python run_qa.py --model_name_or_path /data1/vchua/tld-poc/bert-base-squadv1-local-hybrid-compiled --optimize_model_before_eval --optimized_checkpoint /data1/vchua/tld-poc/bert-base-squadv1-local-hybrid-filled-lt-compiled --dataset_name squad --do_eval --do_train --evaluation_strategy steps --eval_steps 250 --learning_rate 3e-5 --teacher bert-large-uncased-whole-word-masking-finetuned-squad --teacher_ratio 0.9 --lr_scheduler_type cosine_with_restarts --warmup_ratio 0.25 --cosine_cycles 1 --num_train_epochs 5 --per_device_eval_batch_size 128 --per_device_train_batch_size 16 --max_seq_length 384 --doc_stride 128 --save_steps 250 --nncf_config nncf_bert_squad_sparsity.json --logging_steps 1 --overwrite_output_dir --run_name run10-bert-squad-cropped-qat-customkd-lt-5eph --output_dir /data1/vchua/tld-poc-csr-dgx1-03//run10-bert-squad-cropped-qat-customkd-lt-5eph --max_steps 25 ### End of CMD --- 01/16/2022 13:20:03 - WARNING - __main__ - Process rank: -1, device: cuda:0, n_gpu: 1distributed training: False, 16-bits training: False 01/16/2022 13:20:03 - INFO - __main__ - Training/evaluation parameters TrainingArguments( _n_gpu=1, adafactor=False, adam_beta1=0.9, adam_beta2=0.999, adam_epsilon=1e-08, cosine_cycles=1, dataloader_drop_last=False, dataloader_num_workers=0, dataloader_pin_memory=True, ddp_find_unused_parameters=None, debug=[], deepspeed=None, disable_tqdm=False, distill_temp=2.0, do_eval=True, do_predict=False, do_train=True, eval_accumulation_steps=None, eval_steps=250, evaluation_strategy=IntervalStrategy.STEPS, fp16=False, fp16_backend=auto, fp16_full_eval=False, fp16_opt_level=O1, gradient_accumulation_steps=1, greater_is_better=None, group_by_length=False, ignore_data_skip=False, label_names=None, label_smoothing_factor=0.0, learning_rate=3e-05, length_column_name=length, load_best_model_at_end=False, local_rank=-1, log_level=-1, log_level_replica=-1, log_on_each_node=True, logging_dir=/data1/vchua/tld-poc-csr-dgx1-03//run10-bert-squad-cropped-qat-customkd-lt-5eph/runs/Jan16_13-20-03_csr-dgx1-03, logging_first_step=False, logging_steps=1, logging_strategy=IntervalStrategy.STEPS, lr_scheduler_type=SchedulerType.COSINE_WITH_RESTARTS, max_grad_norm=1.0, max_steps=25, metric_for_best_model=None, mp_parameters=, nncf_config=nncf_bert_squad_sparsity.json, no_cuda=False, num_train_epochs=5.0, optimize_model_before_eval=True, optimized_checkpoint=/data1/vchua/tld-poc/bert-base-squadv1-local-hybrid-filled-lt-compiled, output_dir=/data1/vchua/tld-poc-csr-dgx1-03//run10-bert-squad-cropped-qat-customkd-lt-5eph, overwrite_output_dir=True, past_index=-1, per_device_eval_batch_size=128, per_device_train_batch_size=16, prediction_loss_only=False, push_to_hub=False, push_to_hub_model_id=run10-bert-squad-cropped-qat-customkd-lt-5eph, push_to_hub_organization=None, push_to_hub_token=None, qat_checkpoint=None, remove_unused_columns=True, report_to=['mlflow', 'tensorboard', 'wandb'], resume_from_checkpoint=None, run_name=run10-bert-squad-cropped-qat-customkd-lt-5eph, save_on_each_node=False, save_steps=250, save_strategy=IntervalStrategy.STEPS, save_total_limit=None, seed=42, sharded_ddp=[], skip_memory_metrics=True, teacher=bert-large-uncased-whole-word-masking-finetuned-squad, teacher_ratio=0.9, to_onnx=None, tpu_metrics_debug=False, tpu_num_cores=None, use_legacy_prediction_loop=False, warmup_ratio=0.25, warmup_steps=0, weight_decay=0.0, ) 01/16/2022 13:20:03 - INFO - datasets.builder - No config specified, defaulting to first: squad/plain_text 01/16/2022 13:20:03 - INFO - datasets.info - Loading Dataset Infos from /home/vchua/.cache/huggingface/modules/datasets_modules/datasets/squad/d6ec3ceb99ca480ce37cdd35555d6cb2511d223b9150cce08a837ef62ffea453 01/16/2022 13:20:03 - INFO - datasets.builder - Overwrite dataset info from restored data version. 01/16/2022 13:20:03 - INFO - datasets.info - Loading Dataset info from /home/vchua/.cache/huggingface/datasets/squad/plain_text/1.0.0/d6ec3ceb99ca480ce37cdd35555d6cb2511d223b9150cce08a837ef62ffea453 01/16/2022 13:20:03 - WARNING - datasets.builder - Reusing dataset squad (/home/vchua/.cache/huggingface/datasets/squad/plain_text/1.0.0/d6ec3ceb99ca480ce37cdd35555d6cb2511d223b9150cce08a837ef62ffea453) 01/16/2022 13:20:03 - INFO - datasets.info - Loading Dataset info from /home/vchua/.cache/huggingface/datasets/squad/plain_text/1.0.0/d6ec3ceb99ca480ce37cdd35555d6cb2511d223b9150cce08a837ef62ffea453 /home/vchua/tld-poc/nncf/nncf/torch/dynamic_graph/patch_pytorch.py:163: UserWarning: Not patching unique_dim since it is missing in this version of PyTorch warnings.warn("Not patching {} since it is missing in this version of PyTorch".format(op_name)) /home/vchua/tld-poc/nncf/nncf/torch/dynamic_graph/patch_pytorch.py:163: UserWarning: Not patching unique_dim since it is missing in this version of PyTorch warnings.warn("Not patching {} since it is missing in this version of PyTorch".format(op_name)) 0%| | 0/2 [00:00> loading configuration file /data1/vchua/tld-poc/bert-base-squadv1-local-hybrid-compiled/config.json [INFO|configuration_utils.py:581] 2022-01-16 13:20:03,990 >> Model config BertConfig { "_name_or_path": "bert-base-uncased", "architectures": [ "BertForQuestionAnswering" ], "attention_probs_dropout_prob": 0.1, "classifier_dropout": null, "gradient_checkpointing": false, "hidden_act": "gelu", "hidden_dropout_prob": 0.1, "hidden_size": 768, "initializer_range": 0.02, "intermediate_size": 3072, "layer_norm_eps": 1e-12, "max_position_embeddings": 512, "model_type": "bert", "num_attention_heads": 12, "num_hidden_layers": 12, "pad_token_id": 0, "position_embedding_type": "absolute", "pruned_heads": { "0": [ 0, 2, 4, 5, 6, 7, 11 ], "1": [ 0, 2, 3, 5, 6, 7, 8 ], "2": [ 8, 4, 7 ], "3": [ 2, 4, 6 ], "4": [ 1, 2, 11 ], "5": [ 1, 2, 5, 6, 7, 11 ], "6": [ 0, 2, 3, 7, 10 ], "7": [ 1, 3, 6, 7, 11 ], "8": [ 0, 3, 4, 5, 8 ], "9": [ 1, 3, 4, 5, 7, 9, 10 ], "10": [ 1, 4, 5, 6, 7, 8 ], "11": [ 4, 5, 7, 8, 10, 11 ] }, "torch_dtype": "float32", "transformers_version": "4.9.1", "type_vocab_size": 2, "use_cache": true, "vocab_size": 30522 } [INFO|tokenization_utils_base.py:1664] 2022-01-16 13:20:03,991 >> Didn't find file /data1/vchua/tld-poc/bert-base-squadv1-local-hybrid-compiled/added_tokens.json. We won't load it. [INFO|tokenization_utils_base.py:1728] 2022-01-16 13:20:03,991 >> loading file /data1/vchua/tld-poc/bert-base-squadv1-local-hybrid-compiled/vocab.txt [INFO|tokenization_utils_base.py:1728] 2022-01-16 13:20:03,991 >> loading file /data1/vchua/tld-poc/bert-base-squadv1-local-hybrid-compiled/tokenizer.json [INFO|tokenization_utils_base.py:1728] 2022-01-16 13:20:03,991 >> loading file None [INFO|tokenization_utils_base.py:1728] 2022-01-16 13:20:03,991 >> loading file /data1/vchua/tld-poc/bert-base-squadv1-local-hybrid-compiled/special_tokens_map.json [INFO|tokenization_utils_base.py:1728] 2022-01-16 13:20:03,991 >> loading file /data1/vchua/tld-poc/bert-base-squadv1-local-hybrid-compiled/tokenizer_config.json 01/16/2022 13:20:04 - WARNING - datasets.arrow_dataset - Loading cached processed dataset at /home/vchua/.cache/huggingface/datasets/squad/plain_text/1.0.0/d6ec3ceb99ca480ce37cdd35555d6cb2511d223b9150cce08a837ef62ffea453/cache-65c15074471a5f81.arrow 01/16/2022 13:20:04 - WARNING - datasets.arrow_dataset - Loading cached processed dataset at /home/vchua/.cache/huggingface/datasets/squad/plain_text/1.0.0/d6ec3ceb99ca480ce37cdd35555d6cb2511d223b9150cce08a837ef62ffea453/cache-c7d0554dffe9c91c.arrow [INFO|configuration_utils.py:545] 2022-01-16 13:20:05,280 >> loading configuration file https://huggingface.co/bert-large-uncased-whole-word-masking-finetuned-squad/resolve/main/config.json from cache at /home/vchua/.cache/huggingface/transformers/402f6d8c99fdd3bffd354782842e2b5a6be81f80ab630591051ebc78ca726f39.ebffac96fee44dbe30674c204dd3d3f358c1b8c33100281ecdd688514f41410a [INFO|configuration_utils.py:581] 2022-01-16 13:20:05,281 >> Model config BertConfig { "architectures": [ "BertForQuestionAnswering" ], "attention_probs_dropout_prob": 0.1, "gradient_checkpointing": false, "hidden_act": "gelu", "hidden_dropout_prob": 0.1, "hidden_size": 1024, "initializer_range": 0.02, "intermediate_size": 4096, "layer_norm_eps": 1e-12, "max_position_embeddings": 512, "model_type": "bert", "num_attention_heads": 16, "num_hidden_layers": 24, "pad_token_id": 0, "position_embedding_type": "absolute", "transformers_version": "4.9.1", "type_vocab_size": 2, "use_cache": true, "vocab_size": 30522 } [INFO|modeling_utils.py:1280] 2022-01-16 13:20:05,631 >> loading weights file https://huggingface.co/bert-large-uncased-whole-word-masking-finetuned-squad/resolve/main/pytorch_model.bin from cache at /home/vchua/.cache/huggingface/transformers/28a060c1e2e1216bd9c8f5222ce38ce916c4829b8b05e027fe91510f3fd4da7e.50fc4a146342b3a6a99b185af3d5b70163b64d45790be64d9124dcccbcd3915e [INFO|modeling_utils.py:1574] 2022-01-16 13:20:08,823 >> All model checkpoint weights were used when initializing BertForQuestionAnswering. [INFO|modeling_utils.py:1582] 2022-01-16 13:20:08,823 >> All the weights of BertForQuestionAnswering were initialized from the model checkpoint at bert-large-uncased-whole-word-masking-finetuned-squad. If your task is similar to the task the model of the checkpoint was trained on, you can already use BertForQuestionAnswering for predictions without further training. [INFO|modeling_utils.py:1278] 2022-01-16 13:20:08,825 >> loading weights file /data1/vchua/tld-poc/bert-base-squadv1-local-hybrid-compiled/pytorch_model.bin [INFO|modeling_utils.py:1574] 2022-01-16 13:20:10,073 >> All model checkpoint weights were used when initializing BertForQuestionAnswering. [INFO|modeling_utils.py:1582] 2022-01-16 13:20:10,073 >> All the weights of BertForQuestionAnswering were initialized from the model checkpoint at /data1/vchua/tld-poc/bert-base-squadv1-local-hybrid-compiled. If your task is similar to the task the model of the checkpoint was trained on, you can already use BertForQuestionAnswering for predictions without further training. removed heads 0, total_heads=81, percentage removed=0.0 bert.encoder.layer.0.intermediate.dense, sparsity = 93.98 bert.encoder.layer.0.output.dense, sparsity = 93.98 bert.encoder.layer.1.intermediate.dense, sparsity = 89.75 bert.encoder.layer.1.output.dense, sparsity = 89.75 bert.encoder.layer.2.intermediate.dense, sparsity = 88.96 bert.encoder.layer.2.output.dense, sparsity = 88.96 bert.encoder.layer.3.intermediate.dense, sparsity = 88.02 bert.encoder.layer.3.output.dense, sparsity = 88.02 bert.encoder.layer.4.intermediate.dense, sparsity = 87.43 bert.encoder.layer.4.output.dense, sparsity = 87.43 bert.encoder.layer.5.intermediate.dense, sparsity = 89.06 bert.encoder.layer.5.output.dense, sparsity = 89.06 bert.encoder.layer.6.intermediate.dense, sparsity = 90.89 bert.encoder.layer.6.output.dense, sparsity = 90.89 bert.encoder.layer.7.intermediate.dense, sparsity = 93.13 bert.encoder.layer.7.output.dense, sparsity = 93.13 bert.encoder.layer.8.intermediate.dense, sparsity = 96.48 bert.encoder.layer.8.output.dense, sparsity = 96.48 bert.encoder.layer.9.intermediate.dense, sparsity = 98.27 bert.encoder.layer.9.output.dense, sparsity = 98.27 bert.encoder.layer.10.intermediate.dense, sparsity = 97.20 bert.encoder.layer.10.output.dense, sparsity = 97.20 bert.encoder.layer.11.intermediate.dense, sparsity = 96.58 bert.encoder.layer.11.output.dense, sparsity = 96.58 WARNING:nncf:Graphviz is not installed - only the .dot model visualization format will be used. Install pygraphviz into your Python environment and graphviz system-wide to enable PNG rendering. INFO:nncf:Wrapping module BertForQuestionAnswering/BertModel[bert]/BertEmbeddings[embeddings]/Embedding[word_embeddings] by BertForQuestionAnswering/BertModel[bert]/BertEmbeddings[embeddings]/NNCFEmbedding[word_embeddings] INFO:nncf:Wrapping module BertForQuestionAnswering/BertModel[bert]/BertEmbeddings[embeddings]/Embedding[position_embeddings] by BertForQuestionAnswering/BertModel[bert]/BertEmbeddings[embeddings]/NNCFEmbedding[position_embeddings] INFO:nncf:Wrapping module BertForQuestionAnswering/BertModel[bert]/BertEmbeddings[embeddings]/Embedding[token_type_embeddings] by BertForQuestionAnswering/BertModel[bert]/BertEmbeddings[embeddings]/NNCFEmbedding[token_type_embeddings] INFO:nncf:Wrapping module BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[0]/BertAttention[attention]/BertSelfAttention[self]/Linear[query] by BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[0]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[query] INFO:nncf:Wrapping module BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[0]/BertAttention[attention]/BertSelfAttention[self]/Linear[key] by BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[0]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[key] INFO:nncf:Wrapping module BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[0]/BertAttention[attention]/BertSelfAttention[self]/Linear[value] by BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[0]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[value] INFO:nncf:Wrapping module BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[0]/BertAttention[attention]/BertSelfOutput[output]/Linear[dense] by BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[0]/BertAttention[attention]/BertSelfOutput[output]/NNCFLinear[dense] INFO:nncf:Wrapping module BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[0]/BertIntermediate[intermediate]/Linear[dense] by BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[0]/BertIntermediate[intermediate]/NNCFLinear[dense] INFO:nncf:Wrapping module BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[0]/BertOutput[output]/Linear[dense] by BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[0]/BertOutput[output]/NNCFLinear[dense] INFO:nncf:Wrapping module BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[1]/BertAttention[attention]/BertSelfAttention[self]/Linear[query] by BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[1]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[query] INFO:nncf:Wrapping module BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[1]/BertAttention[attention]/BertSelfAttention[self]/Linear[key] by BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[1]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[key] INFO:nncf:Wrapping module BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[1]/BertAttention[attention]/BertSelfAttention[self]/Linear[value] by BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[1]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[value] INFO:nncf:Wrapping module BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[1]/BertAttention[attention]/BertSelfOutput[output]/Linear[dense] by BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[1]/BertAttention[attention]/BertSelfOutput[output]/NNCFLinear[dense] INFO:nncf:Wrapping module BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[1]/BertIntermediate[intermediate]/Linear[dense] by BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[1]/BertIntermediate[intermediate]/NNCFLinear[dense] INFO:nncf:Wrapping module BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[1]/BertOutput[output]/Linear[dense] by BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[1]/BertOutput[output]/NNCFLinear[dense] INFO:nncf:Wrapping module BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[2]/BertAttention[attention]/BertSelfAttention[self]/Linear[query] by BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[2]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[query] INFO:nncf:Wrapping module BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[2]/BertAttention[attention]/BertSelfAttention[self]/Linear[key] by BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[2]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[key] INFO:nncf:Wrapping module BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[2]/BertAttention[attention]/BertSelfAttention[self]/Linear[value] by BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[2]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[value] INFO:nncf:Wrapping module BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[2]/BertAttention[attention]/BertSelfOutput[output]/Linear[dense] by BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[2]/BertAttention[attention]/BertSelfOutput[output]/NNCFLinear[dense] INFO:nncf:Wrapping module BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[2]/BertIntermediate[intermediate]/Linear[dense] by BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[2]/BertIntermediate[intermediate]/NNCFLinear[dense] INFO:nncf:Wrapping module BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[2]/BertOutput[output]/Linear[dense] by BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[2]/BertOutput[output]/NNCFLinear[dense] INFO:nncf:Wrapping module BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[3]/BertAttention[attention]/BertSelfAttention[self]/Linear[query] by BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[3]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[query] INFO:nncf:Wrapping module BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[3]/BertAttention[attention]/BertSelfAttention[self]/Linear[key] by BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[3]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[key] INFO:nncf:Wrapping module BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[3]/BertAttention[attention]/BertSelfAttention[self]/Linear[value] by BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[3]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[value] INFO:nncf:Wrapping module BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[3]/BertAttention[attention]/BertSelfOutput[output]/Linear[dense] by BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[3]/BertAttention[attention]/BertSelfOutput[output]/NNCFLinear[dense] INFO:nncf:Wrapping module BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[3]/BertIntermediate[intermediate]/Linear[dense] by BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[3]/BertIntermediate[intermediate]/NNCFLinear[dense] INFO:nncf:Wrapping module BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[3]/BertOutput[output]/Linear[dense] by BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[3]/BertOutput[output]/NNCFLinear[dense] INFO:nncf:Wrapping module BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[4]/BertAttention[attention]/BertSelfAttention[self]/Linear[query] by BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[4]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[query] INFO:nncf:Wrapping module BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[4]/BertAttention[attention]/BertSelfAttention[self]/Linear[key] by BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[4]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[key] INFO:nncf:Wrapping module BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[4]/BertAttention[attention]/BertSelfAttention[self]/Linear[value] by BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[4]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[value] INFO:nncf:Wrapping module BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[4]/BertAttention[attention]/BertSelfOutput[output]/Linear[dense] by BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[4]/BertAttention[attention]/BertSelfOutput[output]/NNCFLinear[dense] INFO:nncf:Wrapping module BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[4]/BertIntermediate[intermediate]/Linear[dense] by BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[4]/BertIntermediate[intermediate]/NNCFLinear[dense] INFO:nncf:Wrapping module BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[4]/BertOutput[output]/Linear[dense] by BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[4]/BertOutput[output]/NNCFLinear[dense] INFO:nncf:Wrapping module BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[5]/BertAttention[attention]/BertSelfAttention[self]/Linear[query] by BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[5]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[query] INFO:nncf:Wrapping module BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[5]/BertAttention[attention]/BertSelfAttention[self]/Linear[key] by BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[5]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[key] INFO:nncf:Wrapping module BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[5]/BertAttention[attention]/BertSelfAttention[self]/Linear[value] by BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[5]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[value] INFO:nncf:Wrapping module BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[5]/BertAttention[attention]/BertSelfOutput[output]/Linear[dense] by BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[5]/BertAttention[attention]/BertSelfOutput[output]/NNCFLinear[dense] INFO:nncf:Wrapping module BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[5]/BertIntermediate[intermediate]/Linear[dense] by BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[5]/BertIntermediate[intermediate]/NNCFLinear[dense] INFO:nncf:Wrapping module BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[5]/BertOutput[output]/Linear[dense] by BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[5]/BertOutput[output]/NNCFLinear[dense] INFO:nncf:Wrapping module BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[6]/BertAttention[attention]/BertSelfAttention[self]/Linear[query] by BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[6]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[query] INFO:nncf:Wrapping module BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[6]/BertAttention[attention]/BertSelfAttention[self]/Linear[key] by BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[6]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[key] INFO:nncf:Wrapping module BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[6]/BertAttention[attention]/BertSelfAttention[self]/Linear[value] by BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[6]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[value] INFO:nncf:Wrapping module BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[6]/BertAttention[attention]/BertSelfOutput[output]/Linear[dense] by BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[6]/BertAttention[attention]/BertSelfOutput[output]/NNCFLinear[dense] INFO:nncf:Wrapping module BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[6]/BertIntermediate[intermediate]/Linear[dense] by BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[6]/BertIntermediate[intermediate]/NNCFLinear[dense] INFO:nncf:Wrapping module BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[6]/BertOutput[output]/Linear[dense] by BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[6]/BertOutput[output]/NNCFLinear[dense] INFO:nncf:Wrapping module BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[7]/BertAttention[attention]/BertSelfAttention[self]/Linear[query] by BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[7]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[query] INFO:nncf:Wrapping module BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[7]/BertAttention[attention]/BertSelfAttention[self]/Linear[key] by BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[7]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[key] INFO:nncf:Wrapping module BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[7]/BertAttention[attention]/BertSelfAttention[self]/Linear[value] by BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[7]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[value] INFO:nncf:Wrapping module BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[7]/BertAttention[attention]/BertSelfOutput[output]/Linear[dense] by BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[7]/BertAttention[attention]/BertSelfOutput[output]/NNCFLinear[dense] INFO:nncf:Wrapping module BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[7]/BertIntermediate[intermediate]/Linear[dense] by BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[7]/BertIntermediate[intermediate]/NNCFLinear[dense] INFO:nncf:Wrapping module BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[7]/BertOutput[output]/Linear[dense] by BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[7]/BertOutput[output]/NNCFLinear[dense] INFO:nncf:Wrapping module BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[8]/BertAttention[attention]/BertSelfAttention[self]/Linear[query] by BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[8]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[query] INFO:nncf:Wrapping module BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[8]/BertAttention[attention]/BertSelfAttention[self]/Linear[key] by BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[8]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[key] INFO:nncf:Wrapping module BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[8]/BertAttention[attention]/BertSelfAttention[self]/Linear[value] by BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[8]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[value] INFO:nncf:Wrapping module BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[8]/BertAttention[attention]/BertSelfOutput[output]/Linear[dense] by BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[8]/BertAttention[attention]/BertSelfOutput[output]/NNCFLinear[dense] INFO:nncf:Wrapping module BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[8]/BertIntermediate[intermediate]/Linear[dense] by BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[8]/BertIntermediate[intermediate]/NNCFLinear[dense] INFO:nncf:Wrapping module BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[8]/BertOutput[output]/Linear[dense] by BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[8]/BertOutput[output]/NNCFLinear[dense] INFO:nncf:Wrapping module BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[9]/BertAttention[attention]/BertSelfAttention[self]/Linear[query] by BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[9]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[query] INFO:nncf:Wrapping module BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[9]/BertAttention[attention]/BertSelfAttention[self]/Linear[key] by BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[9]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[key] INFO:nncf:Wrapping module BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[9]/BertAttention[attention]/BertSelfAttention[self]/Linear[value] by BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[9]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[value] INFO:nncf:Wrapping module BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[9]/BertAttention[attention]/BertSelfOutput[output]/Linear[dense] by BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[9]/BertAttention[attention]/BertSelfOutput[output]/NNCFLinear[dense] INFO:nncf:Wrapping module BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[9]/BertIntermediate[intermediate]/Linear[dense] by BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[9]/BertIntermediate[intermediate]/NNCFLinear[dense] INFO:nncf:Wrapping module BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[9]/BertOutput[output]/Linear[dense] by BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[9]/BertOutput[output]/NNCFLinear[dense] INFO:nncf:Wrapping module BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[10]/BertAttention[attention]/BertSelfAttention[self]/Linear[query] by BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[10]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[query] INFO:nncf:Wrapping module BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[10]/BertAttention[attention]/BertSelfAttention[self]/Linear[key] by BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[10]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[key] INFO:nncf:Wrapping module BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[10]/BertAttention[attention]/BertSelfAttention[self]/Linear[value] by BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[10]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[value] INFO:nncf:Wrapping module BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[10]/BertAttention[attention]/BertSelfOutput[output]/Linear[dense] by BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[10]/BertAttention[attention]/BertSelfOutput[output]/NNCFLinear[dense] INFO:nncf:Wrapping module BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[10]/BertIntermediate[intermediate]/Linear[dense] by BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[10]/BertIntermediate[intermediate]/NNCFLinear[dense] INFO:nncf:Wrapping module BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[10]/BertOutput[output]/Linear[dense] by BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[10]/BertOutput[output]/NNCFLinear[dense] INFO:nncf:Wrapping module BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[11]/BertAttention[attention]/BertSelfAttention[self]/Linear[query] by BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[11]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[query] INFO:nncf:Wrapping module BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[11]/BertAttention[attention]/BertSelfAttention[self]/Linear[key] by BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[11]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[key] INFO:nncf:Wrapping module BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[11]/BertAttention[attention]/BertSelfAttention[self]/Linear[value] by BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[11]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[value] INFO:nncf:Wrapping module BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[11]/BertAttention[attention]/BertSelfOutput[output]/Linear[dense] by BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[11]/BertAttention[attention]/BertSelfOutput[output]/NNCFLinear[dense] INFO:nncf:Wrapping module BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[11]/BertIntermediate[intermediate]/Linear[dense] by BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[11]/BertIntermediate[intermediate]/NNCFLinear[dense] INFO:nncf:Wrapping module BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[11]/BertOutput[output]/Linear[dense] by BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[11]/BertOutput[output]/NNCFLinear[dense] INFO:nncf:Wrapping module BertForQuestionAnswering/Linear[qa_outputs] by BertForQuestionAnswering/NNCFLinear[qa_outputs] WARNING:nncf:Preset quantizer parameters {'mode'} explicitly overrided. WARNING:nncf:Preset quantizer parameters {'mode'} explicitly overrided. WARNING:nncf:Could not find an associated input activation quantizer for a weighted node with quantizable weights: BertForQuestionAnswering/BertModel[bert]/BertEmbeddings[embeddings]/NNCFEmbedding[word_embeddings]/embedding_0 WARNING:nncf:Could not find an associated input activation quantizer for a weighted node with quantizable weights: BertForQuestionAnswering/BertModel[bert]/BertEmbeddings[embeddings]/NNCFEmbedding[position_embeddings]/embedding_0 WARNING:nncf:Could not find an associated input activation quantizer for a weighted node with quantizable weights: BertForQuestionAnswering/BertModel[bert]/BertEmbeddings[embeddings]/NNCFEmbedding[token_type_embeddings]/embedding_0 WARNING:nncf:Attempted to use weight quantizer of 4 BertForQuestionAnswering/BertModel[bert]/BertEmbeddings[embeddings]/NNCFEmbedding[word_embeddings]/embedding_0 to quantize input of {'6 BertForQuestionAnswering/BertModel[bert]/BertEmbeddings[embeddings]/__add___0'}, but no compatible configs were found. WARNING:nncf:Attempted to use weight quantizer of 5 BertForQuestionAnswering/BertModel[bert]/BertEmbeddings[embeddings]/NNCFEmbedding[token_type_embeddings]/embedding_0 to quantize input of {'6 BertForQuestionAnswering/BertModel[bert]/BertEmbeddings[embeddings]/__add___0'}, but no compatible configs were found. WARNING:nncf:Attempted to use weight quantizer of 7 BertForQuestionAnswering/BertModel[bert]/BertEmbeddings[embeddings]/NNCFEmbedding[position_embeddings]/embedding_0 to quantize input of {'8 BertForQuestionAnswering/BertModel[bert]/BertEmbeddings[embeddings]/__iadd___0'}, but no compatible configs were found. WARNING:nncf:Could not find an associated input activation quantizer for a weighted node with quantizable weights: BertForQuestionAnswering/BertModel[bert]/BertEmbeddings[embeddings]/NNCFEmbedding[word_embeddings]/embedding_0 WARNING:nncf:Could not find an associated input activation quantizer for a weighted node with quantizable weights: BertForQuestionAnswering/BertModel[bert]/BertEmbeddings[embeddings]/NNCFEmbedding[position_embeddings]/embedding_0 WARNING:nncf:Could not find an associated input activation quantizer for a weighted node with quantizable weights: BertForQuestionAnswering/BertModel[bert]/BertEmbeddings[embeddings]/NNCFEmbedding[token_type_embeddings]/embedding_0 WARNING:nncf:Attempted to use weight quantizer of 4 BertForQuestionAnswering/BertModel[bert]/BertEmbeddings[embeddings]/NNCFEmbedding[word_embeddings]/embedding_0 to quantize input of {'6 BertForQuestionAnswering/BertModel[bert]/BertEmbeddings[embeddings]/__add___0'}, but no compatible configs were found. WARNING:nncf:Attempted to use weight quantizer of 5 BertForQuestionAnswering/BertModel[bert]/BertEmbeddings[embeddings]/NNCFEmbedding[token_type_embeddings]/embedding_0 to quantize input of {'6 BertForQuestionAnswering/BertModel[bert]/BertEmbeddings[embeddings]/__add___0'}, but no compatible configs were found. WARNING:nncf:Attempted to use weight quantizer of 7 BertForQuestionAnswering/BertModel[bert]/BertEmbeddings[embeddings]/NNCFEmbedding[position_embeddings]/embedding_0 to quantize input of {'8 BertForQuestionAnswering/BertModel[bert]/BertEmbeddings[embeddings]/__iadd___0'}, but no compatible configs were found. WARNING:nncf:NNCFNetwork( (nncf_module): BertForQuestionAnswering( (bert): BertModel( (embeddings): BertEmbeddings( (word_embeddings): NNCFEmbedding( 30522, 768, padding_idx=0 (pre_ops): ModuleDict() (post_ops): ModuleDict() ) (position_embeddings): NNCFEmbedding( 512, 768 (pre_ops): ModuleDict() (post_ops): ModuleDict() ) (token_type_embeddings): NNCFEmbedding( 2, 768 (pre_ops): ModuleDict() (post_ops): ModuleDict() ) (LayerNorm): LayerNorm((768,), eps=1e-12, elementwise_affine=True) (dropout): Dropout(p=0.1, inplace=False) ) (encoder): BertEncoder( (layer): ModuleList( (0): BertLayer( (attention): BertAttention( (self): BertSelfAttention( (query): NNCFLinear( in_features=768, out_features=320, bias=True (pre_ops): ModuleDict() (post_ops): ModuleDict() ) (key): NNCFLinear( in_features=768, out_features=320, bias=True (pre_ops): ModuleDict() (post_ops): ModuleDict() ) (value): NNCFLinear( in_features=768, out_features=320, bias=True (pre_ops): ModuleDict() (post_ops): ModuleDict() ) (dropout): Dropout(p=0.1, inplace=False) ) (output): BertSelfOutput( (dense): NNCFLinear( in_features=320, out_features=768, bias=True (pre_ops): ModuleDict() (post_ops): ModuleDict() ) (LayerNorm): LayerNorm((768,), eps=1e-12, elementwise_affine=True) (dropout): Dropout(p=0.1, inplace=False) ) ) (intermediate): BertIntermediate( (dense): NNCFLinear( in_features=768, out_features=185, bias=True (pre_ops): ModuleDict() (post_ops): ModuleDict() ) ) (output): BertOutput( (dense): NNCFLinear( in_features=185, out_features=768, bias=True (pre_ops): ModuleDict() (post_ops): ModuleDict() ) (LayerNorm): LayerNorm((768,), eps=1e-12, elementwise_affine=True) (dropout): Dropout(p=0.1, inplace=False) ) ) (1): BertLayer( (attention): BertAttention( (self): BertSelfAttention( (query): NNCFLinear( in_features=768, out_features=320, bias=True (pre_ops): ModuleDict() (post_ops): ModuleDict() ) (key): NNCFLinear( in_features=768, out_features=320, bias=True (pre_ops): ModuleDict() (post_ops): ModuleDict() ) (value): NNCFLinear( in_features=768, out_features=320, bias=True (pre_ops): ModuleDict() (post_ops): ModuleDict() ) (dropout): Dropout(p=0.1, inplace=False) ) (output): BertSelfOutput( (dense): NNCFLinear( in_features=320, out_features=768, bias=True (pre_ops): ModuleDict() (post_ops): ModuleDict() ) (LayerNorm): LayerNorm((768,), eps=1e-12, elementwise_affine=True) (dropout): Dropout(p=0.1, inplace=False) ) ) (intermediate): BertIntermediate( (dense): NNCFLinear( in_features=768, out_features=315, bias=True (pre_ops): ModuleDict() (post_ops): ModuleDict() ) ) (output): BertOutput( (dense): NNCFLinear( in_features=315, out_features=768, bias=True (pre_ops): ModuleDict() (post_ops): ModuleDict() ) (LayerNorm): LayerNorm((768,), eps=1e-12, elementwise_affine=True) (dropout): Dropout(p=0.1, inplace=False) ) ) (2): BertLayer( (attention): BertAttention( (self): BertSelfAttention( (query): NNCFLinear( in_features=768, out_features=576, bias=True (pre_ops): ModuleDict() (post_ops): ModuleDict() ) (key): NNCFLinear( in_features=768, out_features=576, bias=True (pre_ops): ModuleDict() (post_ops): ModuleDict() ) (value): NNCFLinear( in_features=768, out_features=576, bias=True (pre_ops): ModuleDict() (post_ops): ModuleDict() ) (dropout): Dropout(p=0.1, inplace=False) ) (output): BertSelfOutput( (dense): NNCFLinear( in_features=576, out_features=768, bias=True (pre_ops): ModuleDict() (post_ops): ModuleDict() ) (LayerNorm): LayerNorm((768,), eps=1e-12, elementwise_affine=True) (dropout): Dropout(p=0.1, inplace=False) ) ) (intermediate): BertIntermediate( (dense): NNCFLinear( in_features=768, out_features=339, bias=True (pre_ops): ModuleDict() (post_ops): ModuleDict() ) ) (output): BertOutput( (dense): NNCFLinear( in_features=339, out_features=768, bias=True (pre_ops): ModuleDict() (post_ops): ModuleDict() ) (LayerNorm): LayerNorm((768,), eps=1e-12, elementwise_affine=True) (dropout): Dropout(p=0.1, inplace=False) ) ) (3): BertLayer( (attention): BertAttention( (self): BertSelfAttention( (query): NNCFLinear( in_features=768, out_features=576, bias=True (pre_ops): ModuleDict() (post_ops): ModuleDict() ) (key): NNCFLinear( in_features=768, out_features=576, bias=True (pre_ops): ModuleDict() (post_ops): ModuleDict() ) (value): NNCFLinear( in_features=768, out_features=576, bias=True (pre_ops): ModuleDict() (post_ops): ModuleDict() ) (dropout): Dropout(p=0.1, inplace=False) ) (output): BertSelfOutput( (dense): NNCFLinear( in_features=576, out_features=768, bias=True (pre_ops): ModuleDict() (post_ops): ModuleDict() ) (LayerNorm): LayerNorm((768,), eps=1e-12, elementwise_affine=True) (dropout): Dropout(p=0.1, inplace=False) ) ) (intermediate): BertIntermediate( (dense): NNCFLinear( in_features=768, out_features=368, bias=True (pre_ops): ModuleDict() (post_ops): ModuleDict() ) ) (output): BertOutput( (dense): NNCFLinear( in_features=368, out_features=768, bias=True (pre_ops): ModuleDict() (post_ops): ModuleDict() ) (LayerNorm): LayerNorm((768,), eps=1e-12, elementwise_affine=True) (dropout): Dropout(p=0.1, inplace=False) ) ) (4): BertLayer( (attention): BertAttention( (self): BertSelfAttention( (query): NNCFLinear( in_features=768, out_features=576, bias=True (pre_ops): ModuleDict() (post_ops): ModuleDict() ) (key): NNCFLinear( in_features=768, out_features=576, bias=True (pre_ops): ModuleDict() (post_ops): ModuleDict() ) (value): NNCFLinear( in_features=768, out_features=576, bias=True (pre_ops): ModuleDict() (post_ops): ModuleDict() ) (dropout): Dropout(p=0.1, inplace=False) ) (output): BertSelfOutput( (dense): NNCFLinear( in_features=576, out_features=768, bias=True (pre_ops): ModuleDict() (post_ops): ModuleDict() ) (LayerNorm): LayerNorm((768,), eps=1e-12, elementwise_affine=True) (dropout): Dropout(p=0.1, inplace=False) ) ) (intermediate): BertIntermediate( (dense): NNCFLinear( in_features=768, out_features=386, bias=True (pre_ops): ModuleDict() (post_ops): ModuleDict() ) ) (output): BertOutput( (dense): NNCFLinear( in_features=386, out_features=768, bias=True (pre_ops): ModuleDict() (post_ops): ModuleDict() ) (LayerNorm): LayerNorm((768,), eps=1e-12, elementwise_affine=True) (dropout): Dropout(p=0.1, inplace=False) ) ) (5): BertLayer( (attention): BertAttention( (self): BertSelfAttention( (query): NNCFLinear( in_features=768, out_features=384, bias=True (pre_ops): ModuleDict() (post_ops): ModuleDict() ) (key): NNCFLinear( in_features=768, out_features=384, bias=True (pre_ops): ModuleDict() (post_ops): ModuleDict() ) (value): NNCFLinear( in_features=768, out_features=384, bias=True (pre_ops): ModuleDict() (post_ops): ModuleDict() ) (dropout): Dropout(p=0.1, inplace=False) ) (output): BertSelfOutput( (dense): NNCFLinear( in_features=384, out_features=768, bias=True (pre_ops): ModuleDict() (post_ops): ModuleDict() ) (LayerNorm): LayerNorm((768,), eps=1e-12, elementwise_affine=True) (dropout): Dropout(p=0.1, inplace=False) ) ) (intermediate): BertIntermediate( (dense): NNCFLinear( in_features=768, out_features=336, bias=True (pre_ops): ModuleDict() (post_ops): ModuleDict() ) ) (output): BertOutput( (dense): NNCFLinear( in_features=336, out_features=768, bias=True (pre_ops): ModuleDict() (post_ops): ModuleDict() ) (LayerNorm): LayerNorm((768,), eps=1e-12, elementwise_affine=True) (dropout): Dropout(p=0.1, inplace=False) ) ) (6): BertLayer( (attention): BertAttention( (self): BertSelfAttention( (query): NNCFLinear( in_features=768, out_features=448, bias=True (pre_ops): ModuleDict() (post_ops): ModuleDict() ) (key): NNCFLinear( in_features=768, out_features=448, bias=True (pre_ops): ModuleDict() (post_ops): ModuleDict() ) (value): NNCFLinear( in_features=768, out_features=448, bias=True (pre_ops): ModuleDict() (post_ops): ModuleDict() ) (dropout): Dropout(p=0.1, inplace=False) ) (output): BertSelfOutput( (dense): NNCFLinear( in_features=448, out_features=768, bias=True (pre_ops): ModuleDict() (post_ops): ModuleDict() ) (LayerNorm): LayerNorm((768,), eps=1e-12, elementwise_affine=True) (dropout): Dropout(p=0.1, inplace=False) ) ) (intermediate): BertIntermediate( (dense): NNCFLinear( in_features=768, out_features=280, bias=True (pre_ops): ModuleDict() (post_ops): ModuleDict() ) ) (output): BertOutput( (dense): NNCFLinear( in_features=280, out_features=768, bias=True (pre_ops): ModuleDict() (post_ops): ModuleDict() ) (LayerNorm): LayerNorm((768,), eps=1e-12, elementwise_affine=True) (dropout): Dropout(p=0.1, inplace=False) ) ) (7): BertLayer( (attention): BertAttention( (self): BertSelfAttention( (query): NNCFLinear( in_features=768, out_features=448, bias=True (pre_ops): ModuleDict() (post_ops): ModuleDict() ) (key): NNCFLinear( in_features=768, out_features=448, bias=True (pre_ops): ModuleDict() (post_ops): ModuleDict() ) (value): NNCFLinear( in_features=768, out_features=448, bias=True (pre_ops): ModuleDict() (post_ops): ModuleDict() ) (dropout): Dropout(p=0.1, inplace=False) ) (output): BertSelfOutput( (dense): NNCFLinear( in_features=448, out_features=768, bias=True (pre_ops): ModuleDict() (post_ops): ModuleDict() ) (LayerNorm): LayerNorm((768,), eps=1e-12, elementwise_affine=True) (dropout): Dropout(p=0.1, inplace=False) ) ) (intermediate): BertIntermediate( (dense): NNCFLinear( in_features=768, out_features=211, bias=True (pre_ops): ModuleDict() (post_ops): ModuleDict() ) ) (output): BertOutput( (dense): NNCFLinear( in_features=211, out_features=768, bias=True (pre_ops): ModuleDict() (post_ops): ModuleDict() ) (LayerNorm): LayerNorm((768,), eps=1e-12, elementwise_affine=True) (dropout): Dropout(p=0.1, inplace=False) ) ) (8): BertLayer( (attention): BertAttention( (self): BertSelfAttention( (query): NNCFLinear( in_features=768, out_features=448, bias=True (pre_ops): ModuleDict() (post_ops): ModuleDict() ) (key): NNCFLinear( in_features=768, out_features=448, bias=True (pre_ops): ModuleDict() (post_ops): ModuleDict() ) (value): NNCFLinear( in_features=768, out_features=448, bias=True (pre_ops): ModuleDict() (post_ops): ModuleDict() ) (dropout): Dropout(p=0.1, inplace=False) ) (output): BertSelfOutput( (dense): NNCFLinear( in_features=448, out_features=768, bias=True (pre_ops): ModuleDict() (post_ops): ModuleDict() ) (LayerNorm): LayerNorm((768,), eps=1e-12, elementwise_affine=True) (dropout): Dropout(p=0.1, inplace=False) ) ) (intermediate): BertIntermediate( (dense): NNCFLinear( in_features=768, out_features=108, bias=True (pre_ops): ModuleDict() (post_ops): ModuleDict() ) ) (output): BertOutput( (dense): NNCFLinear( in_features=108, out_features=768, bias=True (pre_ops): ModuleDict() (post_ops): ModuleDict() ) (LayerNorm): LayerNorm((768,), eps=1e-12, elementwise_affine=True) (dropout): Dropout(p=0.1, inplace=False) ) ) (9): BertLayer( (attention): BertAttention( (self): BertSelfAttention( (query): NNCFLinear( in_features=768, out_features=320, bias=True (pre_ops): ModuleDict() (post_ops): ModuleDict() ) (key): NNCFLinear( in_features=768, out_features=320, bias=True (pre_ops): ModuleDict() (post_ops): ModuleDict() ) (value): NNCFLinear( in_features=768, out_features=320, bias=True (pre_ops): ModuleDict() (post_ops): ModuleDict() ) (dropout): Dropout(p=0.1, inplace=False) ) (output): BertSelfOutput( (dense): NNCFLinear( in_features=320, out_features=768, bias=True (pre_ops): ModuleDict() (post_ops): ModuleDict() ) (LayerNorm): LayerNorm((768,), eps=1e-12, elementwise_affine=True) (dropout): Dropout(p=0.1, inplace=False) ) ) (intermediate): BertIntermediate( (dense): NNCFLinear( in_features=768, out_features=53, bias=True (pre_ops): ModuleDict() (post_ops): ModuleDict() ) ) (output): BertOutput( (dense): NNCFLinear( in_features=53, out_features=768, bias=True (pre_ops): ModuleDict() (post_ops): ModuleDict() ) (LayerNorm): LayerNorm((768,), eps=1e-12, elementwise_affine=True) (dropout): Dropout(p=0.1, inplace=False) ) ) (10): BertLayer( (attention): BertAttention( (self): BertSelfAttention( (query): NNCFLinear( in_features=768, out_features=384, bias=True (pre_ops): ModuleDict() (post_ops): ModuleDict() ) (key): NNCFLinear( in_features=768, out_features=384, bias=True (pre_ops): ModuleDict() (post_ops): ModuleDict() ) (value): NNCFLinear( in_features=768, out_features=384, bias=True (pre_ops): ModuleDict() (post_ops): ModuleDict() ) (dropout): Dropout(p=0.1, inplace=False) ) (output): BertSelfOutput( (dense): NNCFLinear( in_features=384, out_features=768, bias=True (pre_ops): ModuleDict() (post_ops): ModuleDict() ) (LayerNorm): LayerNorm((768,), eps=1e-12, elementwise_affine=True) (dropout): Dropout(p=0.1, inplace=False) ) ) (intermediate): BertIntermediate( (dense): NNCFLinear( in_features=768, out_features=86, bias=True (pre_ops): ModuleDict() (post_ops): ModuleDict() ) ) (output): BertOutput( (dense): NNCFLinear( in_features=86, out_features=768, bias=True (pre_ops): ModuleDict() (post_ops): ModuleDict() ) (LayerNorm): LayerNorm((768,), eps=1e-12, elementwise_affine=True) (dropout): Dropout(p=0.1, inplace=False) ) ) (11): BertLayer( (attention): BertAttention( (self): BertSelfAttention( (query): NNCFLinear( in_features=768, out_features=384, bias=True (pre_ops): ModuleDict() (post_ops): ModuleDict() ) (key): NNCFLinear( in_features=768, out_features=384, bias=True (pre_ops): ModuleDict() (post_ops): ModuleDict() ) (value): NNCFLinear( in_features=768, out_features=384, bias=True (pre_ops): ModuleDict() (post_ops): ModuleDict() ) (dropout): Dropout(p=0.1, inplace=False) ) (output): BertSelfOutput( (dense): NNCFLinear( in_features=384, out_features=768, bias=True (pre_ops): ModuleDict() (post_ops): ModuleDict() ) (LayerNorm): LayerNorm((768,), eps=1e-12, elementwise_affine=True) (dropout): Dropout(p=0.1, inplace=False) ) ) (intermediate): BertIntermediate( (dense): NNCFLinear( in_features=768, out_features=105, bias=True (pre_ops): ModuleDict() (post_ops): ModuleDict() ) ) (output): BertOutput( (dense): NNCFLinear( in_features=105, out_features=768, bias=True (pre_ops): ModuleDict() (post_ops): ModuleDict() ) (LayerNorm): LayerNorm((768,), eps=1e-12, elementwise_affine=True) (dropout): Dropout(p=0.1, inplace=False) ) ) ) ) ) (qa_outputs): NNCFLinear( in_features=768, out_features=2, bias=True (pre_ops): ModuleDict() (post_ops): ModuleDict() ) ) ) INFO:nncf:Collecting tensor statistics ████████ | 1 / 2 INFO:nncf:Collecting tensor statistics ████████████████ | 2 / 2 INFO:nncf:Set sign: True and scale: [0.3102, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEmbeddings[embeddings]/NNCFEmbedding[word_embeddings]/embedding_0 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEmbeddings[embeddings]/NNCFEmbedding[word_embeddings]/embedding_0 INFO:nncf:Set sign: True and scale: [0.6038, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEmbeddings[embeddings]/NNCFEmbedding[token_type_embeddings]/embedding_0 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEmbeddings[embeddings]/NNCFEmbedding[token_type_embeddings]/embedding_0 INFO:nncf:Set sign: True and scale: [0.7471, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEmbeddings[embeddings]/__add___0 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEmbeddings[embeddings]/__add___0 INFO:nncf:Set sign: True and scale: [0.1617, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEmbeddings[embeddings]/NNCFEmbedding[position_embeddings]/embedding_0 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEmbeddings[embeddings]/NNCFEmbedding[position_embeddings]/embedding_0 INFO:nncf:Set sign: True and scale: [0.7471, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEmbeddings[embeddings]/__iadd___0 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEmbeddings[embeddings]/__iadd___0 INFO:nncf:Set sign: True and scale: [6.4642, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[0]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[query]/linear_0 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[0]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[query]/linear_0 INFO:nncf:Set sign: True and scale: [6.3328, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[0]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[key]/linear_0 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[0]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[key]/linear_0 INFO:nncf:Set sign: True and scale: [13.2152, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[0]/BertAttention[attention]/BertSelfAttention[self]/__truediv___0 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[0]/BertAttention[attention]/BertSelfAttention[self]/__truediv___0 INFO:nncf:Set sign: False and scale: [0.9097, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[0]/BertAttention[attention]/BertSelfAttention[self]/Softmax/softmax_0 INFO:nncf:Performing unsigned activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[0]/BertAttention[attention]/BertSelfAttention[self]/Softmax/softmax_0 INFO:nncf:Set sign: True and scale: [2.6530, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[0]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[value]/linear_0 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[0]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[value]/linear_0 INFO:nncf:Set sign: True and scale: [1.5105, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[0]/BertAttention[attention]/BertSelfAttention[self]/matmul_1 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[0]/BertAttention[attention]/BertSelfAttention[self]/matmul_1 INFO:nncf:Set sign: True and scale: [1.0678, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[0]/BertAttention[attention]/BertSelfOutput[output]/NNCFLinear[dense]/linear_0 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[0]/BertAttention[attention]/BertSelfOutput[output]/NNCFLinear[dense]/linear_0 INFO:nncf:Set sign: True and scale: [4.1636, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[0]/BertAttention[attention]/BertSelfOutput[output]/__add___0 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[0]/BertAttention[attention]/BertSelfOutput[output]/__add___0 INFO:nncf:Set sign: True and scale: [7.5213, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[0]/BertIntermediate[intermediate]/gelu_0 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[0]/BertIntermediate[intermediate]/gelu_0 INFO:nncf:Set sign: True and scale: [7.2468, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[0]/BertOutput[output]/NNCFLinear[dense]/linear_0 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[0]/BertOutput[output]/NNCFLinear[dense]/linear_0 INFO:nncf:Set sign: True and scale: [26.0014, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[0]/BertOutput[output]/__add___0 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[0]/BertOutput[output]/__add___0 INFO:nncf:Set sign: True and scale: [7.3508, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[1]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[query]/linear_0 