0.7777291536331177,
|
| 382 |
+
"learning_rate": 0.000314832612625101,
|
| 383 |
+
"loss": 0.5412663650512696,
|
| 384 |
+
"mean_token_accuracy": 0.8195344746112824,
|
| 385 |
+
"num_tokens": 4102434.0,
|
| 386 |
+
"step": 1550
|
| 387 |
+
},
|
| 388 |
+
{
|
| 389 |
+
"entropy": 0.578739655315876,
|
| 390 |
+
"epoch": 3.9226993865030675,
|
| 391 |
+
"grad_norm": 0.8103565573692322,
|
| 392 |
+
"learning_rate": 0.0003075807628056167,
|
| 393 |
+
"loss": 0.5605202865600586,
|
| 394 |
+
"mean_token_accuracy": 0.815134603381157,
|
| 395 |
+
"num_tokens": 4234037.0,
|
| 396 |
+
"step": 1600
|
| 397 |
+
},
|
| 398 |
+
{
|
| 399 |
+
"epoch": 4.0,
|
| 400 |
+
"eval_entropy": 0.6183825404303415,
|
| 401 |
+
"eval_mean_token_accuracy": 0.7838981601170131,
|
| 402 |
+
"eval_not_syn_loss": 0.6814553737640381,
|
| 403 |
+
"eval_not_syn_runtime": 56.4858,
|
| 404 |
+
"eval_not_syn_samples_per_second": 24.714,
|
| 405 |
+
"eval_not_syn_steps_per_second": 3.098,
|
| 406 |
+
"eval_num_tokens": 4320784.0,
|
| 407 |
+
"step": 1632
|
| 408 |
+
},
|
| 409 |
+
{
|
| 410 |
+
"epoch": 4.0,
|
| 411 |
+
"eval_entropy": 0.5816301565510886,
|
| 412 |
+
"eval_mean_token_accuracy": 0.8187636685371399,
|
| 413 |
+
"eval_num_tokens": 4320784.0,
|
| 414 |
+
"eval_syn_loss": 0.6366032361984253,
|
| 415 |
+
"eval_syn_runtime": 58.1317,
|
| 416 |
+
"eval_syn_samples_per_second": 24.014,
|
| 417 |
+
"eval_syn_steps_per_second": 3.01,
|
| 418 |
+
"step": 1632
|
| 419 |
+
},
|
| 420 |
+
{
|
| 421 |
+
"entropy": 0.5312421954039371,
|
| 422 |
+
"epoch": 4.044171779141104,
|
| 423 |
+
"grad_norm": 0.6530773043632507,
|
| 424 |
+
"learning_rate": 0.00030013535610559226,
|
| 425 |
+
"loss": 0.5023544311523438,
|
| 426 |
+
"mean_token_accuracy": 0.8297024351177793,
|
| 427 |
+
"num_tokens": 4368268.0,
|
| 428 |
+
"step": 1650
|
| 429 |
+
},
|
| 430 |
+
{
|
| 431 |
+
"entropy": 0.4853435277938843,
|
| 432 |
+
"epoch": 4.166871165644172,
|
| 433 |
+
"grad_norm": 1.0383695363998413,
|
| 434 |
+
"learning_rate": 0.00029251001501843485,
|
| 435 |
+
"loss": 0.45016471862792967,
|
| 436 |
+
"mean_token_accuracy": 0.8418685424327851,
|
| 437 |
+
"num_tokens": 4494525.0,
|
| 438 |
+
"step": 1700
|
| 439 |
+
},
|
| 440 |
+
{
|
| 441 |
+
"entropy": 0.48153585255146025,
|
| 442 |
+
"epoch": 4.289570552147239,
|
| 443 |
+
"grad_norm": 0.7215332388877869,
|
| 444 |
+
"learning_rate": 0.00028471869125462477,
|
| 445 |
+
"loss": 0.4460752487182617,
|
| 446 |
+
"mean_token_accuracy": 0.8421206694841384,
|
| 447 |
+
"num_tokens": 4631985.0,
|
| 448 |
+
"step": 1750
|
| 449 |
+
},
|
| 450 |
+
{
|
| 451 |
+
"entropy": 0.49278041124343874,
|
| 452 |
+
"epoch": 4.412269938650307,
|
| 453 |
+
"grad_norm": 0.7771472334861755,
|
| 454 |
+
"learning_rate": 0.0002767756402149588,
|
| 455 |
+
"loss": 0.45794551849365234,
|
| 456 |
+
"mean_token_accuracy": 0.8404830145835877,
|
| 457 |
+
"num_tokens": 4766323.0,
|
| 458 |
+
"step": 1800
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"entropy": 0.5082326257228851,
|
| 462 |
+
"epoch": 4.534969325153375,
|
| 463 |
+
"grad_norm": 0.9560676217079163,
|
| 464 |
+
"learning_rate": 0.00026869539490814704,
|
| 465 |
+
"loss": 0.4632451629638672,
|
| 466 |
+
"mean_token_accuracy": 0.8386451864242553,
|
| 467 |
+
"num_tokens": 4895606.0,
|
| 468 |
+
"step": 1850
|
| 469 |
+
},
|
| 470 |
+
{
|
| 471 |
+
"entropy": 0.5081156292557716,
|
| 472 |
+
"epoch": 4.6576687116564415,
|
| 473 |
+
"grad_norm": 0.7719491720199585,
|
| 474 |
+
"learning_rate": 0.0002604927393604828,
|
| 475 |
+
"loss": 0.4630893325805664,
|
| 476 |
+
"mean_token_accuracy": 0.839306333065033,
|
| 477 |
+
"num_tokens": 5025912.0,
|
| 478 |
+
"step": 1900
|
| 479 |
+
},
|
| 480 |
+
{
|
| 481 |
+
"entropy": 0.5121548187732696,
|
| 482 |
+
"epoch": 4.780368098159509,
|
| 483 |
+
"grad_norm": 0.8171904683113098,
|
| 484 |
+
"learning_rate": 0.00025218268156623985,
|
| 485 |
+
"loss": 0.46541053771972657,
|
| 486 |
+
"mean_token_accuracy": 0.8369353520870209,
|
| 487 |
+
"num_tokens": 5162972.0,
|
| 488 |
+
"step": 1950
|
| 489 |
+
},
|
| 490 |
+
{
|
| 491 |
+
"entropy": 0.5085560208559037,
|
| 492 |
+
"epoch": 4.903067484662577,
|
| 493 |
+
"grad_norm": 0.5919133424758911,
|
| 494 |
+
"learning_rate": 0.00024378042602828492,
|
| 495 |
+
"loss": 0.46405506134033203,
|
| 496 |
+
"mean_token_accuracy": 0.8379521882534027,
|
| 497 |
+
"num_tokens": 5297935.0,
|
| 498 |
+
"step": 2000
|
| 499 |
+
},
|
| 500 |
+
{
|
| 501 |
+
"epoch": 5.0,
|
| 502 |
+
"eval_entropy": 0.5431888212476458,
|
| 503 |
+
"eval_mean_token_accuracy": 0.7749751152311053,
|
| 504 |
+
"eval_not_syn_loss": 0.7035888433456421,
|
| 505 |
+
"eval_not_syn_runtime": 56.5207,
|
| 506 |
+
"eval_not_syn_samples_per_second": 24.699,
|
| 507 |
+
"eval_not_syn_steps_per_second": 3.096,
|
| 508 |
+
"eval_num_tokens": 5400980.0,
|
| 509 |
+
"step": 2040
|
| 510 |
+
},
|
| 511 |
+
{
|
| 512 |
+
"epoch": 5.0,
|
| 513 |
+
"eval_entropy": 0.5098190077713558,
|
| 514 |
+
"eval_mean_token_accuracy": 0.832349077633449,
|
| 515 |
+
"eval_num_tokens": 5400980.0,
|
| 516 |
+
"eval_syn_loss": 0.6227898597717285,
|
| 517 |
+
"eval_syn_runtime": 57.6978,
|
| 518 |
+
"eval_syn_samples_per_second": 24.195,
|
| 519 |
+
"eval_syn_steps_per_second": 3.033,
|
| 520 |
+
"step": 2040
|
| 521 |
+
},
|
| 522 |
+
{
|
| 523 |
+
"entropy": 0.4960306563762703,
|
| 524 |
+
"epoch": 5.024539877300613,
|
| 525 |
+
"grad_norm": 0.9411171078681946,
|
| 526 |
+
"learning_rate": 0.0002353013459391488,
|
| 527 |
+
"loss": 0.4484004211425781,
|
| 528 |
+
"mean_token_accuracy": 0.8438825571175778,
|
| 529 |
+
"num_tokens": 5427362.0,
|
| 530 |
+
"step": 2050
|
| 531 |
+
},
|
| 532 |
+
{
|
| 533 |
+
"entropy": 0.3916636416316032,
|
| 534 |
+
"epoch": 5.147239263803681,
|
| 535 |
+
"grad_norm": 0.8929405808448792,
|
| 536 |
+
"learning_rate": 0.00022676095505345462,
|
| 537 |
+
"loss": 0.3446478271484375,
|
| 538 |
+
"mean_token_accuracy": 0.8732858788967133,
|
| 539 |
+
"num_tokens": 5564186.0,
|
| 540 |
+
"step": 2100
|
| 541 |
+
},
|
| 542 |
+
{
|
| 543 |
+
"entropy": 0.39636621803045274,
|
| 544 |
+
"epoch": 5.269938650306749,
|
| 545 |
+
"grad_norm": 0.69569331407547,
|
| 546 |
+
"learning_rate": 0.0002181748793031656,
|
| 547 |
+
"loss": 0.3548049163818359,
|
| 548 |
+
"mean_token_accuracy": 0.8699262255430221,
|
| 549 |
+
"num_tokens": 5694728.0,
|
| 550 |
+
"step": 2150
|
| 551 |
+
},
|
| 552 |
+
{
|
| 553 |
+
"entropy": 0.4028259950876236,
|
| 554 |
+
"epoch": 5.392638036809816,
|
| 555 |
+
"grad_norm": 0.7605823874473572,
|
| 556 |
+
"learning_rate": 0.00020955882820758868,
|
| 557 |
+
"loss": 0.35218334197998047,
|
| 558 |
+
"mean_token_accuracy": 0.8704028391838073,
|
| 559 |
+
"num_tokens": 5832581.0,
|
| 560 |
+
"step": 2200
|
| 561 |
+
},
|
| 562 |
+
{
|
| 563 |
+
"entropy": 0.4091260200738907,
|
| 564 |
+
"epoch": 5.515337423312883,
|
| 565 |
+
"grad_norm": 0.8990920782089233,
|
| 566 |
+
"learning_rate": 0.00020092856613044126,
|
| 567 |
+
"loss": 0.36372703552246094,
|
| 568 |
+
"mean_token_accuracy": 0.8663889598846436,
|
| 569 |
+
"num_tokens": 5962417.0,
|
| 570 |
+
"step": 2250
|
| 571 |
+
},
|
| 572 |
+
{
|
| 573 |
+
"entropy": 0.4057882487773895,
|
| 574 |
+
"epoch": 5.638036809815951,
|
| 575 |
+
"grad_norm": 0.9872731566429138,
|
| 576 |
+
"learning_rate": 0.0001922998834365732,
|
| 577 |
+
"loss": 0.3588364791870117,
|
| 578 |
+
"mean_token_accuracy": 0.8686526268720627,
|
| 579 |
+
"num_tokens": 6092338.0,
|
| 580 |
+
"step": 2300
|
| 581 |
+
},
|
| 582 |
+
{
|
| 583 |
+
"entropy": 0.41128933161497117,
|
| 584 |
+
"epoch": 5.7607361963190185,
|
| 585 |
+
"grad_norm": 0.8626382350921631,
|
| 586 |
+
"learning_rate": 0.0001836885676011143,
|
| 587 |
+
"loss": 0.36508998870849607,
|
| 588 |
+
"mean_token_accuracy": 0.8664334756135941,
|
| 589 |
+
"num_tokens": 6226327.0,
|
| 590 |
+
"step": 2350
|
| 591 |
+
},
|
| 592 |
+
{
|
| 593 |
+
"entropy": 0.40721403241157533,
|
| 594 |
+
"epoch": 5.883435582822086,
|
| 595 |
+
"grad_norm": 0.775476336479187,
|
| 596 |
+
"learning_rate": 0.00017511037432391027,
|
| 597 |
+
"loss": 0.36363380432128906,
|
| 598 |
+
"mean_token_accuracy": 0.8656364333629608,
|
| 599 |
+
"num_tokens": 6359083.0,
|
| 600 |
+
"step": 2400
|
| 601 |
+
},
|
| 602 |
+
{
|
| 603 |
+
"epoch": 6.0,
|
| 604 |
+
"eval_entropy": 0.4985327124595642,
|
| 605 |
+
"eval_mean_token_accuracy": 0.8019610796655927,
|
| 606 |
+
"eval_not_syn_loss": 0.717173159122467,
|
| 607 |
+
"eval_not_syn_runtime": 56.4169,
|
| 608 |
+
"eval_not_syn_samples_per_second": 24.744,
|
| 609 |
+
"eval_not_syn_steps_per_second": 3.102,
|
| 610 |
+
"eval_num_tokens": 6481176.0,
|
| 611 |
+
"step": 2448
|
| 612 |
+
},
|
| 613 |
+
{
|
| 614 |
+
"epoch": 6.0,
|
| 615 |
+
"eval_entropy": 0.4651233986445836,
|
| 616 |
+
"eval_mean_token_accuracy": 0.8059938202585493,
|
| 617 |
+
"eval_num_tokens": 6481176.0,
|
| 618 |
+
"eval_syn_loss": 0.6778863668441772,
|
| 619 |
+
"eval_syn_runtime": 58.5069,
|
| 620 |
+
