Upload 3 files

config.json

@@ -0,0 +1,113 @@
+{
+  "_name_or_path": "",
+  "activation_dropout": 0.1,
+  "activation_function": "gelu",
+  "add_bias_logits": false,
+  "add_cross_attention": false,
+  "add_final_layer_norm": false,
+  "architectures": [
+    "BartForConditionalGeneration"
+  ],
+  "attention_dropout": 0.1,
+  "bad_words_ids": null,
+  "bos_token_id": 0,
+  "chunk_size_feed_forward": 0,
+  "classif_dropout": 0.1,
+  "classifier_dropout": 0.0,
+  "clip_val": 2.5,
+  "cross_attention_hidden_size": null,
+  "d_model": 1024,
+  "decoder_attention_heads": 16,
+  "decoder_ffn_dim": 4096,
+  "decoder_layerdrop": 0.0,
+  "decoder_layers": 1,
+  "decoder_start_token_id": 2,
+  "diversity_penalty": 0.0,
+  "do_sample": false,
+  "dropout": 0.1,
+  "early_stopping": true,
+  "encoder_attention_heads": 16,
+  "encoder_ffn_dim": 4096,
+  "encoder_layerdrop": 0.0,
+  "encoder_layers": 3,
+  "encoder_no_repeat_ngram_size": 0,
+  "eos_token_id": 2,
+  "finetuning_task": null,
+  "forced_bos_token_id": 0,
+  "forced_eos_token_id": 2,
+  "gradient_checkpointing": false,
+  "id2label": {
+    "0": "LABEL_0",
+    "1": "LABEL_1",
+    "2": "LABEL_2"
+  },
+  "init_std": 0.02,
+  "input_bits": 8,
+  "is_decoder": false,
+  "is_encoder_decoder": true,
+  "label2id": {
+    "LABEL_0": 0,
+    "LABEL_1": 1,
+    "LABEL_2": 2
+  },
+  "length_penalty": 1.0,
+  "max_length": 20,
+  "max_position_embeddings": 1024,
+  "min_length": 0,
+  "model_type": "bart",
+  "no_repeat_ngram_size": 3,
+  "normalize_before": false,
+  "num_beam_groups": 1,
+  "num_beams": 4,
+  "num_hidden_layers": 12,
+  "num_return_sequences": 1,
+  "output_attentions": false,
+  "output_hidden_states": false,
+  "output_scores": false,
+  "pad_token_id": 1,
+  "prefix": null,
+  "problem_type": null,
+  "pruned_heads": {},
+  "quantize_act": true,
+  "remove_invalid_values": false,
+  "repetition_penalty": 1.0,
+  "return_dict": true,
+  "return_dict_in_generate": false,
+  "scale_embedding": false,
+  "sep_token_id": null,
+  "task_specific_params": {
+    "summarization": {
+      "length_penalty": 1.0,
+      "max_length": 128,
+      "min_length": 12,
+      "num_beams": 4
+    },
+    "summarization_cnn": {
+      "length_penalty": 2.0,
+      "max_length": 142,
+      "min_length": 56,
+      "num_beams": 4
+    },
+    "summarization_xsum": {
+      "length_penalty": 1.0,
+      "max_length": 62,
+      "min_length": 11,
+      "num_beams": 6
+    }
+  },
+  "temperature": 1.0,
+  "tie_encoder_decoder": false,
+  "tie_word_embeddings": true,
+  "tokenizer_class": null,
+  "top_k": 50,
+  "top_p": 1.0,
+  "torch_dtype": "float32",
+  "torchscript": false,
+  "transformers_version": "4.7.0.dev0",
+  "typical_p": 1.0,
+  "use_bfloat16": false,
+  "use_cache": true,
+  "vocab_size": 50265,
+  "vocab_size_or_config_json_file": -1,
+  "weight_bits": 16
+}

pytorch_model.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:82d38521aafb43be1c0b9f93be9c8f4011558be849ca4a9414bb949a32b09d0a
+size 432907017

training.log

@@ -0,0 +1,78 @@
+08/13/2023 00:40:06 - INFO - __main__ -   Distributed environment: NO
+Num processes: 1
+Process index: 0
+Local process index: 0
+Device: cuda
+Use FP16 precision: False
+