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[1]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[query]/linear_0 INFO:nncf:Set sign: True and scale: [6.6369, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[1]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[key]/linear_0 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[1]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[key]/linear_0 INFO:nncf:Set sign: True and scale: [22.0350, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[1]/BertAttention[attention]/BertSelfAttention[self]/__truediv___0 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[1]/BertAttention[attention]/BertSelfAttention[self]/__truediv___0 INFO:nncf:Set sign: False and scale: [0.9929, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[1]/BertAttention[attention]/BertSelfAttention[self]/Softmax/softmax_0 INFO:nncf:Performing unsigned activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[1]/BertAttention[attention]/BertSelfAttention[self]/Softmax/softmax_0 INFO:nncf:Set sign: True and scale: [4.6145, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[1]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[value]/linear_0 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[1]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[value]/linear_0 INFO:nncf:Set sign: True and scale: [2.3576, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[1]/BertAttention[attention]/BertSelfAttention[self]/matmul_1 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[1]/BertAttention[attention]/BertSelfAttention[self]/matmul_1 INFO:nncf:Set sign: True and scale: [1.0576, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[1]/BertAttention[attention]/BertSelfOutput[output]/NNCFLinear[dense]/linear_0 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[1]/BertAttention[attention]/BertSelfOutput[output]/NNCFLinear[dense]/linear_0 INFO:nncf:Set sign: True and scale: [6.1697, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[1]/BertAttention[attention]/BertSelfOutput[output]/__add___0 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[1]/BertAttention[attention]/BertSelfOutput[output]/__add___0 INFO:nncf:Set sign: True and scale: [8.2951, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[1]/BertIntermediate[intermediate]/gelu_0 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[1]/BertIntermediate[intermediate]/gelu_0 INFO:nncf:Set sign: True and scale: [2.3020, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[1]/BertOutput[output]/NNCFLinear[dense]/linear_0 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[1]/BertOutput[output]/NNCFLinear[dense]/linear_0 INFO:nncf:Set sign: True and scale: [32.6324, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[1]/BertOutput[output]/__add___0 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[1]/BertOutput[output]/__add___0 INFO:nncf:Set sign: True and scale: [9.5079, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[2]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[query]/linear_0 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[2]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[query]/linear_0 INFO:nncf:Set sign: True and scale: [8.5606, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[2]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[key]/linear_0 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[2]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[key]/linear_0 INFO:nncf:Set sign: True and scale: [43.0158, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[2]/BertAttention[attention]/BertSelfAttention[self]/__truediv___0 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[2]/BertAttention[attention]/BertSelfAttention[self]/__truediv___0 INFO:nncf:Set sign: False and scale: [0.9999, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[2]/BertAttention[attention]/BertSelfAttention[self]/Softmax/softmax_0 INFO:nncf:Performing unsigned activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[2]/BertAttention[attention]/BertSelfAttention[self]/Softmax/softmax_0 INFO:nncf:Set sign: True and scale: [4.9327, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[2]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[value]/linear_0 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[2]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[value]/linear_0 INFO:nncf:Set sign: True and scale: [2.1434, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[2]/BertAttention[attention]/BertSelfAttention[self]/matmul_1 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[2]/BertAttention[attention]/BertSelfAttention[self]/matmul_1 INFO:nncf:Set sign: True and scale: [1.1261, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[2]/BertAttention[attention]/BertSelfOutput[output]/NNCFLinear[dense]/linear_0 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[2]/BertAttention[attention]/BertSelfOutput[output]/NNCFLinear[dense]/linear_0 INFO:nncf:Set sign: True and scale: [7.7405, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[2]/BertAttention[attention]/BertSelfOutput[output]/__add___0 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[2]/BertAttention[attention]/BertSelfOutput[output]/__add___0 INFO:nncf:Set sign: True and scale: [5.6863, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[2]/BertIntermediate[intermediate]/gelu_0 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[2]/BertIntermediate[intermediate]/gelu_0 INFO:nncf:Set sign: True and scale: [1.7192, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[2]/BertOutput[output]/NNCFLinear[dense]/linear_0 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[2]/BertOutput[output]/NNCFLinear[dense]/linear_0 INFO:nncf:Set sign: True and scale: [27.8657, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[2]/BertOutput[output]/__add___0 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[2]/BertOutput[output]/__add___0 INFO:nncf:Set sign: True and scale: [6.3400, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[3]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[query]/linear_0 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[3]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[query]/linear_0 INFO:nncf:Set sign: True and scale: [6.3571, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[3]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[key]/linear_0 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[3]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[key]/linear_0 INFO:nncf:Set sign: True and scale: [16.8984, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[3]/BertAttention[attention]/BertSelfAttention[self]/__truediv___0 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[3]/BertAttention[attention]/BertSelfAttention[self]/__truediv___0 INFO:nncf:Set sign: False and scale: [0.9659, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[3]/BertAttention[attention]/BertSelfAttention[self]/Softmax/softmax_0 INFO:nncf:Performing unsigned activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[3]/BertAttention[attention]/BertSelfAttention[self]/Softmax/softmax_0 INFO:nncf:Set sign: True and scale: [3.5732, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[3]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[value]/linear_0 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[3]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[value]/linear_0 INFO:nncf:Set sign: True and scale: [2.4142, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[3]/BertAttention[attention]/BertSelfAttention[self]/matmul_1 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[3]/BertAttention[attention]/BertSelfAttention[self]/matmul_1 INFO:nncf:Set sign: True and scale: [1.4166, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[3]/BertAttention[attention]/BertSelfOutput[output]/NNCFLinear[dense]/linear_0 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[3]/BertAttention[attention]/BertSelfOutput[output]/NNCFLinear[dense]/linear_0 INFO:nncf:Set sign: True and scale: [7.1450, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[3]/BertAttention[attention]/BertSelfOutput[output]/__add___0 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[3]/BertAttention[attention]/BertSelfOutput[output]/__add___0 INFO:nncf:Set sign: True and scale: [6.4436, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[3]/BertIntermediate[intermediate]/gelu_0 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[3]/BertIntermediate[intermediate]/gelu_0 INFO:nncf:Set sign: True and scale: [1.7720, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[3]/BertOutput[output]/NNCFLinear[dense]/linear_0 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[3]/BertOutput[output]/NNCFLinear[dense]/linear_0 INFO:nncf:Set sign: True and scale: [18.1251, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[3]/BertOutput[output]/__add___0 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[3]/BertOutput[output]/__add___0 INFO:nncf:Set sign: True and scale: [5.6173, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[4]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[query]/linear_0 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[4]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[query]/linear_0 INFO:nncf:Set sign: True and scale: [6.2416, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[4]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[key]/linear_0 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[4]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[key]/linear_0 INFO:nncf:Set sign: True and scale: [12.2383, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[4]/BertAttention[attention]/BertSelfAttention[self]/__truediv___0 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[4]/BertAttention[attention]/BertSelfAttention[self]/__truediv___0 INFO:nncf:Set sign: False and scale: [0.5249, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[4]/BertAttention[attention]/BertSelfAttention[self]/Softmax/softmax_0 INFO:nncf:Performing unsigned activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[4]/BertAttention[attention]/BertSelfAttention[self]/Softmax/softmax_0 INFO:nncf:Set sign: True and scale: [3.2927, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[4]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[value]/linear_0 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[4]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[value]/linear_0 INFO:nncf:Set sign: True and scale: [2.2221, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[4]/BertAttention[attention]/BertSelfAttention[self]/matmul_1 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[4]/BertAttention[attention]/BertSelfAttention[self]/matmul_1 INFO:nncf:Set sign: True and scale: [1.3042, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[4]/BertAttention[attention]/BertSelfOutput[output]/NNCFLinear[dense]/linear_0 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[4]/BertAttention[attention]/BertSelfOutput[output]/NNCFLinear[dense]/linear_0 INFO:nncf:Set sign: True and scale: [7.5106, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[4]/BertAttention[attention]/BertSelfOutput[output]/__add___0 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[4]/BertAttention[attention]/BertSelfOutput[output]/__add___0 INFO:nncf:Set sign: True and scale: [5.8193, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[4]/BertIntermediate[intermediate]/gelu_0 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[4]/BertIntermediate[intermediate]/gelu_0 INFO:nncf:Set sign: True and scale: [1.5960, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[4]/BertOutput[output]/NNCFLinear[dense]/linear_0 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[4]/BertOutput[output]/NNCFLinear[dense]/linear_0 INFO:nncf:Set sign: True and scale: [16.6634, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[4]/BertOutput[output]/__add___0 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[4]/BertOutput[output]/__add___0 INFO:nncf:Set sign: True and scale: [5.8104, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[5]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[query]/linear_0 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[5]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[query]/linear_0 INFO:nncf:Set sign: True and scale: [7.1760, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[5]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[key]/linear_0 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[5]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[key]/linear_0 INFO:nncf:Set sign: True and scale: [19.2534, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[5]/BertAttention[attention]/BertSelfAttention[self]/__truediv___0 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[5]/BertAttention[attention]/BertSelfAttention[self]/__truediv___0 INFO:nncf:Set sign: False and scale: [0.9088, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[5]/BertAttention[attention]/BertSelfAttention[self]/Softmax/softmax_0 INFO:nncf:Performing unsigned activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[5]/BertAttention[attention]/BertSelfAttention[self]/Softmax/softmax_0 INFO:nncf:Set sign: True and scale: [3.2773, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[5]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[value]/linear_0 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[5]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[value]/linear_0 INFO:nncf:Set sign: True and scale: [2.3789, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[5]/BertAttention[attention]/BertSelfAttention[self]/matmul_1 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[5]/BertAttention[attention]/BertSelfAttention[self]/matmul_1 INFO:nncf:Set sign: True and scale: [1.3509, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[5]/BertAttention[attention]/BertSelfOutput[output]/NNCFLinear[dense]/linear_0 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[5]/BertAttention[attention]/BertSelfOutput[output]/NNCFLinear[dense]/linear_0 INFO:nncf:Set sign: True and scale: [8.0278, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[5]/BertAttention[attention]/BertSelfOutput[output]/__add___0 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[5]/BertAttention[attention]/BertSelfOutput[output]/__add___0 INFO:nncf:Set sign: True and scale: [11.2281, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[5]/BertIntermediate[intermediate]/gelu_0 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[5]/BertIntermediate[intermediate]/gelu_0 INFO:nncf:Set sign: True and scale: [2.5510, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[5]/BertOutput[output]/NNCFLinear[dense]/linear_0 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[5]/BertOutput[output]/NNCFLinear[dense]/linear_0 INFO:nncf:Set sign: True and scale: [13.3404, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[5]/BertOutput[output]/__add___0 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[5]/BertOutput[output]/__add___0 INFO:nncf:Set sign: True and scale: [5.9837, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[6]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[query]/linear_0 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[6]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[query]/linear_0 INFO:nncf:Set sign: True and scale: [7.7394, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[6]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[key]/linear_0 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[6]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[key]/linear_0 INFO:nncf:Set sign: True and scale: [16.1081, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[6]/BertAttention[attention]/BertSelfAttention[self]/__truediv___0 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[6]/BertAttention[attention]/BertSelfAttention[self]/__truediv___0 INFO:nncf:Set sign: False and scale: [0.8820, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[6]/BertAttention[attention]/BertSelfAttention[self]/Softmax/softmax_0 INFO:nncf:Performing unsigned activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[6]/BertAttention[attention]/BertSelfAttention[self]/Softmax/softmax_0 INFO:nncf:Set sign: True and scale: [2.8653, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[6]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[value]/linear_0 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[6]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[value]/linear_0 INFO:nncf:Set sign: True and scale: [2.0227, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[6]/BertAttention[attention]/BertSelfAttention[self]/matmul_1 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[6]/BertAttention[attention]/BertSelfAttention[self]/matmul_1 INFO:nncf:Set sign: True and scale: [1.1265, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[6]/BertAttention[attention]/BertSelfOutput[output]/NNCFLinear[dense]/linear_0 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[6]/BertAttention[attention]/BertSelfOutput[output]/NNCFLinear[dense]/linear_0 INFO:nncf:Set sign: True and scale: [7.6260, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[6]/BertAttention[attention]/BertSelfOutput[output]/__add___0 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[6]/BertAttention[attention]/BertSelfOutput[output]/__add___0 INFO:nncf:Set sign: True and scale: [10.1383, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[6]/BertIntermediate[intermediate]/gelu_0 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[6]/BertIntermediate[intermediate]/gelu_0 INFO:nncf:Set sign: True and scale: [2.0738, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[6]/BertOutput[output]/NNCFLinear[dense]/linear_0 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[6]/BertOutput[output]/NNCFLinear[dense]/linear_0 INFO:nncf:Set sign: True and scale: [12.1558, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[6]/BertOutput[output]/__add___0 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[6]/BertOutput[output]/__add___0 INFO:nncf:Set sign: True and scale: [6.2621, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[7]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[query]/linear_0 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[7]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[query]/linear_0 INFO:nncf:Set sign: True and scale: [7.9600, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[7]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[key]/linear_0 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[7]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[key]/linear_0 INFO:nncf:Set sign: True and scale: [15.8902, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[7]/BertAttention[attention]/BertSelfAttention[self]/__truediv___0 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[7]/BertAttention[attention]/BertSelfAttention[self]/__truediv___0 INFO:nncf:Set sign: False and scale: [0.8493, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[7]/BertAttention[attention]/BertSelfAttention[self]/Softmax/softmax_0 INFO:nncf:Performing unsigned activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[7]/BertAttention[attention]/BertSelfAttention[self]/Softmax/softmax_0 INFO:nncf:Set sign: True and scale: [2.7640, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[7]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[value]/linear_0 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[7]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[value]/linear_0 INFO:nncf:Set sign: True and scale: [2.1766, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[7]/BertAttention[attention]/BertSelfAttention[self]/matmul_1 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[7]/BertAttention[attention]/BertSelfAttention[self]/matmul_1 INFO:nncf:Set sign: True and scale: [1.7565, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[7]/BertAttention[attention]/BertSelfOutput[output]/NNCFLinear[dense]/linear_0 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[7]/BertAttention[attention]/BertSelfOutput[output]/NNCFLinear[dense]/linear_0 INFO:nncf:Set sign: True and scale: [6.6154, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[7]/BertAttention[attention]/BertSelfOutput[output]/__add___0 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[7]/BertAttention[attention]/BertSelfOutput[output]/__add___0 INFO:nncf:Set sign: True and scale: [12.0720, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[7]/BertIntermediate[intermediate]/gelu_0 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[7]/BertIntermediate[intermediate]/gelu_0 INFO:nncf:Set sign: True and scale: [2.7441, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[7]/BertOutput[output]/NNCFLinear[dense]/linear_0 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[7]/BertOutput[output]/NNCFLinear[dense]/linear_0 INFO:nncf:Set sign: True and scale: [12.5571, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[7]/BertOutput[output]/__add___0 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[7]/BertOutput[output]/__add___0 INFO:nncf:Set sign: True and scale: [6.7775, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[8]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[query]/linear_0 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[8]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[query]/linear_0 INFO:nncf:Set sign: True and scale: [7.6945, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[8]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[key]/linear_0 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[8]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[key]/linear_0 INFO:nncf:Set sign: True and scale: [15.1916, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[8]/BertAttention[attention]/BertSelfAttention[self]/__truediv___0 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[8]/BertAttention[attention]/BertSelfAttention[self]/__truediv___0 INFO:nncf:Set sign: False and scale: [0.4918, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[8]/BertAttention[attention]/BertSelfAttention[self]/Softmax/softmax_0 INFO:nncf:Performing unsigned activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[8]/BertAttention[attention]/BertSelfAttention[self]/Softmax/softmax_0 INFO:nncf:Set sign: True and scale: [3.3290, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[8]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[value]/linear_0 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[8]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[value]/linear_0 INFO:nncf:Set sign: True and scale: [2.5413, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[8]/BertAttention[attention]/BertSelfAttention[self]/matmul_1 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[8]/BertAttention[attention]/BertSelfAttention[self]/matmul_1 INFO:nncf:Set sign: True and scale: [1.2726, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[8]/BertAttention[attention]/BertSelfOutput[output]/NNCFLinear[dense]/linear_0 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[8]/BertAttention[attention]/BertSelfOutput[output]/NNCFLinear[dense]/linear_0 INFO:nncf:Set sign: True and scale: [5.8699, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[8]/BertAttention[attention]/BertSelfOutput[output]/__add___0 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[8]/BertAttention[attention]/BertSelfOutput[output]/__add___0 INFO:nncf:Set sign: True and scale: [9.0108, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[8]/BertIntermediate[intermediate]/gelu_0 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[8]/BertIntermediate[intermediate]/gelu_0 INFO:nncf:Set sign: True and scale: [3.2770, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[8]/BertOutput[output]/NNCFLinear[dense]/linear_0 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[8]/BertOutput[output]/NNCFLinear[dense]/linear_0 INFO:nncf:Set sign: True and scale: [16.1354, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[8]/BertOutput[output]/__add___0 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[8]/BertOutput[output]/__add___0 INFO:nncf:Set sign: True and scale: [5.5691, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[9]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[query]/linear_0 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[9]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[query]/linear_0 INFO:nncf:Set sign: True and scale: [8.0957, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[9]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[key]/linear_0 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[9]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[key]/linear_0 INFO:nncf:Set sign: True and scale: [12.0098, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[9]/BertAttention[attention]/BertSelfAttention[self]/__truediv___0 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[9]/BertAttention[attention]/BertSelfAttention[self]/__truediv___0 INFO:nncf:Set sign: False and scale: [0.4007, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[9]/BertAttention[attention]/BertSelfAttention[self]/Softmax/softmax_0 INFO:nncf:Performing unsigned activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[9]/BertAttention[attention]/BertSelfAttention[self]/Softmax/softmax_0 INFO:nncf:Set sign: True and scale: [2.9702, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[9]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[value]/linear_0 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[9]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[value]/linear_0 INFO:nncf:Set sign: True and scale: [2.1663, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[9]/BertAttention[attention]/BertSelfAttention[self]/matmul_1 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[9]/BertAttention[attention]/BertSelfAttention[self]/matmul_1 INFO:nncf:Set sign: True and scale: [1.2682, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[9]/BertAttention[attention]/BertSelfOutput[output]/NNCFLinear[dense]/linear_0 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[9]/BertAttention[attention]/BertSelfOutput[output]/NNCFLinear[dense]/linear_0 INFO:nncf:Set sign: True and scale: [5.5587, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[9]/BertAttention[attention]/BertSelfOutput[output]/__add___0 