"eval_syn_samples_per_second": 23.86,
|
| 621 |
+
"eval_syn_steps_per_second": 2.991,
|
| 622 |
+
"step": 2448
|
| 623 |
+
}
|
| 624 |
+
],
|
| 625 |
+
"logging_steps": 50,
|
| 626 |
+
"max_steps": 4080,
|
| 627 |
+
"num_input_tokens_seen": 0,
|
| 628 |
+
"num_train_epochs": 10,
|
| 629 |
+
"save_steps": 500,
|
| 630 |
+
"stateful_callbacks": {
|
| 631 |
+
"TrainerControl": {
|
| 632 |
+
"args": {
|
| 633 |
+
"should_epoch_stop": false,
|
| 634 |
+
"should_evaluate": false,
|
| 635 |
+
"should_log": false,
|
| 636 |
+
"should_save": true,
|
| 637 |
+
"should_training_stop": false
|
| 638 |
+
},
|
| 639 |
+
"attributes": {}
|
| 640 |
+
}
|
| 641 |
+
},
|
| 642 |
+
"total_flos": 1.321195971768434e+17,
|
| 643 |
+
"train_batch_size": 8,
|
| 644 |
+
"trial_name": null,
|
| 645 |
+
"trial_params": null
|
| 646 |
+
}
|
substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-2856/README.md
ADDED
|
@@ -0,0 +1,209 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
base_model: Qwen/Qwen3.5-2B-Base
|
| 3 |
+
library_name: peft
|
| 4 |
+
pipeline_tag: text-generation
|
| 5 |
+
tags:
|
| 6 |
+
- base_model:adapter:Qwen/Qwen3.5-2B-Base
|
| 7 |
+
- lora
|
| 8 |
+
- sft
|
| 9 |
+
- transformers
|
| 10 |
+
- trl
|
| 11 |
+
---
|
| 12 |
+
|
| 13 |
+
# Model Card for Model ID
|
| 14 |
+
|
| 15 |
+
<!-- Provide a quick summary of what the model is/does. -->
|
| 16 |
+
|
| 17 |
+
|
| 18 |
+
|
| 19 |
+
## Model Details
|
| 20 |
+
|
| 21 |
+
### Model Description
|
| 22 |
+
|
| 23 |
+
<!-- Provide a longer summary of what this model is. -->
|
| 24 |
+
|
| 25 |
+
|
| 26 |
+
|
| 27 |
+
- **Developed by:** [More Information Needed]
|
| 28 |
+
- **Funded by [optional]:** [More Information Needed]
|
| 29 |
+
- **Shared by [optional]:** [More Information Needed]
|
| 30 |
+
- **Model type:** [More Information Needed]
|
| 31 |
+
- **Language(s) (NLP):** [More Information Needed]
|
| 32 |
+
- **License:** [More Information Needed]
|
| 33 |
+
- **Finetuned from model [optional]:** [More Information Needed]
|
| 34 |
+
|
| 35 |
+
### Model Sources [optional]
|
| 36 |
+
|
| 37 |
+
<!-- Provide the basic links for the model. -->
|
| 38 |
+
|
| 39 |
+
- **Repository:** [More Information Needed]
|
| 40 |
+
- **Paper [optional]:** [More Information Needed]
|
| 41 |
+
- **Demo [optional]:** [More Information Needed]
|
| 42 |
+
|
| 43 |
+
## Uses
|
| 44 |
+
|
| 45 |
+
<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
|
| 46 |
+
|
| 47 |
+
### Direct Use
|
| 48 |
+
|
| 49 |
+
<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
|
| 50 |
+
|
| 51 |
+
[More Information Needed]
|
| 52 |
+
|
| 53 |
+
### Downstream Use [optional]
|
| 54 |
+
|
| 55 |
+
<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
|
| 56 |
+
|
| 57 |
+
[More Information Needed]
|
| 58 |
+
|
| 59 |
+
### Out-of-Scope Use
|
| 60 |
+
|
| 61 |
+
<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
|
| 62 |
+
|
| 63 |
+
[More Information Needed]
|
| 64 |
+
|
| 65 |
+
## Bias, Risks, and Limitations
|
| 66 |
+
|
| 67 |
+
<!-- This section is meant to convey both technical and sociotechnical limitations. -->
|
| 68 |
+
|
| 69 |
+
[More Information Needed]
|
| 70 |
+
|
| 71 |
+
### Recommendations
|
| 72 |
+
|
| 73 |
+
<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
|
| 74 |
+
|
| 75 |
+
Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
|
| 76 |
+
|
| 77 |
+
## How to Get Started with the Model
|
| 78 |
+
|
| 79 |
+
Use the code below to get started with the model.
|
| 80 |
+
|
| 81 |
+
[More Information Needed]
|
| 82 |
+
|
| 83 |
+
## Training Details
|
| 84 |
+
|
| 85 |
+
### Training Data
|
| 86 |
+
|
| 87 |
+
<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
|
| 88 |
+
|
| 89 |
+
[More Information Needed]
|
| 90 |
+
|
| 91 |
+
### Training Procedure
|
| 92 |
+
|
| 93 |
+
<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
|
| 94 |
+
|
| 95 |
+
#### Preprocessing [optional]
|
| 96 |
+
|
| 97 |
+
[More Information Needed]
|
| 98 |
+
|
| 99 |
+
|
| 100 |
+
#### Training Hyperparameters
|
| 101 |
+
|
| 102 |
+
- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
|
| 103 |
+
|
| 104 |
+
#### Speeds, Sizes, Times [optional]
|
| 105 |
+
|
| 106 |
+
<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
|
| 107 |
+
|
| 108 |
+
[More Information Needed]
|
| 109 |
+
|
| 110 |
+
## Evaluation
|
| 111 |
+
|
| 112 |
+
<!-- This section describes the evaluation protocols and provides the results. -->
|
| 113 |
+
|
| 114 |
+
### Testing Data, Factors & Metrics
|
| 115 |
+
|
| 116 |
+
#### Testing Data
|
| 117 |
+
|
| 118 |
+
<!-- This should link to a Dataset Card if possible. -->
|
| 119 |
+
|
| 120 |
+
[More Information Needed]
|
| 121 |
+
|
| 122 |
+
#### Factors
|
| 123 |
+
|
| 124 |
+
<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
|
| 125 |
+
|
| 126 |
+
[More Information Needed]
|
| 127 |
+
|
| 128 |
+
#### Metrics
|
| 129 |
+
|
| 130 |
+
<!-- These are the evaluation metrics being used, ideally with a description of why. -->
|
| 131 |
+
|
| 132 |
+
[More Information Needed]
|
| 133 |
+
|
| 134 |
+
### Results
|
| 135 |
+
|
| 136 |
+
[More Information Needed]
|
| 137 |
+
|
| 138 |
+
#### Summary
|
| 139 |
+
|
| 140 |
+
|
| 141 |
+
|
| 142 |
+
## Model Examination [optional]
|
| 143 |
+
|
| 144 |
+
<!-- Relevant interpretability work for the model goes here -->
|
| 145 |
+
|
| 146 |
+
[More Information Needed]
|
| 147 |
+
|
| 148 |
+
## Environmental Impact
|
| 149 |
+
|
| 150 |
+
<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
|
| 151 |
+
|
| 152 |
+
Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
|
| 153 |
+
|
| 154 |
+
- **Hardware Type:** [More Information Needed]
|
| 155 |
+
- **Hours used:** [More Information Needed]
|
| 156 |
+
- **Cloud Provider:** [More Information Needed]
|
| 157 |
+
- **Compute Region:** [More Information Needed]
|
| 158 |
+
- **Carbon Emitted:** [More Information Needed]
|
| 159 |
+
|
| 160 |
+
## Technical Specifications [optional]
|
| 161 |
+
|
| 162 |
+
### Model Architecture and Objective
|
| 163 |
+
|
| 164 |
+
[More Information Needed]
|
| 165 |
+
|
| 166 |
+
### Compute Infrastructure
|
| 167 |
+
|
| 168 |
+
[More Information Needed]
|
| 169 |
+
|
| 170 |
+
#### Hardware
|
| 171 |
+
|
| 172 |
+
[More Information Needed]
|
| 173 |
+
|
| 174 |
+
#### Software
|
| 175 |
+
|
| 176 |
+
[More Information Needed]
|
| 177 |
+
|
| 178 |
+
## Citation [optional]
|
| 179 |
+
|
| 180 |
+
<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
|
| 181 |
+
|
| 182 |
+
**BibTeX:**
|
| 183 |
+
|
| 184 |
+
[More Information Needed]
|
| 185 |
+
|
| 186 |
+
**APA:**
|
| 187 |
+
|
| 188 |
+
[More Information Needed]
|
| 189 |
+
|
| 190 |
+
## Glossary [optional]
|
| 191 |
+
|
| 192 |
+
<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
|
| 193 |
+
|
| 194 |
+
[More Information Needed]
|
| 195 |
+
|
| 196 |
+
## More Information [optional]
|
| 197 |
+
|
| 198 |
+
[More Information Needed]
|
| 199 |
+
|
| 200 |
+
## Model Card Authors [optional]
|
| 201 |
+
|
| 202 |
+
[More Information Needed]
|
| 203 |
+
|
| 204 |
+
## Model Card Contact
|
| 205 |
+
|
| 206 |
+
[More Information Needed]
|
| 207 |
+
### Framework versions
|
| 208 |
+
|
| 209 |
+
- PEFT 0.19.1
|
substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-2856/adapter_config.json
ADDED
|
@@ -0,0 +1,48 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"alora_invocation_tokens": null,
|
| 3 |
+
"alpha_pattern": {},
|
| 4 |
+
"arrow_config": null,
|
| 5 |
+
"auto_mapping": null,
|
| 6 |
+
"base_model_name_or_path": "Qwen/Qwen3.5-2B-Base",
|
| 7 |
+
"bias": "none",
|
| 8 |
+
"corda_config": null,
|
| 9 |
+
"ensure_weight_tying": false,
|
| 10 |
+
"eva_config": null,
|
| 11 |
+
"exclude_modules": null,
|
| 12 |
+
"fan_in_fan_out": false,
|
| 13 |
+
"inference_mode": true,
|
| 14 |
+
"init_lora_weights": true,
|
| 15 |
+
"layer_replication": null,
|
| 16 |
+
"layers_pattern": null,
|
| 17 |
+
"layers_to_transform": null,
|
| 18 |
+
"loftq_config": {},
|
| 19 |
+
"lora_alpha": 128,
|
| 20 |
+
"lora_bias": false,
|
| 21 |
+
"lora_dropout": 0.05635041589438875,
|
| 22 |
+
"lora_ga_config": null,
|
| 23 |
+
"megatron_config": null,
|
| 24 |
+
"megatron_core": "megatron.core",
|
| 25 |
+
"modules_to_save": null,
|
| 26 |
+
"peft_type": "LORA",
|
| 27 |
+
"peft_version": "0.19.1",
|
| 28 |
+
"qalora_group_size": 16,
|
| 29 |
+
"r": 64,
|
| 30 |
+
"rank_pattern": {},
|
| 31 |
+
"revision": null,
|
| 32 |
+
"target_modules": [
|
| 33 |
+
"q_proj",
|
| 34 |
+
"k_proj",
|
| 35 |
+
"gate_proj",
|
| 36 |
+
"up_proj",
|
| 37 |
+
"o_proj",
|
| 38 |
+
"down_proj",
|
| 39 |
+
"v_proj"
|
| 40 |
+
],
|
| 41 |
+
"target_parameters": null,
|
| 42 |
+
"task_type": "CAUSAL_LM",
|
| 43 |
+
"trainable_token_indices": null,
|
| 44 |
+
"use_bdlora": null,
|
| 45 |
+
"use_dora": false,
|
| 46 |
+
"use_qalora": false,
|
| 47 |
+
"use_rslora": false
|
| 48 |
+
}
|
substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-2856/chat_template.jinja