+08/13/2023 00:40:06 - WARNING - __main__ -   Namespace(dataset_name='s-nlp/paradetox', dataset_config_name=None, train_file=None, ignore_pad_token_for_loss=True, max_source_length=1024, source_prefix=None, preprocessing_num_workers=None, overwrite_cache=None, max_target_length=128, val_max_target_length=None, pad_to_max_length=False, model_name_or_path='s-nlp/bart-base-detox', config_name=None, tokenizer_name=None, text_column=None, summary_column=None, use_slow_tokenizer=False, per_device_train_batch_size=8, per_device_eval_batch_size=4, learning_rate=3e-05, weight_decay=0.0, num_train_epochs=10, max_train_steps=None, gradient_accumulation_steps=2, lr_scheduler_type=<SchedulerType.LINEAR: 'linear'>, warmup_ratio=0.05, output_dir='./output_s-nlp/paradetox_bart_base_detox/16_8_3_1_10_3e-05_fp16', seed=28, model_type=None, teacher_model='s-nlp/bart-base-detox', student_model='s-nlp/bart-base-detox', pred_distill=True, intermediate_distill=True, weight_bits=16, input_bits=8, clip_val=2.5, length_penalty=150, max_length=62, min_length=11, num_beams=6, do_train=True, do_test=True, test_teacher=False, distill_encoder=3, distill_decoder=1, log_steps=20, local_rank=0, weighted=False, new_distill_map=False, task_weight=1, logits_weight=1, hid_weight=1)
+08/13/2023 00:40:24 - INFO - __main__ -   ***** Running training *****
+08/13/2023 00:40:24 - INFO - __main__ -     Num examples = 19546
+08/13/2023 00:40:24 - INFO - __main__ -     Num Epochs = 10
+08/13/2023 00:40:24 - INFO - __main__ -     Instantaneous batch size per device = 8
+08/13/2023 00:40:24 - INFO - __main__ -     Total train batch size (w. parallel, distributed & accumulation) = 16
+08/13/2023 00:40:24 - INFO - __main__ -     Gradient Accumulation steps = 2
+08/13/2023 00:40:24 - INFO - __main__ -     Total optimization steps = 24440
+08/13/2023 00:40:24 - INFO - __main__ -     student encoder layers = 3
+08/13/2023 00:40:24 - INFO - __main__ -     student decoder layers = 1
+08/13/2023 00:40:24 - INFO - __main__ -     student encoder layers [0, 1, 2] is mapped with teacher encoder layers [0, 2, 5]
+08/13/2023 00:40:24 - INFO - __main__ -     student decoder layers [0] is mapped with teacher decoder layers [5]
+08/13/2023 00:48:32 - INFO - __main__ -   evaluation result: {'accuracy': 0.9501243829727173, 'similarity': 0.5612009167671204, 'fluency': 0.8357802033424377, 'joint': 0.4501223564147949, 'chrF': 22.59420505922646} 
+08/13/2023 00:57:03 - INFO - __main__ -   evaluation result: {'accuracy': 0.9501243829727173, 'similarity': 0.5612009167671204, 'fluency': 0.8357802033424377, 'joint': 0.4501223564147949, 'chrF': 54.978123489133125} 
+08/13/2023 01:05:13 - INFO - __main__ -   evaluation result: {'accuracy': 0.9501243829727173, 'similarity': 0.5612009167671204, 'fluency': 0.8357802033424377, 'joint': 0.4501223564147949, 'chrF': 62.023727727095206} 
+08/13/2023 01:13:26 - INFO - __main__ -   evaluation result: {'accuracy': 0.9501243829727173, 'similarity': 0.5612009167671204, 'fluency': 0.8357802033424377, 'joint': 0.4501223564147949, 'chrF': 63.97172971159594} 
+08/13/2023 01:21:23 - INFO - __main__ -   evaluation result: {'accuracy': 0.9501243829727173, 'similarity': 0.5612009167671204, 'fluency': 0.8357802033424377, 'joint': 0.4501223564147949, 'chrF': 64.5360970481176} 
+08/13/2023 01:29:39 - INFO - __main__ -   evaluation result: {'accuracy': 0.9501243829727173, 'similarity': 0.5612009167671204, 'fluency': 0.8357802033424377, 'joint': 0.4501223564147949, 'chrF': 64.94717856121133} 
+08/13/2023 01:37:40 - INFO - __main__ -   evaluation result: {'accuracy': 0.9501243829727173, 'similarity': 0.5612009167671204, 'fluency': 0.8357802033424377, 'joint': 0.4501223564147949, 'chrF': 64.59791205000518} 
+08/13/2023 01:45:57 - INFO - __main__ -   evaluation result: {'accuracy': 0.9501243829727173, 'similarity': 0.5612009167671204, 'fluency': 0.8357802033424377, 'joint': 0.4501223564147949, 'chrF': 65.09188210629405} 
+08/13/2023 01:54:25 - INFO - __main__ -   evaluation result: {'accuracy': 0.9501243829727173, 'similarity': 0.5612009167671204, 'fluency': 0.8357802033424377, 'joint': 0.4501223564147949, 'chrF': 65.43488256903395} 