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[9]/BertAttention[attention]/BertSelfOutput[output]/__add___0 INFO:nncf:Set sign: True and scale: [13.8907, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[9]/BertIntermediate[intermediate]/gelu_0 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[9]/BertIntermediate[intermediate]/gelu_0 INFO:nncf:Set sign: True and scale: [2.9744, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[9]/BertOutput[output]/NNCFLinear[dense]/linear_0 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[9]/BertOutput[output]/NNCFLinear[dense]/linear_0 INFO:nncf:Set sign: True and scale: [11.1289, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[9]/BertOutput[output]/__add___0 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[9]/BertOutput[output]/__add___0 INFO:nncf:Set sign: True and scale: [4.5430, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[10]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[query]/linear_0 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[10]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[query]/linear_0 INFO:nncf:Set sign: True and scale: [7.9071, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[10]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[key]/linear_0 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[10]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[key]/linear_0 INFO:nncf:Set sign: True and scale: [11.5003, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[10]/BertAttention[attention]/BertSelfAttention[self]/__truediv___0 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[10]/BertAttention[attention]/BertSelfAttention[self]/__truediv___0 INFO:nncf:Set sign: False and scale: [0.5586, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[10]/BertAttention[attention]/BertSelfAttention[self]/Softmax/softmax_0 INFO:nncf:Performing unsigned activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[10]/BertAttention[attention]/BertSelfAttention[self]/Softmax/softmax_0 INFO:nncf:Set sign: True and scale: [6.2925, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[10]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[value]/linear_0 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[10]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[value]/linear_0 INFO:nncf:Set sign: True and scale: [5.2843, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[10]/BertAttention[attention]/BertSelfAttention[self]/matmul_1 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[10]/BertAttention[attention]/BertSelfAttention[self]/matmul_1 INFO:nncf:Set sign: True and scale: [1.9610, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[10]/BertAttention[attention]/BertSelfOutput[output]/NNCFLinear[dense]/linear_0 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[10]/BertAttention[attention]/BertSelfOutput[output]/NNCFLinear[dense]/linear_0 INFO:nncf:Set sign: True and scale: [6.5934, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[10]/BertAttention[attention]/BertSelfOutput[output]/__add___0 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[10]/BertAttention[attention]/BertSelfOutput[output]/__add___0 INFO:nncf:Set sign: True and scale: [9.1569, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[10]/BertIntermediate[intermediate]/gelu_0 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[10]/BertIntermediate[intermediate]/gelu_0 INFO:nncf:Set sign: True and scale: [3.3628, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[10]/BertOutput[output]/NNCFLinear[dense]/linear_0 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[10]/BertOutput[output]/NNCFLinear[dense]/linear_0 INFO:nncf:Set sign: True and scale: [9.9668, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[10]/BertOutput[output]/__add___0 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[10]/BertOutput[output]/__add___0 INFO:nncf:Set sign: True and scale: [4.3727, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[11]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[query]/linear_0 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[11]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[query]/linear_0 INFO:nncf:Set sign: True and scale: [6.3919, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[11]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[key]/linear_0 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[11]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[key]/linear_0 INFO:nncf:Set sign: True and scale: [11.2429, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[11]/BertAttention[attention]/BertSelfAttention[self]/__truediv___0 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[11]/BertAttention[attention]/BertSelfAttention[self]/__truediv___0 INFO:nncf:Set sign: False and scale: [0.5345, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[11]/BertAttention[attention]/BertSelfAttention[self]/Softmax/softmax_0 INFO:nncf:Performing unsigned activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[11]/BertAttention[attention]/BertSelfAttention[self]/Softmax/softmax_0 INFO:nncf:Set sign: True and scale: [4.5022, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[11]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[value]/linear_0 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[11]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[value]/linear_0 INFO:nncf:Set sign: True and scale: [3.6806, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[11]/BertAttention[attention]/BertSelfAttention[self]/matmul_1 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[11]/BertAttention[attention]/BertSelfAttention[self]/matmul_1 INFO:nncf:Set sign: True and scale: [1.8066, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[11]/BertAttention[attention]/BertSelfOutput[output]/NNCFLinear[dense]/linear_0 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[11]/BertAttention[attention]/BertSelfOutput[output]/NNCFLinear[dense]/linear_0 INFO:nncf:Set sign: True and scale: [7.7527, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[11]/BertAttention[attention]/BertSelfOutput[output]/__add___0 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[11]/BertAttention[attention]/BertSelfOutput[output]/__add___0 INFO:nncf:Set sign: True and scale: [7.1442, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[11]/BertIntermediate[intermediate]/gelu_0 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[11]/BertIntermediate[intermediate]/gelu_0 INFO:nncf:Set sign: True and scale: [3.5706, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[11]/BertOutput[output]/NNCFLinear[dense]/linear_0 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[11]/BertOutput[output]/NNCFLinear[dense]/linear_0 INFO:nncf:Set sign: True and scale: [14.4304, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[11]/BertOutput[output]/__add___0 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[11]/BertOutput[output]/__add___0 INFO:nncf:Set sign: True and scale: [2.8359, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[11]/BertOutput[output]/LayerNorm[LayerNorm]/layer_norm_0 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[11]/BertOutput[output]/LayerNorm[LayerNorm]/layer_norm_0 INFO:nncf:Set sign: True and scale: [4.1685, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEmbeddings[embeddings]/LayerNorm[LayerNorm]/layer_norm_0 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEmbeddings[embeddings]/LayerNorm[LayerNorm]/layer_norm_0 INFO:nncf:Set sign: True and scale: [26.1356, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[0]/BertAttention[attention]/BertSelfOutput[output]/LayerNorm[LayerNorm]/layer_norm_0 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[0]/BertAttention[attention]/BertSelfOutput[output]/LayerNorm[LayerNorm]/layer_norm_0 INFO:nncf:Set sign: True and scale: [6.5463, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[0]/BertOutput[output]/LayerNorm[LayerNorm]/layer_norm_0 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[0]/BertOutput[output]/LayerNorm[LayerNorm]/layer_norm_0 INFO:nncf:Set sign: True and scale: [31.8059, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[1]/BertAttention[attention]/BertSelfOutput[output]/LayerNorm[LayerNorm]/layer_norm_0 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[1]/BertAttention[attention]/BertSelfOutput[output]/LayerNorm[LayerNorm]/layer_norm_0 INFO:nncf:Set sign: True and scale: [7.8890, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[1]/BertOutput[output]/LayerNorm[LayerNorm]/layer_norm_0 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[1]/BertOutput[output]/LayerNorm[LayerNorm]/layer_norm_0 INFO:nncf:Set sign: True and scale: [26.9679, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[2]/BertAttention[attention]/BertSelfOutput[output]/LayerNorm[LayerNorm]/layer_norm_0 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[2]/BertAttention[attention]/BertSelfOutput[output]/LayerNorm[LayerNorm]/layer_norm_0 INFO:nncf:Set sign: True and scale: [7.5738, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[2]/BertOutput[output]/LayerNorm[LayerNorm]/layer_norm_0 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[2]/BertOutput[output]/LayerNorm[LayerNorm]/layer_norm_0 INFO:nncf:Set sign: True and scale: [17.8367, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[3]/BertAttention[attention]/BertSelfOutput[output]/LayerNorm[LayerNorm]/layer_norm_0 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[3]/BertAttention[attention]/BertSelfOutput[output]/LayerNorm[LayerNorm]/layer_norm_0 INFO:nncf:Set sign: True and scale: [7.5737, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[3]/BertOutput[output]/LayerNorm[LayerNorm]/layer_norm_0 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[3]/BertOutput[output]/LayerNorm[LayerNorm]/layer_norm_0 INFO:nncf:Set sign: True and scale: [15.8130, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[4]/BertAttention[attention]/BertSelfOutput[output]/LayerNorm[LayerNorm]/layer_norm_0 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[4]/BertAttention[attention]/BertSelfOutput[output]/LayerNorm[LayerNorm]/layer_norm_0 INFO:nncf:Set sign: True and scale: [7.9781, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[4]/BertOutput[output]/LayerNorm[LayerNorm]/layer_norm_0 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[4]/BertOutput[output]/LayerNorm[LayerNorm]/layer_norm_0 INFO:nncf:Set sign: True and scale: [12.5568, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[5]/BertAttention[attention]/BertSelfOutput[output]/LayerNorm[LayerNorm]/layer_norm_0 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[5]/BertAttention[attention]/BertSelfOutput[output]/LayerNorm[LayerNorm]/layer_norm_0 INFO:nncf:Set sign: True and scale: [7.7418, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[5]/BertOutput[output]/LayerNorm[LayerNorm]/layer_norm_0 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[5]/BertOutput[output]/LayerNorm[LayerNorm]/layer_norm_0 INFO:nncf:Set sign: True and scale: [10.9951, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[6]/BertAttention[attention]/BertSelfOutput[output]/LayerNorm[LayerNorm]/layer_norm_0 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[6]/BertAttention[attention]/BertSelfOutput[output]/LayerNorm[LayerNorm]/layer_norm_0 INFO:nncf:Set sign: True and scale: [6.9427, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[6]/BertOutput[output]/LayerNorm[LayerNorm]/layer_norm_0 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[6]/BertOutput[output]/LayerNorm[LayerNorm]/layer_norm_0 INFO:nncf:Set sign: True and scale: [13.5330, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[7]/BertAttention[attention]/BertSelfOutput[output]/LayerNorm[LayerNorm]/layer_norm_0 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[7]/BertAttention[attention]/BertSelfOutput[output]/LayerNorm[LayerNorm]/layer_norm_0 INFO:nncf:Set sign: True and scale: [6.1915, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[7]/BertOutput[output]/LayerNorm[LayerNorm]/layer_norm_0 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[7]/BertOutput[output]/LayerNorm[LayerNorm]/layer_norm_0 INFO:nncf:Set sign: True and scale: [14.7554, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[8]/BertAttention[attention]/BertSelfOutput[output]/LayerNorm[LayerNorm]/layer_norm_0 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[8]/BertAttention[attention]/BertSelfOutput[output]/LayerNorm[LayerNorm]/layer_norm_0 INFO:nncf:Set sign: True and scale: [5.7933, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[8]/BertOutput[output]/LayerNorm[LayerNorm]/layer_norm_0 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[8]/BertOutput[output]/LayerNorm[LayerNorm]/layer_norm_0 INFO:nncf:Set sign: True and scale: [10.3499, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[9]/BertAttention[attention]/BertSelfOutput[output]/LayerNorm[LayerNorm]/layer_norm_0 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[9]/BertAttention[attention]/BertSelfOutput[output]/LayerNorm[LayerNorm]/layer_norm_0 INFO:nncf:Set sign: True and scale: [6.5209, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[9]/BertOutput[output]/LayerNorm[LayerNorm]/layer_norm_0 