ADDED
|
@@ -0,0 +1,154 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{%- set image_count = namespace(value=0) %}
|
| 2 |
+
{%- set video_count = namespace(value=0) %}
|
| 3 |
+
{%- macro render_content(content, do_vision_count, is_system_content=false) %}
|
| 4 |
+
{%- if content is string %}
|
| 5 |
+
{{- content }}
|
| 6 |
+
{%- elif content is iterable and content is not mapping %}
|
| 7 |
+
{%- for item in content %}
|
| 8 |
+
{%- if 'image' in item or 'image_url' in item or item.type == 'image' %}
|
| 9 |
+
{%- if is_system_content %}
|
| 10 |
+
{{- raise_exception('System message cannot contain images.') }}
|
| 11 |
+
{%- endif %}
|
| 12 |
+
{%- if do_vision_count %}
|
| 13 |
+
{%- set image_count.value = image_count.value + 1 %}
|
| 14 |
+
{%- endif %}
|
| 15 |
+
{%- if add_vision_id %}
|
| 16 |
+
{{- 'Picture ' ~ image_count.value ~ ': ' }}
|
| 17 |
+
{%- endif %}
|
| 18 |
+
{{- '<|vision_start|><|image_pad|><|vision_end|>' }}
|
| 19 |
+
{%- elif 'video' in item or item.type == 'video' %}
|
| 20 |
+
{%- if is_system_content %}
|
| 21 |
+
{{- raise_exception('System message cannot contain videos.') }}
|
| 22 |
+
{%- endif %}
|
| 23 |
+
{%- if do_vision_count %}
|
| 24 |
+
{%- set video_count.value = video_count.value + 1 %}
|
| 25 |
+
{%- endif %}
|
| 26 |
+
{%- if add_vision_id %}
|
| 27 |
+
{{- 'Video ' ~ video_count.value ~ ': ' }}
|
| 28 |
+
{%- endif %}
|
| 29 |
+
{{- '<|vision_start|><|video_pad|><|vision_end|>' }}
|
| 30 |
+
{%- elif 'text' in item %}
|
| 31 |
+
{{- item.text }}
|
| 32 |
+
{%- else %}
|
| 33 |
+
{{- raise_exception('Unexpected item type in content.') }}
|
| 34 |
+
{%- endif %}
|
| 35 |
+
{%- endfor %}
|
| 36 |
+
{%- elif content is none or content is undefined %}
|
| 37 |
+
{{- '' }}
|
| 38 |
+
{%- else %}
|
| 39 |
+
{{- raise_exception('Unexpected content type.') }}
|
| 40 |
+
{%- endif %}
|
| 41 |
+
{%- endmacro %}
|
| 42 |
+
{%- if not messages %}
|
| 43 |
+
{{- raise_exception('No messages provided.') }}
|
| 44 |
+
{%- endif %}
|
| 45 |
+
{%- if tools and tools is iterable and tools is not mapping %}
|
| 46 |
+
{{- '<|im_start|>system\n' }}
|
| 47 |
+
{{- "# Tools\n\nYou have access to the following functions:\n\n<tools>" }}
|
| 48 |
+
{%- for tool in tools %}
|
| 49 |
+
{{- "\n" }}
|
| 50 |
+
{{- tool | tojson }}
|
| 51 |
+
{%- endfor %}
|
| 52 |
+
{{- "\n</tools>" }}
|
| 53 |
+
{{- '\n\nIf you choose to call a function ONLY reply in the following format with NO suffix:\n\n<tool_call>\n<function=example_function_name>\n<parameter=example_parameter_1>\nvalue_1\n</parameter>\n<parameter=example_parameter_2>\nThis is the value for the second parameter\nthat can span\nmultiple lines\n</parameter>\n</function>\n</tool_call>\n\n<IMPORTANT>\nReminder:\n- Function calls MUST follow the specified format: an inner <function=...></function> block must be nested within <tool_call></tool_call> XML tags\n- Required parameters MUST be specified\n- You may provide optional reasoning for your function call in natural language BEFORE the function call, but NOT after\n- If there is no function call available, answer the question like normal with your current knowledge and do not tell the user about function calls\n</IMPORTANT>' }}
|
| 54 |
+
{%- if messages[0].role == 'system' %}
|
| 55 |
+
{%- set content = render_content(messages[0].content, false, true)|trim %}
|
| 56 |
+
{%- if content %}
|
| 57 |
+
{{- '\n\n' + content }}
|
| 58 |
+
{%- endif %}
|
| 59 |
+
{%- endif %}
|
| 60 |
+
{{- '<|im_end|>\n' }}
|
| 61 |
+
{%- else %}
|
| 62 |
+
{%- if messages[0].role == 'system' %}
|
| 63 |
+
{%- set content = render_content(messages[0].content, false, true)|trim %}
|
| 64 |
+
{{- '<|im_start|>system\n' + content + '<|im_end|>\n' }}
|
| 65 |
+
{%- endif %}
|
| 66 |
+
{%- endif %}
|
| 67 |
+
{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
|
| 68 |
+
{%- for message in messages[::-1] %}
|
| 69 |
+
{%- set index = (messages|length - 1) - loop.index0 %}
|
| 70 |
+
{%- if ns.multi_step_tool and message.role == "user" %}
|
| 71 |
+
{%- set content = render_content(message.content, false)|trim %}
|
| 72 |
+
{%- if not(content.startswith('<tool_response>') and content.endswith('</tool_response>')) %}
|
| 73 |
+
{%- set ns.multi_step_tool = false %}
|
| 74 |
+
{%- set ns.last_query_index = index %}
|
| 75 |
+
{%- endif %}
|
| 76 |
+
{%- endif %}
|
| 77 |
+
{%- endfor %}
|
| 78 |
+
{%- if ns.multi_step_tool %}
|
| 79 |
+
{{- raise_exception('No user query found in messages.') }}
|
| 80 |
+
{%- endif %}
|
| 81 |
+
{%- for message in messages %}
|
| 82 |
+
{%- set content = render_content(message.content, true)|trim %}
|
| 83 |
+
{%- if message.role == "system" %}
|
| 84 |
+
{%- if not loop.first %}
|
| 85 |
+
{{- raise_exception('System message must be at the beginning.') }}
|
| 86 |
+
{%- endif %}
|
| 87 |
+
{%- elif message.role == "user" %}
|
| 88 |
+
{{- '<|im_start|>' + message.role + '\n' + content + '<|im_end|>' + '\n' }}
|
| 89 |
+
{%- elif message.role == "assistant" %}
|
| 90 |
+
{%- set reasoning_content = '' %}
|
| 91 |
+
{%- if message.reasoning_content is string %}
|
| 92 |
+
{%- set reasoning_content = message.reasoning_content %}
|
| 93 |
+
{%- else %}
|
| 94 |
+
{%- if '</think>' in content %}
|
| 95 |
+
{%- set reasoning_content = content.split('</think>')[0].rstrip('\n').split('<think>')[-1].lstrip('\n') %}
|
| 96 |
+
{%- set content = content.split('</think>')[-1].lstrip('\n') %}
|
| 97 |
+
{%- endif %}
|
| 98 |
+
{%- endif %}
|
| 99 |
+
{%- set reasoning_content = reasoning_content|trim %}
|
| 100 |
+
{%- if loop.index0 > ns.last_query_index %}
|
| 101 |
+
{{- '<|im_start|>' + message.role + '\n<think>\n' + reasoning_content + '\n</think>\n\n' + content }}
|
| 102 |
+
{%- else %}
|
| 103 |
+
{{- '<|im_start|>' + message.role + '\n' + content }}
|
| 104 |
+
{%- endif %}
|
| 105 |
+
{%- if message.tool_calls and message.tool_calls is iterable and message.tool_calls is not mapping %}
|
| 106 |
+
{%- for tool_call in message.tool_calls %}
|
| 107 |
+
{%- if tool_call.function is defined %}
|
| 108 |
+
{%- set tool_call = tool_call.function %}
|
| 109 |
+
{%- endif %}
|
| 110 |
+
{%- if loop.first %}
|
| 111 |
+
{%- if content|trim %}
|
| 112 |
+
{{- '\n\n<tool_call>\n<function=' + tool_call.name + '>\n' }}
|
| 113 |
+
{%- else %}
|
| 114 |
+
{{- '<tool_call>\n<function=' + tool_call.name + '>\n' }}
|
| 115 |
+
{%- endif %}
|
| 116 |
+
{%- else %}
|
| 117 |
+
{{- '\n<tool_call>\n<function=' + tool_call.name + '>\n' }}
|
| 118 |
+
{%- endif %}
|
| 119 |
+
{%- if tool_call.arguments is defined %}
|
| 120 |
+
{%- for args_name, args_value in tool_call.arguments|items %}
|
| 121 |
+
{{- '<parameter=' + args_name + '>\n' }}
|
| 122 |
+
{%- set args_value = args_value | tojson | safe if args_value is mapping or (args_value is sequence and args_value is not string) else args_value | string %}
|
| 123 |
+
{{- args_value }}
|
| 124 |
+
{{- '\n</parameter>\n' }}
|
| 125 |
+
{%- endfor %}
|
| 126 |
+
{%- endif %}
|
| 127 |
+
{{- '</function>\n</tool_call>' }}
|
| 128 |
+
{%- endfor %}
|
| 129 |
+
{%- endif %}
|
| 130 |
+
{{- '<|im_end|>\n' }}
|
| 131 |
+
{%- elif message.role == "tool" %}
|
| 132 |
+
{%- if loop.previtem and loop.previtem.role != "tool" %}
|
| 133 |
+
{{- '<|im_start|>user' }}
|
| 134 |
+
{%- endif %}
|
| 135 |
+
{{- '\n<tool_response>\n' }}
|
| 136 |
+
{{- content }}
|
| 137 |
+
{{- '\n</tool_response>' }}
|
| 138 |
+
{%- if not loop.last and loop.nextitem.role != "tool" %}
|
| 139 |
+
{{- '<|im_end|>\n' }}
|
| 140 |
+
{%- elif loop.last %}
|
| 141 |
+
{{- '<|im_end|>\n' }}
|
| 142 |
+
{%- endif %}
|
| 143 |
+
{%- else %}
|
| 144 |
+
{{- raise_exception('Unexpected message role.') }}
|
| 145 |
+
{%- endif %}
|
| 146 |
+
{%- endfor %}
|
| 147 |
+
{%- if add_generation_prompt %}
|
| 148 |
+
{{- '<|im_start|>assistant\n' }}
|
| 149 |
+
{%- if enable_thinking is defined and enable_thinking is true %}
|
| 150 |
+
{{- '<think>\n' }}
|
| 151 |
+
{%- else %}
|
| 152 |
+
{{- '<think>\n\n</think>\n\n' }}
|
| 153 |
+
{%- endif %}
|
| 154 |
+
{%- endif %}
|
substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-2856/tokenizer_config.json
ADDED
|
@@ -0,0 +1,31 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"add_prefix_space": false,
|
| 3 |
+
"audio_bos_token": "<|audio_start|>",
|
| 4 |
+
"audio_eos_token": "<|audio_end|>",
|
| 5 |
+
"audio_token": "<|audio_pad|>",
|
| 6 |
+
"backend": "tokenizers",
|
| 7 |
+
"bos_token": null,
|
| 8 |
+
"clean_up_tokenization_spaces": false,
|
| 9 |
+
"eos_token": "<|endoftext|>",
|
| 10 |
+
"errors": "replace",
|
| 11 |
+
"image_token": "<|image_pad|>",
|
| 12 |
+
"is_local": false,
|
| 13 |
+
"model_max_length": 262144,
|
| 14 |
+
"model_specific_special_tokens": {
|
| 15 |
+
"audio_bos_token": "<|audio_start|>",
|
| 16 |
+
"audio_eos_token": "<|audio_end|>",
|
| 17 |
+
"audio_token": "<|audio_pad|>",
|
| 18 |
+
"image_token": "<|image_pad|>",
|
| 19 |
+
"video_token": "<|video_pad|>",