+08/13/2023 02:02:48 - INFO - __main__ -   evaluation result: {'accuracy': 0.9501243829727173, 'similarity': 0.5612009167671204, 'fluency': 0.8357802033424377, 'joint': 0.4501223564147949, 'chrF': 65.13437994345524} 
+08/13/2023 13:33:25 - WARNING - __main__ -   You're running a t5 model but didn't provide a source prefix, which is the expected, e.g. with `--source_prefix 'summarize: ' `
+08/13/2023 13:33:25 - INFO - __main__ -   Distributed environment: NO
+Num processes: 1
+Process index: 0
+Local process index: 0
+Device: cuda
+Use FP16 precision: False
+
+08/13/2023 13:33:25 - WARNING - __main__ -   Namespace(dataset_name='s-nlp/paradetox', dataset_config_name=None, train_file=None, ignore_pad_token_for_loss=True, max_source_length=1024, source_prefix=None, preprocessing_num_workers=None, overwrite_cache=None, max_target_length=128, val_max_target_length=None, pad_to_max_length=False, model_name_or_path='t5-large', config_name=None, tokenizer_name=None, text_column=None, summary_column=None, use_slow_tokenizer=False, per_device_train_batch_size=8, per_device_eval_batch_size=4, learning_rate=3e-05, weight_decay=0.0, num_train_epochs=10, max_train_steps=None, gradient_accumulation_steps=2, lr_scheduler_type=<SchedulerType.LINEAR: 'linear'>, warmup_ratio=0.05, output_dir='./output_s-nlp/paradetox_bart_base_detox/16_8_3_1_10_3e-05_fp16', seed=28, model_type=None, teacher_model='t5-large', student_model='t5-large', pred_distill=True, intermediate_distill=True, weight_bits=16, input_bits=8, clip_val=2.5, length_penalty=150, max_length=62, min_length=11, num_beams=6, do_train=True, do_test=True, test_teacher=False, distill_encoder=3, distill_decoder=1, log_steps=20, local_rank=0, weighted=False, new_distill_map=False, task_weight=1, logits_weight=1, hid_weight=1)
+08/13/2023 13:33:49 - INFO - __main__ -   ***** Running training *****
+08/13/2023 13:33:49 - INFO - __main__ -     Num examples = 19546
+08/13/2023 13:33:49 - INFO - __main__ -     Num Epochs = 10
+08/13/2023 13:33:49 - INFO - __main__ -     Instantaneous batch size per device = 8
+08/13/2023 13:33:49 - INFO - __main__ -     Total train batch size (w. parallel, distributed & accumulation) = 16
+08/13/2023 13:33:49 - INFO - __main__ -     Gradient Accumulation steps = 2
+08/13/2023 13:33:49 - INFO - __main__ -     Total optimization steps = 24440
+08/13/2023 13:33:49 - INFO - __main__ -     student encoder layers = 3
+08/13/2023 13:33:49 - INFO - __main__ -     student decoder layers = 1
+08/13/2023 13:33:49 - INFO - __main__ -     student encoder layers [0, 1, 2] is mapped with teacher encoder layers [0, 2, 5]
+08/13/2023 13:33:49 - INFO - __main__ -     student decoder layers [0] is mapped with teacher decoder layers [5]
+08/13/2023 18:13:36 - INFO - __main__ -   Distributed environment: NO
+Num processes: 1
+Process index: 0
+Local process index: 0
+Device: cuda
+Use FP16 precision: False
+
+08/13/2023 18:13:36 - WARNING - __main__ -   Namespace(dataset_name='s-nlp/paradetox', dataset_config_name=None, train_file=None, ignore_pad_token_for_loss=True, max_source_length=1024, source_prefix=None, preprocessing_num_workers=None, overwrite_cache=None, max_target_length=128, val_max_target_length=None, pad_to_max_length=False, model_name_or_path='facebook/bart-large', config_name=None, tokenizer_name=None, text_column=None, summary_column=None, use_slow_tokenizer=False, per_device_train_batch_size=8, per_device_eval_batch_size=4, learning_rate=3e-05, weight_decay=0.0, num_train_epochs=10, max_train_steps=None, gradient_accumulation_steps=2, lr_scheduler_type=<SchedulerType.LINEAR: 