ShengdingHu commited on
Commit
754cbd0
·
1 Parent(s): da4524f

Training in progress, epoch 1

Browse files
pytorch_model.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:c67eb4abd0d251308ad555227026025bfb77f69415f7261c90b3ec50e631c286
3
  size 1084131
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1f54e6d71ebebd51f486085171b81aa0fb763e753bf0389a4bf4b88a84db1edd
3
  size 1084131
runs/Jan31_19-19-11_node1/events.out.tfevents.1643628019.node1 CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:e29fb7f517dcd92b769c62e13607b54f3f2e7b790884f98090194e4e045581d9
3
- size 7125
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8ed6da3ab720f3d8fc0ba8d7e196f4ad858818970be1b2856ceff4dfa31dc2dd
3
+ size 9286
runs/Jan31_19-25-22_node1/1643628441.83444/events.out.tfevents.1643628441.node1 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f15637b9505c5534f88affd168b16fc59d025a5cb935b513b5be546fe728a336
3
+ size 5008
runs/Jan31_19-25-22_node1/events.out.tfevents.1643628441.node1 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6124074f80c03d988cef88cf51d0a8c19488e571920c53823e5027798e558c82
3
+ size 4305
tokenizer.json CHANGED
The diff for this file is too large to render. See raw diff
 
training_args.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:d5671f63c7b06e7b055be2e554d1b04117f631e20ceccceb2c6bdc0d98e02557
3
  size 3183
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c2a6c9356208e65c852641943db784f8160e696fb6cd5e449c34a8681ea7871e
3
  size 3183
training_config.json CHANGED
@@ -1 +1 @@
1
- {"dataset_config_name": ["en"], "delta_type": "bitfit", "do_eval": true, "do_test": true, "do_train": true, "eval_dataset_config_name": ["en"], "eval_dataset_name": "cola", "evaluation_strategy": "epoch", "greater_is_better": true, "learning_rate": 0.003, "load_best_model_at_end": true, "max_source_length": 128, "metric_for_best_model": "average_metrics", "model_name_or_path": "../../../../plm_cache/t5-base", "num_train_epochs": 20, "output_dir": "outputs/bitfit/t5-base/cola", "overwrite_output_dir": true, "per_device_eval_batch_size": 100, "per_device_train_batch_size": 100, "predict_with_generate": true, "push_to_hub": true, "save_strategy": "epoch", "save_total_limit": 1, "seed": 42, "split_validation_test": true, "task_name": "cola", "test_dataset_config_name": ["en"], "test_dataset_name": "cola", "tokenizer_name": "../../../../plm_cache/t5-base", "warmup_steps": 0}
 
1
+ {"dataset_config_name": ["en"], "delta_type": "bitfit", "do_eval": true, "do_test": true, "do_train": true, "eval_dataset_config_name": ["en"], "eval_dataset_name": "cola", "evaluation_strategy": "epoch", "greater_is_better": true, "learning_rate": 0.0003, "load_best_model_at_end": true, "max_source_length": 128, "metric_for_best_model": "average_metrics", "model_name_or_path": "../../../../plm_cache/t5-base", "num_train_epochs": 20, "output_dir": "outputs/bitfit/t5-base/cola", "overwrite_output_dir": true, "per_device_eval_batch_size": 100, "per_device_train_batch_size": 100, "predict_with_generate": true, "push_to_hub": true, "save_strategy": "epoch", "save_total_limit": 1, "seed": 42, "split_validation_test": true, "task_name": "cola", "test_dataset_config_name": ["en"], "test_dataset_name": "cola", "tokenizer_name": "../../../../plm_cache/t5-base", "warmup_steps": 500}