|
| 20 |
+
"vision_bos_token": "<|vision_start|>",
|
| 21 |
+
"vision_eos_token": "<|vision_end|>"
|
| 22 |
+
},
|
| 23 |
+
"pad_token": "<|endoftext|>",
|
| 24 |
+
"pretokenize_regex": "(?i:'s|'t|'re|'ve|'m|'ll|'d)|[^\\r\\n\\p{L}\\p{N}]?[\\p{L}\\p{M}]+|\\p{N}| ?[^\\s\\p{L}\\p{M}\\p{N}]+[\\r\\n]*|\\s*[\\r\\n]+|\\s+(?!\\S)|\\s+",
|
| 25 |
+
"split_special_tokens": false,
|
| 26 |
+
"tokenizer_class": "TokenizersBackend",
|
| 27 |
+
"unk_token": null,
|
| 28 |
+
"video_token": "<|video_pad|>",
|
| 29 |
+
"vision_bos_token": "<|vision_start|>",
|
| 30 |
+
"vision_eos_token": "<|vision_end|>"
|
| 31 |
+
}
|
substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-2856/trainer_state.json
ADDED
|
@@ -0,0 +1,758 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": null,
|
| 3 |
+
"best_metric": null,
|
| 4 |
+
"best_model_checkpoint": null,
|
| 5 |
+
"epoch": 7.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 2856,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"entropy": 1.8153233402967452,
|
| 14 |
+
"epoch": 0.12269938650306748,
|
| 15 |
+
"grad_norm": 1.8510568141937256,
|
| 16 |
+
"learning_rate": 4.8469594667113335e-05,
|
| 17 |
+
"loss": 1.7217323303222656,
|
| 18 |
+
"mean_token_accuracy": 0.6236635231971741,
|
| 19 |
+
"num_tokens": 134329.0,
|
| 20 |
+
"step": 50
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"entropy": 0.97831438601017,
|
| 24 |
+
"epoch": 0.24539877300613497,
|
| 25 |
+
"grad_norm": 1.889868974685669,
|
| 26 |
+
"learning_rate": 9.792836473559631e-05,
|
| 27 |
+
"loss": 0.9050540924072266,
|
| 28 |
+
"mean_token_accuracy": 0.7377463465929032,
|
| 29 |
+
"num_tokens": 262407.0,
|
| 30 |
+
"step": 100
|
| 31 |
+
},
|
| 32 |
+
{
|
| 33 |
+
"entropy": 0.8743131357431412,
|
| 34 |
+
"epoch": 0.36809815950920244,
|
| 35 |
+
"grad_norm": 1.1766732931137085,
|
| 36 |
+
"learning_rate": 0.0001473871348040793,
|
| 37 |
+
"loss": 0.8129763031005859,
|
| 38 |
+
"mean_token_accuracy": 0.7583417356014251,
|
| 39 |
+
"num_tokens": 401136.0,
|
| 40 |
+
"step": 150
|
| 41 |
+
},
|
| 42 |
+
{
|
| 43 |
+
"entropy": 0.8273966419696808,
|
| 44 |
+
"epoch": 0.49079754601226994,
|
| 45 |
+
"grad_norm": 1.1311014890670776,
|
| 46 |
+
"learning_rate": 0.00019684590487256229,
|
| 47 |
+
"loss": 0.7711280822753906,
|
| 48 |
+
"mean_token_accuracy": 0.7668228060007095,
|
| 49 |
+
"num_tokens": 531647.0,
|
| 50 |
+
"step": 200
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"entropy": 0.8101438617706299,
|
| 54 |
+
"epoch": 0.6134969325153374,
|
| 55 |
+
"grad_norm": 0.8905950784683228,
|
| 56 |
+
"learning_rate": 0.0002463046749410453,
|
| 57 |
+
"loss": 0.7609222412109375,
|
| 58 |
+
"mean_token_accuracy": 0.7717174577713013,
|
| 59 |
+
"num_tokens": 666127.0,
|
| 60 |
+
"step": 250
|
| 61 |
+
},
|
| 62 |
+
{
|
| 63 |
+
"entropy": 0.7885055243968964,
|
| 64 |
+
"epoch": 0.7361963190184049,
|
| 65 |
+
"grad_norm": 1.1944245100021362,
|
| 66 |
+
"learning_rate": 0.00029576344500952824,
|
| 67 |
+
"loss": 0.7477743530273437,
|
| 68 |
+
"mean_token_accuracy": 0.773262197971344,
|
| 69 |
+
"num_tokens": 799292.0,
|
| 70 |
+
"step": 300
|
| 71 |
+
},
|
| 72 |
+
{
|
| 73 |
+
"entropy": 0.7858527797460556,
|
| 74 |
+
"epoch": 0.8588957055214724,
|
| 75 |
+
"grad_norm": 1.1311068534851074,
|
| 76 |
+
"learning_rate": 0.00034522221507801124,
|
| 77 |
+
"loss": 0.7411377716064453,
|
| 78 |
+
"mean_token_accuracy": 0.7745680212974548,
|
| 79 |
+
"num_tokens": 930614.0,
|
| 80 |
+
"step": 350
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"entropy": 0.7879797518253326,
|
| 84 |
+
"epoch": 0.9815950920245399,
|
| 85 |
+
"grad_norm": 0.9370436072349548,
|
| 86 |
+
"learning_rate": 0.00039468098514649423,
|
| 87 |
+
"loss": 0.7507785034179687,
|
| 88 |
+
"mean_token_accuracy": 0.7747444450855255,
|
| 89 |
+
"num_tokens": 1061654.0,
|
| 90 |
+
"step": 400
|
| 91 |
+
},
|
| 92 |
+
{
|
| 93 |
+
"epoch": 1.0,
|
| 94 |
+
"eval_entropy": 0.8527693561145238,
|
| 95 |
+
"eval_mean_token_accuracy": 0.7730878329277039,
|
| 96 |
+
"eval_not_syn_loss": 0.781921923160553,
|
| 97 |
+
"eval_not_syn_runtime": 56.9157,
|
| 98 |
+
"eval_not_syn_samples_per_second": 24.527,
|
| 99 |
+
"eval_not_syn_steps_per_second": 3.075,
|
| 100 |
+
"eval_num_tokens": 1080196.0,
|
| 101 |
+
"step": 408
|
| 102 |
+
},
|
| 103 |
+
{
|
| 104 |
+
"epoch": 1.0,
|
| 105 |
+
"eval_entropy": 0.7952005583899362,
|
| 106 |
+
"eval_mean_token_accuracy": 0.7705784467288426,
|
| 107 |
+
"eval_num_tokens": 1080196.0,
|
| 108 |
+
"eval_syn_loss": 0.7461121082305908,
|
| 109 |
+
"eval_syn_runtime": 59.5758,
|
| 110 |
+
"eval_syn_samples_per_second": 23.432,
|
| 111 |
+
"eval_syn_steps_per_second": 2.937,
|
| 112 |
+
"step": 408
|
| 113 |
+
},
|
| 114 |
+
{
|
| 115 |
+
"entropy": 0.7564875923021875,
|
| 116 |
+
"epoch": 1.1030674846625768,
|
| 117 |
+
"grad_norm": 1.1334599256515503,
|
| 118 |
+
"learning_rate": 0.00040345942966973637,
|
| 119 |
+
"loss": 0.7208955383300781,
|
| 120 |
+
"mean_token_accuracy": 0.7788144470465304,
|
| 121 |
+
"num_tokens": 1199170.0,
|
| 122 |
+
"step": 450
|
| 123 |
+
},
|
| 124 |
+
{
|
| 125 |
+
"entropy": 0.7427570706605912,
|
| 126 |
+
"epoch": 1.2257668711656442,
|
| 127 |
+
"grad_norm": 1.0777021646499634,
|
| 128 |
+
"learning_rate": 0.00040297229629224153,
|
| 129 |
+
"loss": 0.7049168395996094,
|
| 130 |
+
"mean_token_accuracy": 0.7830300116539002,
|
| 131 |
+
"num_tokens": 1334298.0,
|
| 132 |
+
"step": 500
|
| 133 |
+
},
|
| 134 |
+
{
|
| 135 |
+
"entropy": 0.7533353108167649,
|
| 136 |
+
"epoch": 1.3484662576687116,
|
| 137 |
+
"grad_norm": 1.3439561128616333,
|
| 138 |
+
"learning_rate": 0.00040211707285041045,
|
| 139 |
+
"loss": 0.7102288818359375,
|
| 140 |
+
"mean_token_accuracy": 0.7818449640274048,
|
| 141 |
+
"num_tokens": 1463189.0,
|
| 142 |
+
"step": 550
|
| 143 |
+
},
|
| 144 |
+
{
|
| 145 |
+
"entropy": 0.7666342407464981,
|
| 146 |
+
"epoch": 1.471165644171779,
|
| 147 |
+
"grad_norm": 1.1329729557037354,
|
| 148 |
+
"learning_rate": 0.00040089532410439227,
|
| 149 |
+
"loss": 0.7270996856689453,
|
| 150 |
+
"mean_token_accuracy": 0.778404277563095,
|
| 151 |
+
"num_tokens": 1590310.0,
|
| 152 |
+
"step": 600
|
| 153 |
+
},
|
| 154 |
+
{
|
| 155 |
+
"entropy": 0.7359934949874878,
|
| 156 |
+
"epoch": 1.5938650306748468,
|
| 157 |
+
"grad_norm": 1.2227331399917603,
|
| 158 |
+
"learning_rate": 0.0003993092854276068,
|
| 159 |
+
"loss": 0.6996555328369141,
|
| 160 |
+
"mean_token_accuracy": 0.7834958267211914,
|
| 161 |
+
"num_tokens": 1721836.0,
|
| 162 |
+
"step": 650
|
| 163 |
+
},
|
| 164 |
+
{
|
| 165 |
+
"entropy": 0.7278932738304138,
|
| 166 |
+
"epoch": 1.716564417177914,
|
| 167 |
+
"grad_norm": 0.8753179311752319,
|
| 168 |
+
"learning_rate": 0.0003973618587167924,
|
| 169 |
+
"loss": 0.694002456665039,
|
| 170 |
+
"mean_token_accuracy": 0.7849677371978759,
|
| 171 |
+
"num_tokens": 1853507.0,
|
| 172 |
+
"step": 700
|
| 173 |
+
},
|
| 174 |
+
{
|
| 175 |
+
"entropy": 0.7190143716335297,
|
| 176 |
+
"epoch": 1.8392638036809816,
|
| 177 |
+
"grad_norm": 0.8524742126464844,
|
| 178 |
+
"learning_rate": 0.0003950566070825483,
|
| 179 |
+
"loss": 0.6890143585205079,
|
| 180 |
+
"mean_token_accuracy": 0.7865350896120071,
|
| 181 |
+
"num_tokens": 1988113.0,
|
| 182 |
+
"step": 750
|
| 183 |
+
},
|
| 184 |
+
{
|
| 185 |
+
"entropy": 0.7072037667036056,
|
| 186 |
+
"epoch": 1.961963190184049,
|
| 187 |
+
"grad_norm": 0.877348005771637,
|
| 188 |
+
"learning_rate": 0.00039239774833008714,
|
| 189 |
+
"loss": 0.6795162200927735,
|
| 190 |
+
"mean_token_accuracy": 0.7885233855247498,
|
| 191 |
+
"num_tokens": 2121314.0,
|
| 192 |
+
"step": 800
|
| 193 |
+
},
|
| 194 |
+
{
|
| 195 |
+
"epoch": 2.0,
|
| 196 |
+
"eval_entropy": 0.818787784235818,
|
| 197 |
+
"eval_mean_token_accuracy": 0.7573730857031686,
|
| 198 |
+
"eval_not_syn_loss": 0.7486392855644226,
|
| 199 |
+
"eval_not_syn_runtime": 56.7426,
|
| 200 |
+
"eval_not_syn_samples_per_second": 24.602,
|
| 201 |
+
"eval_not_syn_steps_per_second": 3.084,
|
| 202 |
+
"eval_num_tokens": 2160392.0,
|
| 203 |
+
"step": 816
|
| 204 |
+
},
|
| 205 |
+
{
|
| 206 |
+
"epoch": 2.0,
|
| 207 |
+
"eval_entropy": 0.7632522354807173,
|
| 208 |
+
"eval_mean_token_accuracy": 0.8101163625717163,
|
| 209 |
+
"eval_num_tokens": 2160392.0,
|
| 210 |
+
"eval_syn_loss": 0.6829464435577393,
|
| 211 |
+
"eval_syn_runtime": 58.5255,
|
| 212 |
+
"eval_syn_samples_per_second": 23.853,
|
| 213 |
+
"eval_syn_steps_per_second": 2.99,
|
| 214 |
+
"step": 816
|
| 215 |
+
},
|
| 216 |
+
{
|
| 217 |
+
"entropy": 0.6619202801675508,
|
| 218 |
+
"epoch": 2.083435582822086,
|
| 219 |
+
"grad_norm": 0.8124546408653259,
|
| 220 |
+
"learning_rate": 0.00038939014724212456,
|
| 221 |
+
"loss": 0.6311377716064454,
|
| 222 |
+
"mean_token_accuracy": 0.7991988044796567,
|
| 223 |
+
"num_tokens": 2253179.0,
|
| 224 |
+
"step": 850
|
| 225 |
+
},
|
| 226 |
+
{
|
| 227 |
+
"entropy": 0.6668815296888352,
|
| 228 |
+
"epoch": 2.2061349693251535,