'linear'>, warmup_ratio=0.05, output_dir='./output_s-nlp/paradetox_bart_base_detox/16_8_3_1_10_3e-05_fp16', seed=28, model_type=None, teacher_model='facebook/bart-large', student_model='facebook/bart-large', pred_distill=True, intermediate_distill=True, weight_bits=16, input_bits=8, clip_val=2.5, length_penalty=150, max_length=62, min_length=11, num_beams=6, do_train=True, do_test=True, test_teacher=False, distill_encoder=3, distill_decoder=1, log_steps=20, local_rank=0, weighted=False, new_distill_map=False, task_weight=1, logits_weight=1, hid_weight=1)
+08/13/2023 18:13:57 - INFO - __main__ -   ***** Running training *****
+08/13/2023 18:13:57 - INFO - __main__ -     Num examples = 19546
+08/13/2023 18:13:57 - INFO - __main__ -     Num Epochs = 10
+08/13/2023 18:13:57 - INFO - __main__ -     Instantaneous batch size per device = 8
+08/13/2023 18:13:57 - INFO - __main__ -     Total train batch size (w. parallel, distributed & accumulation) = 16
+08/13/2023 18:13:57 - INFO - __main__ -     Gradient Accumulation steps = 2
+08/13/2023 18:13:57 - INFO - __main__ -     Total optimization steps = 24440
+08/13/2023 18:13:57 - INFO - __main__ -     student encoder layers = 3
+08/13/2023 18:13:57 - INFO - __main__ -     student decoder layers = 1
+08/13/2023 18:13:57 - INFO - __main__ -     student encoder layers [0, 1, 2] is mapped with teacher encoder layers [0, 2, 5]
+08/13/2023 18:13:57 - INFO - __main__ -     student decoder layers [0] is mapped with teacher decoder layers [5]
+08/13/2023 18:23:59 - INFO - __main__ -   evaluation result: {'accuracy': 0.9501243829727173, 'similarity': 0.5612009167671204, 'fluency': 0.8357802033424377, 'joint': 0.4501223564147949, 'chrF': 13.927599253697903} 
+08/13/2023 18:33:46 - INFO - __main__ -   evaluation result: {'accuracy': 0.9501243829727173, 'similarity': 0.5612009167671204, 'fluency': 0.8357802033424377, 'joint': 0.4501223564147949, 'chrF': 14.54024753825692} 
+08/13/2023 18:43:47 - INFO - __main__ -   evaluation result: {'accuracy': 0.9501243829727173, 'similarity': 0.5612009167671204, 'fluency': 0.8357802033424377, 'joint': 0.4501223564147949, 'chrF': 15.696559967829156} 
+08/13/2023 18:53:59 - INFO - __main__ -   evaluation result: {'accuracy': 0.9501243829727173, 'similarity': 0.5612009167671204, 'fluency': 0.8357802033424377, 'joint': 0.4501223564147949, 'chrF': 17.028561047497664} 
+08/13/2023 19:04:19 - INFO - __main__ -   evaluation result: {'accuracy': 0.9501243829727173, 'similarity': 0.5612009167671204, 'fluency': 0.8357802033424377, 'joint': 0.4501223564147949, 'chrF': 18.190088344180207} 
+08/13/2023 19:14:34 - INFO - __main__ -   evaluation result: {'accuracy': 0.9501243829727173, 'similarity': 0.5612009167671204, 'fluency': 0.8357802033424377, 'joint': 0.4501223564147949, 'chrF': 19.35744020416755} 
+08/13/2023 19:24:59 - INFO - __main__ -   evaluation result: {'accuracy': 0.9501243829727173, 'similarity': 0.5612009167671204, 'fluency': 0.8357802033424377, 'joint': 0.4501223564147949, 'chrF': 20.69450840196689} 
+08/13/2023 19:35:05 - INFO - __main__ -   evaluation result: {'accuracy': 0.9501243829727173, 'similarity': 0.5612009167671204, 'fluency': 0.8357802033424377, 'joint': 0.4501223564147949, 'chrF': 21.69196456544141} 
+08/13/2023 19:45:14 - INFO - __main__ -   evaluation result: {'accuracy': 0.9501243829727173, 'similarity': 0.5612009167671204, 'fluency': 0.8357802033424377, 'joint': 0.4501223564147949, 'chrF': 22.46896551649043} 
+08/13/2023 19:55:18 - INFO - __main__ -   evaluation result: {'accuracy': 0.9501243829727173, 'similarity': 0.5612009167671204, 'fluency': 0.8357802033424377, 'joint': 0.4501223564147949, 'chrF': 22.655778155996654}