20241212

config.json

@@ -1,5 +1,5 @@
 {
-  "_name_or_path": "assets/facebookai_xlm_roberta_base_pretrain_20240912_e1",
+  "_name_or_path": "assets/facebookai_xlm_roberta_base_pretrain_20241212",
   "architectures": [
     "XLMRobertaForTokenClassification"
   ],
@@ -47,7 +47,7 @@
   "pad_token_id": 1,
   "position_embedding_type": "absolute",
   "torch_dtype": "float32",
-  "transformers_version": "4.44.1",
+  "transformers_version": "4.47.0",
   "type_vocab_size": 1,
   "use_cache": true,
   "vocab_size": 250002

metrics.json

@@ -1,62 +1,62 @@
 {
-    "eval_loss": 0.05851283296942711,
-    "eval_f1": 0.7611489329237343,
-    "eval_recall": 0.7718861209964413,
-    "eval_accuracy": 0.9808362068965517,
-    "eval_precision": 0.752981756409835,
+    "eval_loss": 0.062092795968055725,
+    "eval_f1": 0.7617369641587196,
+    "eval_recall": 0.7967042497831743,
+    "eval_accuracy": 0.9821096951844263,
+    "eval_precision": 0.7301653204843213,
     "eval_classification_report": {
         "EVT": {
-            "precision": 0.7,
-            "recall": 0.7,
-            "f1-score": 0.7,
-            "support": 440
+            "precision": 0.7176220806794055,
+            "recall": 0.7502774694783574,
+            "f1-score": 0.7335865436787845,
+            "support": 901
         },
         "LOC": {
-            "precision": 0.7535301668806161,
-            "recall": 0.8198324022346368,
-            "f1-score": 0.7852842809364547,
-            "support": 716
+            "precision": 0.7591863517060368,
+            "recall": 0.8615040953090097,
+            "f1-score": 0.8071154516916638,
+            "support": 1343
         },
         "ORG": {
-            "precision": 0.8244274809160306,
-            "recall": 0.7397260273972602,
-            "f1-score": 0.7797833935018051,
-            "support": 438
+            "precision": 0.7531831537708129,
+            "recall": 0.77834008097166,
+            "f1-score": 0.7655550024888004,
+            "support": 988
         },
         "PER": {
-            "precision": 0.8765603328710125,
-            "recall": 0.8789986091794159,
-            "f1-score": 0.8777777777777778,
-            "support": 719
+            "precision": 0.8291905672402804,
+            "recall": 0.8690714762859052,
+            "f1-score": 0.8486627527723418,
+            "support": 1497
         },
         "PRD": {
-            "precision": 0.7058823529411765,
-            "recall": 0.6072874493927125,
-            "f1-score": 0.6528835690968444,
-            "support": 494
+            "precision": 0.6442762535477767,
+            "recall": 0.6663405088062623,
+            "f1-score": 0.6551226551226551,
+            "support": 1022
         },
         "micro avg": {
-            "precision": 0.7799129804205946,
-            "recall": 0.766298539365871,
-            "f1-score": 0.7730458221024259,
-            "support": 2807
+            "precision": 0.7498773106494356,
+            "recall": 0.7970787689097548,
+            "f1-score": 0.772757923128793,
+            "support": 5751
         },
         "macro avg": {
-            "precision": 0.7720800667217672,
-            "recall": 0.7491688976408051,
-            "f1-score": 0.7591458042625764,
-            "support": 2807
+            "precision": 0.7406916813888624,
+            "recall": 0.785106726170239,
+            "f1-score": 0.7620084811508491,
+            "support": 5751
         },
         "weighted avg": {
-            "precision": 0.7793301025347138,
-            "recall": 0.766298539365871,
-            "f1-score": 0.7714482995583739,
-            "support": 2807
+            "precision": 0.7494450236903917,
+            "recall": 0.7970787689097548,
+            "f1-score": 0.7722603659139234,
+            "support": 5751
         }
     },
-    "eval_runtime": 1.6151,
-    "eval_samples_per_second": 619.156,
-    "eval_steps_per_second": 4.953,
-    "epoch": 6.005221932114883,
-    "train_loss": 0.0214
+    "eval_runtime": 3.2082,
+    "eval_samples_per_second": 638.365,
+    "eval_steps_per_second": 4.987,
+    "epoch": 18.28821811100292,
+    "train_loss": 0.018
 }

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:562685288efeac89942e310fa97a21abda0c383002b4903b342f9009b21d7b72
+oid sha256:394c9b3c7b10aa527f3bb2646e04cfe2798d9f3abafcbd098d749d8cba7d82b7
 size 1109870108

tokenizer_config.json

@@ -42,10 +42,11 @@
     }
   },
   "bos_token": "<s>",
-  "clean_up_tokenization_spaces": true,
+  "clean_up_tokenization_spaces": false,
   "cls_token": "<s>",
   "do_lower_case": false,
   "eos_token": "</s>",
+  "extra_special_tokens": {},
   "mask_token": "<mask>",
   "model_max_length": 512,
   "pad_token": "<pad>",

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:bf328ac5dce2605d36ef58939ad3b50c0ffd2028e4606c33a8c7b0a763e9b4db
-size 5112
+oid sha256:e342d99ea1971421359dcecd60501450ffb679d7d40b8535d11bf6cc4071a34d
+size 5240

training_args.json

@@ -6,15 +6,15 @@
     "do_predict": false,
     "eval_strategy": "steps",
     "prediction_loss_only": false,
-    "per_device_train_batch_size": 32,
+    "per_device_train_batch_size": 48,
     "per_device_eval_batch_size": 128,
     "per_gpu_train_batch_size": null,
     "per_gpu_eval_batch_size": null,
-    "gradient_accumulation_steps": 1,
+    "gradient_accumulation_steps": 2,
     "eval_accumulation_steps": null,
     "eval_delay": 0,
     "torch_empty_cache_steps": null,
-    "learning_rate": 2e-05,
+    "learning_rate": 1e-05,
     "weight_decay": 0.01,
     "adam_beta1": 0.9,
     "adam_beta2": 0.999,
@@ -93,13 +93,12 @@
     },
     "deepspeed": null,
     "label_smoothing_factor": 0.0,
-    "optim": "adamw_torch",
+    "optim": "ademamix_8bit",
     "optim_args": null,
     "adafactor": false,
     "group_by_length": false,
     "length_column_name": "length",
     "report_to": [
-        "tensorboard",
         "wandb"
     ],
     "ddp_find_unused_parameters": null,
@@ -114,11 +113,12 @@
     "hub_model_id": null,
     "hub_strategy": "every_save",
     "hub_token": null,
-    "hub_private_repo": false,
+    "hub_private_repo": null,
     "hub_always_push": false,
     "gradient_checkpointing": false,
     "gradient_checkpointing_kwargs": null,
     "include_inputs_for_metrics": false,
+    "include_for_metrics": [],
     "eval_do_concat_batches": true,
     "fp16_backend": "auto",
     "evaluation_strategy": null,
@@ -143,5 +143,7 @@
     "optim_target_modules": null,
     "batch_eval_metrics": false,
     "eval_on_start": false,
-    "eval_use_gather_object": false
+    "use_liger_kernel": false,
+    "eval_use_gather_object": false,
+    "average_tokens_across_devices": false
 }