|
| 229 |
+
"grad_norm": 0.8693709969520569,
|
| 230 |
+
"learning_rate": 0.0003860393066780265,
|
| 231 |
+
"loss": 0.6279015731811524,
|
| 232 |
+
"mean_token_accuracy": 0.7983763587474823,
|
| 233 |
+
"num_tokens": 2379965.0,
|
| 234 |
+
"step": 900
|
| 235 |
+
},
|
| 236 |
+
{
|
| 237 |
+
"entropy": 0.6557431703805924,
|
| 238 |
+
"epoch": 2.3288343558282207,
|
| 239 |
+
"grad_norm": 0.8323768377304077,
|
| 240 |
+
"learning_rate": 0.0003823513575054986,
|
| 241 |
+
"loss": 0.6214478683471679,
|
| 242 |
+
"mean_token_accuracy": 0.8012632429599762,
|
| 243 |
+
"num_tokens": 2515593.0,
|
| 244 |
+
"step": 950
|
| 245 |
+
},
|
| 246 |
+
{
|
| 247 |
+
"entropy": 0.6633540654182434,
|
| 248 |
+
"epoch": 2.4515337423312884,
|
| 249 |
+
"grad_norm": 0.8328827023506165,
|
| 250 |
+
"learning_rate": 0.0003783330473832399,
|
| 251 |
+
"loss": 0.6310849761962891,
|
| 252 |
+
"mean_token_accuracy": 0.7995503985881806,
|
| 253 |
+
"num_tokens": 2642790.0,
|
| 254 |
+
"step": 1000
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"entropy": 0.6542247200012207,
|
| 258 |
+
"epoch": 2.574233128834356,
|
| 259 |
+
"grad_norm": 0.8204748630523682,
|
| 260 |
+
"learning_rate": 0.000373991728415085,
|
| 261 |
+
"loss": 0.6195787048339844,
|
| 262 |
+
"mean_token_accuracy": 0.8006457507610321,
|
| 263 |
+
"num_tokens": 2774396.0,
|
| 264 |
+
"step": 1050
|
| 265 |
+
},
|
| 266 |
+
{
|
| 267 |
+
"entropy": 0.6526653742790223,
|
| 268 |
+
"epoch": 2.6969325153374233,
|
| 269 |
+
"grad_norm": 1.065635323524475,
|
| 270 |
+
"learning_rate": 0.0003693353436982218,
|
| 271 |
+
"loss": 0.6206568145751953,
|
| 272 |
+
"mean_token_accuracy": 0.8002620726823807,
|
| 273 |
+
"num_tokens": 2912102.0,
|
| 274 |
+
"step": 1100
|
| 275 |
+
},
|
| 276 |
+
{
|
| 277 |
+
"entropy": 0.654406590461731,
|
| 278 |
+
"epoch": 2.819631901840491,
|
| 279 |
+
"grad_norm": 0.8653210401535034,
|
| 280 |
+
"learning_rate": 0.00036437241279009834,
|
| 281 |
+
"loss": 0.6259954452514649,
|
| 282 |
+
"mean_token_accuracy": 0.8008036535978317,
|
| 283 |
+
"num_tokens": 3042987.0,
|
| 284 |
+
"step": 1150
|
| 285 |
+
},
|
| 286 |
+
{
|
| 287 |
+
"entropy": 0.6310120761394501,
|
| 288 |
+
"epoch": 2.942331288343558,
|
| 289 |
+
"grad_norm": 0.7713245749473572,
|
| 290 |
+
"learning_rate": 0.0003591120161206093,
|
| 291 |
+
"loss": 0.6115898132324219,
|
| 292 |
+
"mean_token_accuracy": 0.8043569129705429,
|
| 293 |
+
"num_tokens": 3178044.0,
|
| 294 |
+
"step": 1200
|
| 295 |
+
},
|
| 296 |
+
{
|
| 297 |
+
"epoch": 3.0,
|
| 298 |
+
"eval_entropy": 0.6591688336644854,
|
| 299 |
+
"eval_mean_token_accuracy": 0.7992405581474304,
|
| 300 |
+
"eval_not_syn_loss": 0.6962469220161438,
|
| 301 |
+
"eval_not_syn_runtime": 56.161,
|
| 302 |
+
"eval_not_syn_samples_per_second": 24.857,
|
| 303 |
+
"eval_not_syn_steps_per_second": 3.116,
|
| 304 |
+
"eval_num_tokens": 3240588.0,
|
| 305 |
+
"step": 1224
|
| 306 |
+
},
|
| 307 |
+
{
|
| 308 |
+
"epoch": 3.0,
|
| 309 |
+
"eval_entropy": 0.6191406740461077,
|
| 310 |
+
"eval_mean_token_accuracy": 0.7810143651281084,
|
| 311 |
+
"eval_num_tokens": 3240588.0,
|
| 312 |
+
"eval_syn_loss": 0.6663697361946106,
|
| 313 |
+
"eval_syn_runtime": 58.4102,
|
| 314 |
+
"eval_syn_samples_per_second": 23.9,
|
| 315 |
+
"eval_syn_steps_per_second": 2.996,
|
| 316 |
+
"step": 1224
|
| 317 |
+
},
|
| 318 |
+
{
|
| 319 |
+
"entropy": 0.5853322166385073,
|
| 320 |
+
"epoch": 3.063803680981595,
|
| 321 |
+
"grad_norm": 0.7313649654388428,
|
| 322 |
+
"learning_rate": 0.00035356377837808084,
|
| 323 |
+
"loss": 0.5616408538818359,
|
| 324 |
+
"mean_token_accuracy": 0.8144418639366073,
|
| 325 |
+
"num_tokens": 3312797.0,
|
| 326 |
+
"step": 1250
|
| 327 |
+
},
|
| 328 |
+
{
|
| 329 |
+
"entropy": 0.5654504173994064,
|
| 330 |
+
"epoch": 3.1865030674846624,
|
| 331 |
+
"grad_norm": 0.713893711566925,
|
| 332 |
+
"learning_rate": 0.0003477378508994529,
|
| 333 |
+
"loss": 0.5305458068847656,
|
| 334 |
+
"mean_token_accuracy": 0.8218590825796127,
|
| 335 |
+
"num_tokens": 3441629.0,
|
| 336 |
+
"step": 1300
|
| 337 |
+
},
|
| 338 |
+
{
|
| 339 |
+
"entropy": 0.5721602493524551,
|
| 340 |
+
"epoch": 3.30920245398773,
|
| 341 |
+
"grad_norm": 0.8487221002578735,
|
| 342 |
+
"learning_rate": 0.00034164489309687927,
|
| 343 |
+
"loss": 0.5437272262573242,
|
| 344 |
+
"mean_token_accuracy": 0.8169907212257386,
|
| 345 |
+
"num_tokens": 3576093.0,
|
| 346 |
+
"step": 1350
|
| 347 |
+
},
|
| 348 |
+
{
|
| 349 |
+
"entropy": 0.5797226822376251,
|
| 350 |
+
"epoch": 3.4319018404907977,
|
| 351 |
+
"grad_norm": 0.9307994842529297,
|
| 352 |
+
"learning_rate": 0.0003352960529547267,
|
| 353 |
+
"loss": 0.5405152130126953,
|
| 354 |
+
"mean_token_accuracy": 0.818563597202301,
|
| 355 |
+
"num_tokens": 3705912.0,
|
| 356 |
+
"step": 1400
|
| 357 |
+
},
|
| 358 |
+
{
|
| 359 |
+
"entropy": 0.5822264245152473,
|
| 360 |
+
"epoch": 3.554601226993865,
|
| 361 |
+
"grad_norm": 0.71127849817276,
|
| 362 |
+
"learning_rate": 0.00032870294663265757,
|
| 363 |
+
"loss": 0.5559784698486329,
|
| 364 |
+
"mean_token_accuracy": 0.8159428876638413,
|
| 365 |
+
"num_tokens": 3832376.0,
|
| 366 |
+
"step": 1450
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"entropy": 0.5702577340602875,
|
| 370 |
+
"epoch": 3.6773006134969326,
|
| 371 |
+
"grad_norm": 0.8090758919715881,
|
| 372 |
+
"learning_rate": 0.0003218776372121157,
|
| 373 |
+
"loss": 0.5466982650756836,
|
| 374 |
+
"mean_token_accuracy": 0.8186432421207428,
|
| 375 |
+
"num_tokens": 3967273.0,
|
| 376 |
+
"step": 1500
|
| 377 |
+
},
|
| 378 |
+
{
|
| 379 |
+
"entropy": 0.5708746567368508,
|
| 380 |
+
"epoch": 3.8,
|
| 381 |
+
"grad_norm": 0.7777291536331177,
|
| 382 |
+
"learning_rate": 0.000314832612625101,
|
| 383 |
+
"loss": 0.5412663650512696,
|
| 384 |
+
"mean_token_accuracy": 0.8195344746112824,
|
| 385 |
+
"num_tokens": 4102434.0,
|
| 386 |
+
"step": 1550
|
| 387 |
+
},
|
| 388 |
+
{
|
| 389 |
+
"entropy": 0.578739655315876,
|
| 390 |
+
"epoch": 3.9226993865030675,
|
| 391 |
+
"grad_norm": 0.8103565573692322,
|
| 392 |
+
"learning_rate": 0.0003075807628056167,
|
| 393 |
+
"loss": 0.5605202865600586,
|
| 394 |
+
"mean_token_accuracy": 0.815134603381157,
|
| 395 |
+
"num_tokens": 4234037.0,
|
| 396 |
+
"step": 1600
|
| 397 |
+
},
|
| 398 |
+
{
|
| 399 |
+
"epoch": 4.0,
|
| 400 |
+
"eval_entropy": 0.6183825404303415,
|
| 401 |
+
"eval_mean_token_accuracy": 0.7838981601170131,
|
| 402 |
+
"eval_not_syn_loss": 0.6814553737640381,
|
| 403 |
+
"eval_not_syn_runtime": 56.4858,
|
| 404 |
+
"eval_not_syn_samples_per_second": 24.714,
|
| 405 |
+
"eval_not_syn_steps_per_second": 3.098,
|
| 406 |
+
"eval_num_tokens": 4320784.0,
|
| 407 |
+
"step": 1632
|
| 408 |
+
},
|
| 409 |
+
{
|
| 410 |
+
"epoch": 4.0,
|
| 411 |
+
"eval_entropy": 0.5816301565510886,
|
| 412 |
+
"eval_mean_token_accuracy": 0.8187636685371399,
|
| 413 |
+
"eval_num_tokens": 4320784.0,
|
| 414 |
+
"eval_syn_loss": 0.6366032361984253,
|
| 415 |
+
"eval_syn_runtime": 58.1317,
|
| 416 |
+
"eval_syn_samples_per_second": 24.014,
|
| 417 |
+
"eval_syn_steps_per_second": 3.01,
|
| 418 |
+
"step": 1632
|
| 419 |
+
},
|
| 420 |
+
{
|
| 421 |
+
"entropy": 0.5312421954039371,
|
| 422 |
+
"epoch": 4.044171779141104,
|
| 423 |
+
"grad_norm": 0.6530773043632507,
|
| 424 |
+
"learning_rate": 0.00030013535610559226,
|
| 425 |
+
"loss": 0.5023544311523438,
|
| 426 |
+
"mean_token_accuracy": 0.8297024351177793,
|
| 427 |
+
"num_tokens": 4368268.0,
|
| 428 |
+
"step": 1650
|
| 429 |
+
},
|
| 430 |
+
{
|
| 431 |
+
"entropy": 0.4853435277938843,
|
| 432 |
+
"epoch": 4.166871165644172,
|
| 433 |
+
"grad_norm": 1.0383695363998413,
|
| 434 |
+
"learning_rate": 0.00029251001501843485,
|
| 435 |
+
"loss": 0.45016471862792967,
|
| 436 |
+
"mean_token_accuracy": 0.8418685424327851,
|
| 437 |
+
"num_tokens": 4494525.0,
|
| 438 |
+
"step": 1700
|
| 439 |
+
},
|
| 440 |
+
{
|
| 441 |
+
"entropy": 0.48153585255146025,
|
| 442 |
+
"epoch": 4.289570552147239,
|
| 443 |
+
"grad_norm": 0.7215332388877869,
|
| 444 |
+
"learning_rate": 0.00028471869125462477,
|
| 445 |
+
"loss": 0.4460752487182617,
|
| 446 |
+
"mean_token_accuracy": 0.8421206694841384,
|
| 447 |
+
"num_tokens": 4631985.0,
|
| 448 |
+
"step": 1750
|
| 449 |
+
},
|
| 450 |
+
{
|
| 451 |
+
"entropy": 0.49278041124343874,
|
| 452 |
+
"epoch": 4.412269938650307,
|
| 453 |
+
"grad_norm": 0.7771472334861755,
|
| 454 |
+
"learning_rate": 0.0002767756402149588,
|
| 455 |
+
"loss": 0.45794551849365234,
|
| 456 |
+
"mean_token_accuracy": 0.8404830145835877,
|
| 457 |
+
"num_tokens": 4766323.0,
|
| 458 |
+
"step": 1800
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"entropy": 0.5082326257228851,
|
| 462 |
+
"epoch": 4.534969325153375,
|
| 463 |
+
"grad_norm": 0.9560676217079163,
|
| 464 |
+
"learning_rate": 0.00026869539490814704,
|
| 465 |
+
"loss": 0.4632451629638672,
|
| 466 |
+
"mean_token_accuracy": 0.8386451864242553,
|
| 467 |
+
"num_tokens": 4895606.0,
|
| 468 |
+
"step": 1850
|
| 469 |
+
},
|
| 470 |
+
{
|
| 471 |
+
"entropy": 0.5081156292557716,
|
| 472 |
+
"epoch": 4.6576687116564415,
|
| 473 |
+
"grad_norm": 0.7719491720199585,
|
| 474 |
+
"learning_rate": 0.0002604927393604828,
|
| 475 |
+
"loss": 0.4630893325805664,
|
| 476 |
+
"mean_token_accuracy": 0.839306333065033,
|
| 477 |
+
"num_tokens": 5025912.0,
|
| 478 |
+
"step": 1900
|
| 479 |
+
},
|
| 480 |
+
{
|
| 481 |
+
"entropy": 0.5121548187732696,
|
| 482 |
+
"epoch": 4.780368098159509,
|
| 483 |
+
"grad_norm": 0.8171904683113098,
|
| 484 |
+
"learning_rate": 0.00025218268156623985,
|
| 485 |
+
"loss": 0.46541053771972657,
|
| 486 |
+
"mean_token_accuracy": 0.8369353520870209,
|
| 487 |
+
"num_tokens": 5162972.0,
|
| 488 |
+
"step": 1950
|
| 489 |
+
},
|
| 490 |
+
{
|
| 491 |
+
"entropy": 0.5085560208559037,
|
| 492 |
+
"epoch": 4.903067484662577,
|
| 493 |
+
"grad_norm": 0.5919133424758911,
|
| 494 |
+
"learning_rate": 0.00024378042602828492,
|
| 495 |
+
"loss": 0.46405506134033203,
|
| 496 |
+
"mean_token_accuracy": 0.8379521882534027,
|
| 497 |
+
"num_tokens": 5297935.0,
|
| 498 |
+
"step": 2000
|
| 499 |
+
},
|
| 500 |
+
{
|
| 501 |
+
"epoch": 5.0,
|
| 502 |
+
"eval_entropy": 0.5431888212476458,
|
| 503 |
+
"eval_mean_token_accuracy": 0.7749751152311053,
|
| 504 |
+
"eval_not_syn_loss": 0.7035888433456421,
|
| 505 |
+
"eval_not_syn_runtime": 56.5207,
|
| 506 |
+
"eval_not_syn_samples_per_second": 24.699,
|
| 507 |
+
"eval_not_syn_steps_per_second": 3.096,
|
| 508 |
+
"eval_num_tokens": 5400980.0,
|
| 509 |
+
"step": 2040
|
| 510 |
+
},
|
| 511 |
+
{
|
| 512 |
+
"epoch": 5.0,
|
| 513 |
+
"eval_entropy": 0.5098190077713558,
|
| 514 |
+
"eval_mean_token_accuracy": 0.832349077633449,
|
| 515 |
+
"eval_num_tokens": 5400980.0,
|
| 516 |
+
"eval_syn_loss": 0.6227898597717285,
|
| 517 |
+
"eval_syn_runtime": 57.6978,
|
| 518 |
+
"eval_syn_samples_per_second": 24.195,
|
| 519 |
+
"eval_syn_steps_per_second": 3.033,
|
| 520 |
+
"step": 2040
|
| 521 |
+
},
|
| 522 |
+
{
|
| 523 |
+
"entropy": 0.4960306563762703,
|
| 524 |
+
"epoch": 5.024539877300613,
|
| 525 |
+
"grad_norm": 0.9411171078681946,
|
| 526 |
+
"learning_rate": 0.0002353013459391488,
|
| 527 |
+
"loss": 0.4484004211425781,
|
| 528 |
+
"mean_token_accuracy": 0.8438825571175778,
|
| 529 |
+
"num_tokens": 5427362.0,
|
| 530 |
+
"step": 2050
|
| 531 |
+
},
|
| 532 |
+
{
|
| 533 |
+
"entropy": 0.3916636416316032,
|
| 534 |
+
"epoch": 5.147239263803681,
|
| 535 |
+
"grad_norm": 0.8929405808448792,
|
| 536 |
+
"learning_rate": 0.00022676095505345462,
|
| 537 |
+
"loss": 0.3446478271484375,
|
| 538 |
+
"mean_token_accuracy": 0.8732858788967133,
|
| 539 |
+
"num_tokens": 5564186.0,
|
| 540 |
+
"step": 2100
|
| 541 |
+
},
|
| 542 |
+
{
|
| 543 |
+
"entropy": 0.39636621803045274,
|
| 544 |
+
"epoch": 5.269938650306749,
|
| 545 |
+
"grad_norm": 0.69569331407547,
|
| 546 |
+
"learning_rate": 0.0002181748793031656,
|
| 547 |
+
"loss": 0.3548049163818359,
|
| 548 |
+
"mean_token_accuracy": 0.8699262255430221,
|
| 549 |
+
"num_tokens": 5694728.0,
|
| 550 |
+
"step": 2150
|
| 551 |
+
},
|
| 552 |
+
{
|
| 553 |
+
"entropy": 0.4028259950876236,
|
| 554 |
+
"epoch": 5.392638036809816,
|
| 555 |
+
"grad_norm": 0.7605823874473572,
|
| 556 |
+
"learning_rate": 0.00020955882820758868,
|
| 557 |
+
"loss": 0.35218334197998047,
|
| 558 |
+
"mean_token_accuracy": 0.8704028391838073,
|
| 559 |
+
"num_tokens": 5832581.0,
|
| 560 |
+
"step": 2200
|
| 561 |
+
},
|
| 562 |
+
{
|
| 563 |
+
"entropy": 0.4091260200738907,
|
| 564 |
+
"epoch": 5.515337423312883,
|
| 565 |
+
"grad_norm": 0.8990920782089233,
|
| 566 |
+
"learning_rate": 0.00020092856613044126,
|
| 567 |
+
"loss": 0.36372703552246094,
|
| 568 |
+
"mean_token_accuracy": 0.8663889598846436,
|
| 569 |
+
"num_tokens": 5962417.0,
|
| 570 |
+
"step": 2250
|
| 571 |
+
},
|
| 572 |
+
{
|
| 573 |
+
"entropy": 0.4057882487773895,
|
| 574 |
+
"epoch": 5.638036809815951,
|
| 575 |
+
"grad_norm": 0.9872731566429138,
|
| 576 |
+
"learning_rate": 0.0001922998834365732,
|
| 577 |
+
"loss": 0.3588364791870117,
|
| 578 |
+
"mean_token_accuracy": 0.8686526268720627,
|
| 579 |
+
"num_tokens": 6092338.0,
|
| 580 |
+
"step": 2300
|
| 581 |
+
},
|
| 582 |
+
{
|
| 583 |
+
"entropy": 0.41128933161497117,
|
| 584 |
+
"epoch": 5.7607361963190185,
|
| 585 |
+
"grad_norm": 0.8626382350921631,
|
| 586 |
+
"learning_rate": 0.0001836885676011143,
|
| 587 |
+
"loss": 0.36508998870849607,
|
| 588 |
+
"mean_token_accuracy": 0.8664334756135941,
|
| 589 |
+
"num_tokens": 6226327.0,
|
| 590 |
+
"step": 2350
|
| 591 |
+
},
|
| 592 |
+
{
|
| 593 |
+
"entropy": 0.40721403241157533,
|
| 594 |
+
"epoch": 5.883435582822086,
|
| 595 |
+
"grad_norm": 0.775476336479187,
|
| 596 |
+
"learning_rate": 0.00017511037432391027,
|
| 597 |
+
"loss": 0.36363380432128906,
|
| 598 |
+
"mean_token_accuracy": 0.8656364333629608,
|
| 599 |
+
"num_tokens": 6359083.0,
|
| 600 |
+
"step": 2400
|
| 601 |
+
},
|
| 602 |
+
{
|
| 603 |
+
"epoch": 6.0,
|
| 604 |
+
"eval_entropy": 0.4985327124595642,
|
| 605 |
+
"eval_mean_token_accuracy": 0.8019610796655927,
|
| 606 |
+
"eval_not_syn_loss": 0.717173159122467,
|
| 607 |
+
"eval_not_syn_runtime": 56.4169,
|
| 608 |
+
"eval_not_syn_samples_per_second": 24.744,
|
| 609 |
+
"eval_not_syn_steps_per_second": 3.102,
|
| 610 |
+
"eval_num_tokens": 6481176.0,
|
| 611 |
+
"step": 2448
|
| 612 |
+
},
|
| 613 |
+
{
|
| 614 |
+
"epoch": 6.0,
|
| 615 |
+
"eval_entropy": 0.4651233986445836,
|
| 616 |
+
"eval_mean_token_accuracy": 0.8059938202585493,
|
| 617 |
+
"eval_num_tokens": 6481176.0,
|
| 618 |
+
"eval_syn_loss": 0.6778863668441772,
|
| 619 |
+
"eval_syn_runtime": 58.5069,
|
| 620 |
+
"eval_syn_samples_per_second": 23.86,
|
| 621 |
+
"eval_syn_steps_per_second": 2.991,
|
| 622 |
+
"step": 2448
|
| 623 |
+
},
|
| 624 |
+
{
|
| 625 |
+
"entropy": 0.40582828509687174,
|
| 626 |
+
"epoch": 6.004907975460123,
|
| 627 |
+
"grad_norm": 0.6788471341133118,
|
| 628 |
+
"learning_rate": 0.0001665809987020958,
|
| 629 |
+
"loss": 0.3566559600830078,
|
| 630 |
+
"mean_token_accuracy": 0.868294438930473,
|
| 631 |
+
"num_tokens": 6486549.0,
|
| 632 |
+
"step": 2450
|
| 633 |
+
},
|
| 634 |
+
{
|
| 635 |
+
"entropy": 0.2933592677116394,
|
| 636 |
+
"epoch": 6.12760736196319,
|
| 637 |
+
"grad_norm": 0.6598446369171143,
|
| 638 |
+
"learning_rate": 0.00015811604651354912,
|
| 639 |
+
"loss": 0.24192693710327148,
|
| 640 |
+
"mean_token_accuracy": 0.9082661873102188,
|
| 641 |
+
"num_tokens": 6622384.0,
|
| 642 |
+
"step": 2500
|
| 643 |
+
},
|
| 644 |
+
{
|
| 645 |
+
"entropy": 0.2975004023313522,
|
| 646 |
+
"epoch": 6.250306748466258,
|
| 647 |
+
"grad_norm": 0.7413120865821838,
|
| 648 |
+
"learning_rate": 0.00014973100566377046,
|
| 649 |
+
"loss": 0.2502699661254883,
|
| 650 |
+
"mean_token_accuracy": 0.9056042343378067,
|
| 651 |
+
"num_tokens": 6758637.0,
|
| 652 |
+
"step": 2550
|
| 653 |
+
},
|
| 654 |
+
{
|
| 655 |
+
"entropy": 0.3052474121749401,
|
| 656 |
+
"epoch": 6.373006134969325,
|
| 657 |
+
"grad_norm": 0.8629281520843506,
|
| 658 |
+
"learning_rate": 0.00014144121784842467,
|
| 659 |
+
"loss": 0.2543815040588379,
|
| 660 |
+
"mean_token_accuracy": 0.9046169209480286,
|
| 661 |
+
"num_tokens": 6888533.0,
|
| 662 |
+
"step": 2600
|
| 663 |
+
},
|
| 664 |
+
{
|
| 665 |
+
"entropy": 0.303673956990242,
|
| 666 |
+
"epoch": 6.495705521472392,
|
| 667 |
+
"grad_norm": 0.8680911660194397,
|
| 668 |
+
"learning_rate": 0.000133261850483398,
|
| 669 |
+
"loss": 0.2529561424255371,
|
| 670 |
+
"mean_token_accuracy": 0.9040863698720932,
|
| 671 |
+
"num_tokens": 7023729.0,
|
| 672 |
+
"step": 2650
|
| 673 |
+
},
|
| 674 |
+
{
|
| 675 |
+
"entropy": 0.2993183335661888,
|
| 676 |
+
"epoch": 6.61840490797546,
|
| 677 |
+
"grad_norm": 0.7648728489875793,
|
| 678 |
+
"learning_rate": 0.00012520786895372493,
|
| 679 |
+
"loss": 0.2517524528503418,
|
| 680 |
+
"mean_token_accuracy": 0.9038773095607757,
|
| 681 |
+
"num_tokens": 7153821.0,
|
| 682 |
+
"step": 2700
|
| 683 |
+
},
|
| 684 |
+
{
|
| 685 |
+
"entropy": 0.3006903246045113,
|
| 686 |
+
"epoch": 6.741104294478528,
|
| 687 |
+
"grad_norm": 0.7433948516845703,
|
| 688 |
+
"learning_rate": 0.00011729400923216141,
|
| 689 |
+
"loss": 0.25459712982177735,
|
| 690 |
+
"mean_token_accuracy": 0.9037652736902237,
|
| 691 |
+
"num_tokens": 7283513.0,
|
| 692 |
+
"step": 2750
|
| 693 |
+
},
|
| 694 |
+
{
|
| 695 |
+
"entropy": 0.30416361182928087,
|
| 696 |
+
"epoch": 6.863803680981595,
|
| 697 |
+
"grad_norm": 0.8744818568229675,
|
| 698 |
+
"learning_rate": 0.00010953475091750244,
|
| 699 |
+
"loss": 0.25621614456176756,
|
| 700 |
+
"mean_token_accuracy": 0.9019701254367828,
|
| 701 |
+
"num_tokens": 7411578.0,
|
| 702 |
+
"step": 2800
|
| 703 |
+
},
|
| 704 |
+
{
|
| 705 |
+
"entropy": 0.2956543755531311,
|
| 706 |
+
"epoch": 6.986503067484662,
|
| 707 |
+
"grad_norm": 0.8242411017417908,
|
| 708 |
+
"learning_rate": 0.00010194429074197415,
|
| 709 |
+
"loss": 0.24829059600830078,
|
| 710 |
+
"mean_token_accuracy": 0.9054388856887817,
|
| 711 |
+
"num_tokens": 7548356.0,
|
| 712 |
+
"step": 2850
|
| 713 |
+
},
|
| 714 |
+
{
|
| 715 |
+
"epoch": 7.0,
|
| 716 |
+
"eval_entropy": 0.4070556444781167,
|
| 717 |
+
"eval_mean_token_accuracy": 0.7848251972879682,
|
| 718 |
+
"eval_not_syn_loss": 0.8054981827735901,
|
| 719 |
+
"eval_not_syn_runtime": 56.6822,
|
| 720 |
+
"eval_not_syn_samples_per_second": 24.629,
|
| 721 |
+
"eval_not_syn_steps_per_second": 3.087,
|
| 722 |
+
"eval_num_tokens": 7561372.0,
|
| 723 |
+
"step": 2856
|
| 724 |
+
},
|
| 725 |
+
{
|
| 726 |
+
"epoch": 7.0,
|
| 727 |
+
"eval_entropy": 0.3784746325016022,
|
| 728 |
+
"eval_mean_token_accuracy": 0.8244705755370004,
|
| 729 |
+
"eval_num_tokens": 7561372.0,
|
| 730 |
+
"eval_syn_loss": 0.7377892732620239,
|
| 731 |
+
"eval_syn_runtime": 58.6534,
|
| 732 |
+
"eval_syn_samples_per_second": 23.801,
|
| 733 |
+
"eval_syn_steps_per_second": 2.984,
|
| 734 |
+
"step": 2856
|
| 735 |
+
}
|
| 736 |
+
],
|
| 737 |
+
"logging_steps": 50,
|
| 738 |
+
"max_steps": 4080,
|
| 739 |
+
"num_input_tokens_seen": 0,
|
| 740 |
+
"num_train_epochs": 10,
|
| 741 |
+
"save_steps": 500,
|
| 742 |
+
"stateful_callbacks": {
|
| 743 |
+
"TrainerControl": {
|
| 744 |
+
"args": {
|
| 745 |
+
"should_epoch_stop": false,
|
| 746 |
+
"should_evaluate": false,
|
| 747 |
+
"should_log": false,
|
| 748 |
+
"should_save": true,
|
| 749 |
+
"should_training_stop": false
|
| 750 |
+
},
|
| 751 |
+
"attributes": {}
|
| 752 |
+
}
|
| 753 |
+
},
|
| 754 |
+
"total_flos": 1.5400107074712845e+17,
|
| 755 |
+
"train_batch_size": 8,
|
| 756 |
+
"trial_name": null,
|
| 757 |
+
"trial_params": null
|
| 758 |
+
}
|
substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-3264/README.md
ADDED
|
@@ -0,0 +1,209 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
base_model: Qwen/Qwen3.5-2B-Base
|
| 3 |
+
library_name: peft
|
| 4 |
+
pipeline_tag: text-generation
|
| 5 |
+
tags:
|
| 6 |
+
- base_model:adapter:Qwen/Qwen3.5-2B-Base
|
| 7 |
+
- lora
|
| 8 |
+
- sft
|
| 9 |
+
- transformers
|
| 10 |
+
- trl
|
| 11 |
+
---
|
| 12 |
+
|
| 13 |
+
# Model Card for Model ID
|
| 14 |
+
|
| 15 |
+
<!-- Provide a quick summary of what the model is/does. -->
|
| 16 |
+
|
| 17 |
+
|
| 18 |
+
|
| 19 |
+
## Model Details
|
| 20 |
+
|
| 21 |
+
### Model Description
|
| 22 |
+
|
| 23 |
+
<!-- Provide a longer summary of what this model is. -->
|
| 24 |
+
|
| 25 |
+
|
| 26 |
+
|
| 27 |
+
- **Developed by:** [More Information Needed]
|
| 28 |
+
- **Funded by [optional]:** [More Information Needed]
|
| 29 |
+
- **Shared by [optional]:** [More Information Needed]
|
| 30 |
+
- **Model type:** [More Information Needed]
|
| 31 |
+
- **Language(s) (NLP):** [More Information Needed]
|
| 32 |
+
- **License:** [More Information Needed]
|
| 33 |
+
- **Finetuned from model [optional]:** [More Information Needed]
|
| 34 |
+
|
| 35 |
+
### Model Sources [optional]
|
| 36 |
+
|
| 37 |
+
<!-- Provide the basic links for the model. -->
|
| 38 |
+
|
| 39 |
+
- **Repository:** [More Information Needed]
|
| 40 |
+
- **Paper [optional]:** [More Information Needed]
|
| 41 |
+
- **Demo [optional]:** [More Information Needed]
|
| 42 |
+
|
| 43 |
+
## Uses
|
| 44 |
+
|
| 45 |
+
<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
|
| 46 |
+
|
| 47 |
+
### Direct Use
|
| 48 |
+
|
| 49 |
+
<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
|
| 50 |
+
|
| 51 |
+
[More Information Needed]
|
| 52 |
+
|
| 53 |
+
### Downstream Use [optional]
|
| 54 |
+
|
| 55 |
+
<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
|
| 56 |
+
|
| 57 |
+
[More Information Needed]
|
| 58 |
+
|
| 59 |
+
### Out-of-Scope Use
|
| 60 |
+
|
| 61 |
+
<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
|
| 62 |
+
|
| 63 |
+
[More Information Needed]
|
| 64 |
+
|
| 65 |
+
## Bias, Risks, and Limitations
|
| 66 |
+
|
| 67 |
+
<!-- This section is meant to convey both technical and sociotechnical limitations. -->
|
| 68 |
+
|
| 69 |
+
[More Information Needed]
|
| 70 |
+
|
| 71 |
+
### Recommendations
|
| 72 |
+
|
| 73 |
+
<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
|
| 74 |
+
|
| 75 |
+
Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
|
| 76 |
+
|
| 77 |
+
## How to Get Started with the Model
|
| 78 |
+
|
| 79 |
+
Use the code below to get started with the model.
|
| 80 |
+
|
| 81 |
+
[More Information Needed]
|
| 82 |
+
|
| 83 |
+
## Training Details
|
| 84 |
+
|
| 85 |
+
### Training Data
|
| 86 |
+
|
| 87 |
+
<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
|
| 88 |
+
|
| 89 |
+
[More Information Needed]
|
| 90 |
+
|
| 91 |
+
### Training Procedure
|
| 92 |
+
|
| 93 |
+
<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
|
| 94 |
+
|
| 95 |
+
#### Preprocessing [optional]
|
| 96 |
+
|
| 97 |
+
[More Information Needed]
|
| 98 |
+
|
| 99 |
+
|
| 100 |
+
#### Training Hyperparameters
|
| 101 |
+
|
| 102 |
+
- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
|
| 103 |
+
|
| 104 |
+
#### Speeds, Sizes, Times [optional]
|
| 105 |
+
|
| 106 |
+
<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
|
| 107 |
+
|
| 108 |
+
[More Information Needed]
|
| 109 |
+
|
| 110 |
+
## Evaluation
|
| 111 |
+
|
| 112 |
+
<!-- This section describes the evaluation protocols and provides the results. -->
|
| 113 |
+
|
| 114 |
+
### Testing Data, Factors & Metrics
|
| 115 |
+
|
| 116 |
+
#### Testing Data
|
| 117 |
+
|
| 118 |
+
<!-- This should link to a Dataset Card if possible. -->
|
| 119 |
+
|
| 120 |
+
[More Information Needed]
|
| 121 |
+
|
| 122 |
+
#### Factors
|
| 123 |
+
|
| 124 |
+
<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
|
| 125 |
+
|
| 126 |
+
[More Information Needed]
|
| 127 |
+
|
| 128 |
+
#### Metrics
|
| 129 |
+
|
| 130 |
+
<!-- These are the evaluation metrics being used, ideally with a description of why. -->
|
| 131 |
+
|
| 132 |
+
[More Information Needed]
|
| 133 |
+
|
| 134 |
+
### Results
|
| 135 |
+
|
| 136 |
+
[More Information Needed]
|
| 137 |
+
|
| 138 |
+
#### Summary
|
| 139 |
+
|
| 140 |
+
|
| 141 |
+
|
| 142 |
+
## Model Examination [optional]
|
| 143 |
+
|
| 144 |
+
<!-- Relevant interpretability work for the model goes here -->
|
| 145 |
+
|
| 146 |
+
[More Information Needed]
|
| 147 |
+
|
| 148 |
+
## Environmental Impact
|
| 149 |
+
|
| 150 |
+
<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
|
| 151 |
+
|
| 152 |
+
Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
|
| 153 |
+
|
| 154 |
+
- **Hardware Type:** [More Information Needed]
|
| 155 |
+
- **Hours used:** [More Information Needed]
|
| 156 |
+
- **Cloud Provider:** [More Information Needed]
|
| 157 |
+
- **Compute Region:** [More Information Needed]
|
| 158 |
+
- **Carbon Emitted:** [More Information Needed]
|
| 159 |
+
|
| 160 |
+
## Technical Specifications [optional]
|
| 161 |
+
|
| 162 |
+
### Model Architecture and Objective
|
| 163 |
+
|
| 164 |
+
[More Information Needed]
|
| 165 |
+
|
| 166 |
+
### Compute Infrastructure
|
| 167 |
+
|
| 168 |
+
[More Information Needed]
|
| 169 |
+
|
| 170 |
+
#### Hardware
|
| 171 |
+
|
| 172 |
+
[More Information Needed]
|
| 173 |
+
|
| 174 |
+
#### Software
|
| 175 |
+
|
| 176 |
+
[More Information Needed]
|
| 177 |
+
|
| 178 |
+
## Citation [optional]
|
| 179 |
+
|
| 180 |
+
<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
|
| 181 |
+
|
| 182 |
+
**BibTeX:**
|
| 183 |
+
|
| 184 |
+
[More Information Needed]
|
| 185 |
+
|
| 186 |
+
**APA:**
|
| 187 |
+
|
| 188 |
+
[More Information Needed]
|
| 189 |
+
|
| 190 |
+
## Glossary [optional]
|
| 191 |
+
|
| 192 |
+
<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
|
| 193 |
+
|
| 194 |
+
[More Information Needed]
|
| 195 |
+
|
| 196 |
+
## More Information [optional]
|
| 197 |
+
|
| 198 |
+
[More Information Needed]
|
| 199 |
+
|
| 200 |
+
## Model Card Authors [optional]
|
| 201 |
+
|
| 202 |
+
[More Information Needed]
|
| 203 |
+
|
| 204 |
+
## Model Card Contact
|
| 205 |
+
|
| 206 |
+
[More Information Needed]
|
| 207 |
+
### Framework versions
|
| 208 |
+
|
| 209 |
+
- PEFT 0.19.1
|
substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-3264/adapter_config.json
ADDED
|
@@ -0,0 +1,48 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"alora_invocation_tokens": null,
|
| 3 |
+
"alpha_pattern": {},
|
| 4 |
+
"arrow_config": null,
|
| 5 |
+
"auto_mapping": null,
|
| 6 |
+
"base_model_name_or_path": "Qwen/Qwen3.5-2B-Base",
|
| 7 |
+
"bias": "none",
|
| 8 |
+
"corda_config": null,
|
| 9 |
+
"ensure_weight_tying": false,
|
| 10 |
+
"eva_config": null,
|
| 11 |
+
"exclude_modules": null,
|
| 12 |
+
"fan_in_fan_out": false,
|
| 13 |
+
"inference_mode": true,
|
| 14 |
+
"init_lora_weights": true,
|
| 15 |
+
"layer_replication": null,
|
| 16 |
+
"layers_pattern": null,
|
| 17 |
+
"layers_to_transform": null,
|
| 18 |
+
"loftq_config": {},
|
| 19 |
+
"lora_alpha": 128,
|
| 20 |
+
"lora_bias": false,
|
| 21 |
+
"lora_dropout": 0.05635041589438875,
|
| 22 |
+
"lora_ga_config": null,
|
| 23 |
+
"megatron_config": null,
|
| 24 |
+
"megatron_core": "megatron.core",
|
| 25 |
+
"modules_to_save": null,
|
| 26 |
+
"peft_type": "LORA",
|
| 27 |
+
"peft_version": "0.19.1",
|
| 28 |
+
"qalora_group_size": 16,
|
| 29 |
+
"r": 64,
|
| 30 |
+
"rank_pattern": {},
|
| 31 |
+
"revision": null,
|
| 32 |
+
"target_modules": [
|
| 33 |
+
"q_proj",
|
| 34 |
+
"k_proj",
|
| 35 |
+
"gate_proj",
|
| 36 |
+
"up_proj",
|
| 37 |
+
"o_proj",
|
| 38 |
+
"down_proj",
|
| 39 |
+
"v_proj"
|
| 40 |
+
],
|
| 41 |
+
"target_parameters": null,
|
| 42 |
+
"task_type": "CAUSAL_LM",
|
| 43 |
+
"trainable_token_indices": null,
|
| 44 |
+
"use_bdlora": null,
|
| 45 |
+
"use_dora": false,
|
| 46 |
+
"use_qalora": false,
|
| 47 |
+
"use_rslora": false
|
| 48 |
+
}
|
substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-3264/chat_template.jinja
ADDED
|
@@ -0,0 +1,154 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{%- set image_count = namespace(value=0) %}
|
| 2 |
+
{%- set video_count = namespace(value=0) %}
|
| 3 |
+
{%- macro render_content(content, do_vision_count, is_system_content=false) %}
|
| 4 |
+
{%- if content is string %}
|
| 5 |
+
{{- content }}
|
| 6 |
+
{%- elif content is iterable and content is not mapping %}
|
| 7 |
+
{%- for item in content %}
|
| 8 |
+
{%- if 'image' in item or 'image_url' in item or item.type == 'image' %}
|
| 9 |
+
{%- if is_system_content %}
|
| 10 |
+
{{- raise_exception('System message cannot contain images.') }}
|
| 11 |
+
{%- endif %}
|
| 12 |
+
{%- if do_vision_count %}
|
| 13 |
+
{%- set image_count.value = image_count.value + 1 %}
|
| 14 |
+
{%- endif %}
|
| 15 |
+
{%- if add_vision_id %}
|
| 16 |
+
{{- 'Picture ' ~ image_count.value ~ ': ' }}
|
| 17 |
+
{%- endif %}
|
| 18 |
+
{{- '<|vision_start|><|image_pad|><|vision_end|>' }}
|
| 19 |
+
{%- elif 'video' in item or item.type == 'video' %}
|
| 20 |
+
{%- if is_system_content %}
|
| 21 |
+
{{- raise_exception('System message cannot contain videos.') }}
|
| 22 |
+
{%- endif %}
|
| 23 |
+
{%- if do_vision_count %}
|
| 24 |
+
{%- set video_count.value = video_count.value + 1 %}
|
| 25 |
+
{%- endif %}
|
| 26 |
+
{%- if add_vision_id %}
|
| 27 |
+
{{- 'Video ' ~ video_count.value ~ ': ' }}
|
| 28 |
+
{%- endif %}
|
| 29 |
+
{{- '<|vision_start|><|video_pad|><|vision_end|>' }}
|
| 30 |
+
{%- elif 'text' in item %}
|
| 31 |
+
{{- item.text }}
|
| 32 |
+
{%- else %}
|
| 33 |
+
{{- raise_exception('Unexpected item type in content.') }}
|
| 34 |
+
{%- endif %}
|
| 35 |
+
{%- endfor %}
|
| 36 |
+
{%- elif content is none or content is undefined %}
|
| 37 |
+
{{- '' }}
|
| 38 |
+
{%- else %}
|
| 39 |
+
{{- raise_exception('Unexpected content type.') }}
|
| 40 |
+
{%- endif %}
|
| 41 |
+
{%- endmacro %}
|
| 42 |
+
{%- if not messages %}
|
| 43 |
+
{{- raise_exception('No messages provided.') }}
|
| 44 |
+
{%- endif %}
|
| 45 |
+
{%- if tools and tools is iterable and tools is not mapping %}
|
| 46 |
+
{{- '<|im_start|>system\n' }}
|
| 47 |
+
{{- "# Tools\n\nYou have access to the following functions:\n\n<tools>" }}
|
| 48 |
+
{%- for tool in tools %}
|
| 49 |
+
{{- "\n" }}
|
| 50 |
+
{{- tool | tojson }}
|
| 51 |
+
{%- endfor %}
|
| 52 |
+
{{- "\n</tools>" }}
|
| 53 |
+
{{- '\n\nIf you choose to call a function ONLY reply in the following format with NO suffix:\n\n<tool_call>\n<function=example_function_name>\n<parameter=example_parameter_1>\nvalue_1\n</parameter>\n<parameter=example_parameter_2>\nThis is the value for the second parameter\nthat can span\nmultiple lines\n</parameter>\n</function>\n</tool_call>\n\n<IMPORTANT>\nReminder:\n- Function calls MUST follow the specified format: an inner <function=...></function> block must be nested within <tool_call></tool_call> XML tags\n- Required parameters MUST be specified\n- You may provide optional reasoning for your function call in natural language BEFORE the function call, but NOT after\n- If there is no function call available, answer the question like normal with your current knowledge and do not tell the user about function calls\n</IMPORTANT>' }}
|
| 54 |
+
{%- if messages[0].role == 'system' %}
|
| 55 |
+
{%- set content = render_content(messages[0].content, false, true)|trim %}
|
| 56 |
+
{%- if content %}
|
| 57 |
+
{{- '\n\n' + content }}
|
| 58 |
+
{%- endif %}
|
| 59 |
+
{%- endif %}
|
| 60 |
+
{{- '<|im_end|>\n' }}
|
| 61 |
+
{%- else %}
|
| 62 |
+
{%- if messages[0].role == 'system' %}
|
| 63 |
+
{%- set content = render_content(messages[0].content, false, true)|trim %}
|
| 64 |
+
{{- '<|im_start|>system\n' + content + '<|im_end|>\n' }}
|
| 65 |
+
{%- endif %}
|
| 66 |
+
{%- endif %}
|
| 67 |
+
{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
|
| 68 |
+
{%- for message in messages[::-1] %}
|
| 69 |
+
{%- set index = (messages|length - 1) - loop.index0 %}
|
| 70 |
+
{%- if ns.multi_step_tool and message.role == "user" %}
|
| 71 |
+
{%- set content = render_content(message.content, false)|trim %}
|
| 72 |
+
{%- if not(content.startswith('<tool_response>') and content.endswith('</tool_response>')) %}
|
| 73 |
+
{%- set ns.multi_step_tool = false %}
|
| 74 |
+
{%- set ns.last_query_index = index %}
|
| 75 |
+
{%- endif %}
|
| 76 |
+
{%- endif %}
|
| 77 |
+
{%- endfor %}
|
| 78 |
+
{%- if ns.multi_step_tool %}
|
| 79 |
+
{{- raise_exception('No user query found in messages.') }}
|
| 80 |
+
{%- endif %}
|
| 81 |
+
{%- for message in messages %}
|
| 82 |
+
{%- set content = render_content(message.content, true)|trim %}
|
| 83 |
+
{%- if message.role == "system" %}
|
| 84 |
+
{%- if not loop.first %}
|
| 85 |
+
{{- raise_exception('System message must be at the beginning.') }}
|
| 86 |
+
{%- endif %}
|
| 87 |
+
{%- elif message.role == "user" %}
|
| 88 |
+
{{- '<|im_start|>' + message.role + '\n' + content + '<|im_end|>' + '\n' }}
|
| 89 |
+
{%- elif message.role == "assistant" %}
|
| 90 |
+
{%- set reasoning_content = '' %}
|
| 91 |
+
{%- if message.reasoning_content is string %}
|
| 92 |
+
{%- set reasoning_content = message.reasoning_content %}
|
| 93 |
+
{%- else %}
|
| 94 |
+
{%- if '</think>' in content %}
|
| 95 |
+
{%- set reasoning_content = content.split('</think>')[0].rstrip('\n').split('<think>')[-1].lstrip('\n') %}
|
| 96 |
+
{%- set content = content.split('</think>')[-1].lstrip('\n') %}
|
| 97 |
+
{%- endif %}
|
| 98 |
+
{%- endif %}
|
| 99 |
+
{%- set reasoning_content = reasoning_content|trim %}
|
| 100 |
+
{%- if loop.index0 > ns.last_query_index %}
|
| 101 |
+
{{- '<|im_start|>' + message.role + '\n<think>\n' + reasoning_content + '\n</think>\n\n' + content }}
|
| 102 |
+
{%- else %}
|
| 103 |
+
{{- '<|im_start|>' + message.role + '\n' + content }}
|
| 104 |
+
{%- endif %}
|
| 105 |
+
{%- if message.tool_calls and message.tool_calls is iterable and message.tool_calls is not mapping %}
|
| 106 |
+
{%- for tool_call in message.tool_calls %}
|
| 107 |
+
{%- if tool_call.function is defined %}
|
| 108 |
+
{%- set tool_call = tool_call.function %}
|
| 109 |
+
{%- endif %}
|
| 110 |
+
{%- if loop.first %}
|
| 111 |
+
{%- if content|trim %}
|
| 112 |
+
{{- '\n\n<tool_call>\n<function=' + tool_call.name + '>\n' }}
|
| 113 |
+
{%- else %}
|
| 114 |
+
{{- '<tool_call>\n<function=' + tool_call.name + '>\n' }}
|
| 115 |
+
{%- endif %}
|
| 116 |
+
{%- else %}
|
| 117 |
+
{{- '\n<tool_call>\n<function=' + tool_call.name + '>\n' }}
|
| 118 |
+
{%- endif %}
|
| 119 |
+
{%- if tool_call.arguments is defined %}
|
| 120 |
+
{%- for args_name, args_value in tool_call.arguments|items %}
|
| 121 |
+
{{- '<parameter=' + args_name + '>\n' }}
|
| 122 |
+
{%- set args_value = args_value | tojson | safe if args_value is mapping or (args_value is sequence and args_value is not string) else args_value | string %}
|
| 123 |
+
{{- args_value }}
|
| 124 |
+
{{- '\n</parameter>\n' }}
|
| 125 |
+
{%- endfor %}
|
| 126 |
+
{%- endif %}
|
| 127 |
+
{{- '</function>\n</tool_call>' }}
|
| 128 |
+
{%- endfor %}
|
| 129 |
+
{%- endif %}
|
| 130 |
+
{{- '<|im_end|>\n' }}
|
| 131 |
+
{%- elif message.role == "tool" %}
|
| 132 |
+
{%- if loop.previtem and loop.previtem.role != "tool" %}
|
| 133 |
+
{{- '<|im_start|>user' }}
|
| 134 |
+
{%- endif %}
|
| 135 |
+
{{- '\n<tool_response>\n' }}
|
| 136 |
+
{{- content }}
|
| 137 |
+
{{- '\n</tool_response>' }}
|
| 138 |
+
{%- if not loop.last and loop.nextitem.role != "tool" %}
|
| 139 |
+
{{- '<|im_end|>\n' }}
|
| 140 |
+
{%- elif loop.last %}
|
| 141 |
+
{{- '<|im_end|>\n' }}
|
| 142 |
+
{%- endif %}
|
| 143 |
+
{%- else %}
|
| 144 |
+
{{- raise_exception('Unexpected message role.') }}
|
| 145 |
+
{%- endif %}
|
| 146 |
+
{%- endfor %}
|
| 147 |
+
{%- if add_generation_prompt %}
|
| 148 |
+
{{- '<|im_start|>assistant\n' }}
|
| 149 |
+
{%- if enable_thinking is defined and enable_thinking is true %}
|
| 150 |
+
{{- '<think>\n' }}
|
| 151 |
+
{%- else %}
|
| 152 |
+
{{- '<think>\n\n</think>\n\n' }}
|
| 153 |
+
{%- endif %}
|
| 154 |
+
{%- endif %}
|