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[9]/BertOutput[output]/LayerNorm[LayerNorm]/layer_norm_0 INFO:nncf:Set sign: True and scale: [9.4660, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[10]/BertAttention[attention]/BertSelfOutput[output]/LayerNorm[LayerNorm]/layer_norm_0 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[10]/BertAttention[attention]/BertSelfOutput[output]/LayerNorm[LayerNorm]/layer_norm_0 INFO:nncf:Set sign: True and scale: [7.7494, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[10]/BertOutput[output]/LayerNorm[LayerNorm]/layer_norm_0 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[10]/BertOutput[output]/LayerNorm[LayerNorm]/layer_norm_0 INFO:nncf:Set sign: True and scale: [14.1016, ] for TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[11]/BertAttention[attention]/BertSelfOutput[output]/LayerNorm[LayerNorm]/layer_norm_0 INFO:nncf:Performing signed activation quantization for: TargetType.OPERATOR_POST_HOOK BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[11]/BertAttention[attention]/BertSelfOutput[output]/LayerNorm[LayerNorm]/layer_norm_0 WARNING:nncf:The overflow issue fix will be applied. Now all weight quantizers will effectively use only 7 bits out of 8 bits. This resolves the overflow issue problem on AVX2 and AVX-512 machines. Please take a look at the documentation for a detailed information. INFO:nncf:Set sign: True and scale: [0.2280, ] for TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEmbeddings[embeddings]/NNCFEmbedding[word_embeddings]/embedding_0 INFO:nncf:Performing signed weight quantization for: TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEmbeddings[embeddings]/NNCFEmbedding[word_embeddings]/embedding_0 INFO:nncf:Set sign: True and scale: [0.1551, ] for TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEmbeddings[embeddings]/NNCFEmbedding[position_embeddings]/embedding_0 INFO:nncf:Performing signed weight quantization for: TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEmbeddings[embeddings]/NNCFEmbedding[position_embeddings]/embedding_0 INFO:nncf:Set sign: True and scale: [0.5894, ] for TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEmbeddings[embeddings]/NNCFEmbedding[token_type_embeddings]/embedding_0 INFO:nncf:Performing signed weight quantization for: TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEmbeddings[embeddings]/NNCFEmbedding[token_type_embeddings]/embedding_0 INFO:nncf:Set sign: True and scale: [0.2171, ] for TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[0]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[query]/linear_0 INFO:nncf:Performing signed weight quantization for: TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[0]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[query]/linear_0 INFO:nncf:Set sign: True and scale: [0.3125, ] for TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[0]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[key]/linear_0 INFO:nncf:Performing signed weight quantization for: TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[0]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[key]/linear_0 INFO:nncf:Set sign: True and scale: [0.1119, ] for TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[0]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[value]/linear_0 INFO:nncf:Performing signed weight quantization for: TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[0]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[value]/linear_0 INFO:nncf:Set sign: True and scale: [0.1805, ] for TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[0]/BertAttention[attention]/BertSelfOutput[output]/NNCFLinear[dense]/linear_0 INFO:nncf:Performing signed weight quantization for: TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[0]/BertAttention[attention]/BertSelfOutput[output]/NNCFLinear[dense]/linear_0 INFO:nncf:Set sign: True and scale: [0.1899, ] for TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[0]/BertIntermediate[intermediate]/NNCFLinear[dense]/linear_0 INFO:nncf:Performing signed weight quantization for: TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[0]/BertIntermediate[intermediate]/NNCFLinear[dense]/linear_0 INFO:nncf:Set sign: True and scale: [0.4780, ] for TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[0]/BertOutput[output]/NNCFLinear[dense]/linear_0 INFO:nncf:Performing signed weight quantization for: TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[0]/BertOutput[output]/NNCFLinear[dense]/linear_0 INFO:nncf:Set sign: True and scale: [0.1869, ] for TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[1]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[query]/linear_0 INFO:nncf:Performing signed weight quantization for: TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[1]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[query]/linear_0 INFO:nncf:Set sign: True and scale: [0.2278, ] for TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[1]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[key]/linear_0 INFO:nncf:Performing signed weight quantization for: TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[1]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[key]/linear_0 INFO:nncf:Set sign: True and scale: [0.1179, ] for TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[1]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[value]/linear_0 INFO:nncf:Performing signed weight quantization for: TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[1]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[value]/linear_0 INFO:nncf:Set sign: True and scale: [0.1659, ] for TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[1]/BertAttention[attention]/BertSelfOutput[output]/NNCFLinear[dense]/linear_0 INFO:nncf:Performing signed weight quantization for: TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[1]/BertAttention[attention]/BertSelfOutput[output]/NNCFLinear[dense]/linear_0 INFO:nncf:Set sign: True and scale: [0.1715, ] for TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[1]/BertIntermediate[intermediate]/NNCFLinear[dense]/linear_0 INFO:nncf:Performing signed weight quantization for: TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[1]/BertIntermediate[intermediate]/NNCFLinear[dense]/linear_0 INFO:nncf:Set sign: True and scale: [0.3295, ] for TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[1]/BertOutput[output]/NNCFLinear[dense]/linear_0 INFO:nncf:Performing signed weight quantization for: TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[1]/BertOutput[output]/NNCFLinear[dense]/linear_0 INFO:nncf:Set sign: True and scale: [0.3058, ] for TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[2]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[query]/linear_0 INFO:nncf:Performing signed weight quantization for: TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[2]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[query]/linear_0 INFO:nncf:Set sign: True and scale: [0.3194, ] for TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[2]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[key]/linear_0 INFO:nncf:Performing signed weight quantization for: TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[2]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[key]/linear_0 INFO:nncf:Set sign: True and scale: [0.1175, ] for TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[2]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[value]/linear_0 INFO:nncf:Performing signed weight quantization for: TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[2]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[value]/linear_0 INFO:nncf:Set sign: True and scale: [0.1345, ] for TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[2]/BertAttention[attention]/BertSelfOutput[output]/NNCFLinear[dense]/linear_0 INFO:nncf:Performing signed weight quantization for: TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[2]/BertAttention[attention]/BertSelfOutput[output]/NNCFLinear[dense]/linear_0 INFO:nncf:Set sign: True and scale: [0.3839, ] for TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[2]/BertIntermediate[intermediate]/NNCFLinear[dense]/linear_0 INFO:nncf:Performing signed weight quantization for: TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[2]/BertIntermediate[intermediate]/NNCFLinear[dense]/linear_0 INFO:nncf:Set sign: True and scale: [0.2657, ] for TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[2]/BertOutput[output]/NNCFLinear[dense]/linear_0 INFO:nncf:Performing signed weight quantization for: TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[2]/BertOutput[output]/NNCFLinear[dense]/linear_0 INFO:nncf:Set sign: True and scale: [0.2068, ] for TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[3]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[query]/linear_0 INFO:nncf:Performing signed weight quantization for: TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[3]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[query]/linear_0 INFO:nncf:Set sign: True and scale: [0.2180, ] for TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[3]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[key]/linear_0 INFO:nncf:Performing signed weight quantization for: TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[3]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[key]/linear_0 INFO:nncf:Set sign: True and scale: [0.1241, ] for TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[3]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[value]/linear_0 INFO:nncf:Performing signed weight quantization for: TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[3]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[value]/linear_0 INFO:nncf:Set sign: True and scale: [0.1482, ] for TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[3]/BertAttention[attention]/BertSelfOutput[output]/NNCFLinear[dense]/linear_0 INFO:nncf:Performing signed weight quantization for: TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[3]/BertAttention[attention]/BertSelfOutput[output]/NNCFLinear[dense]/linear_0 INFO:nncf:Set sign: True and scale: [0.3309, ] for TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[3]/BertIntermediate[intermediate]/NNCFLinear[dense]/linear_0 INFO:nncf:Performing signed weight quantization for: TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[3]/BertIntermediate[intermediate]/NNCFLinear[dense]/linear_0 INFO:nncf:Set sign: True and scale: [0.2841, ] for TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[3]/BertOutput[output]/NNCFLinear[dense]/linear_0 INFO:nncf:Performing signed weight quantization for: TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[3]/BertOutput[output]/NNCFLinear[dense]/linear_0 INFO:nncf:Set sign: True and scale: [0.1797, ] for TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[4]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[query]/linear_0 INFO:nncf:Performing signed weight quantization for: TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[4]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[query]/linear_0 INFO:nncf:Set sign: True and scale: [0.1788, ] for TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[4]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[key]/linear_0 INFO:nncf:Performing signed weight quantization for: TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[4]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[key]/linear_0 INFO:nncf:Set sign: True and scale: [0.1461, ] for TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[4]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[value]/linear_0 INFO:nncf:Performing signed weight quantization for: TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[4]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[value]/linear_0 INFO:nncf:Set sign: True and scale: [0.1489, ] for TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[4]/BertAttention[attention]/BertSelfOutput[output]/NNCFLinear[dense]/linear_0 INFO:nncf:Performing signed weight quantization for: TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[4]/BertAttention[attention]/BertSelfOutput[output]/NNCFLinear[dense]/linear_0 INFO:nncf:Set sign: True and scale: [0.2849, ] for TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[4]/BertIntermediate[intermediate]/NNCFLinear[dense]/linear_0 INFO:nncf:Performing signed weight quantization for: TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[4]/BertIntermediate[intermediate]/NNCFLinear[dense]/linear_0 INFO:nncf:Set sign: True and scale: [0.2799, ] for TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[4]/BertOutput[output]/NNCFLinear[dense]/linear_0 INFO:nncf:Performing signed weight quantization for: TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[4]/BertOutput[output]/NNCFLinear[dense]/linear_0 INFO:nncf:Set sign: True and scale: [0.1738, ] for TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[5]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[query]/linear_0 INFO:nncf:Performing signed weight quantization for: TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[5]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[query]/linear_0 INFO:nncf:Set sign: True and scale: [0.1816, ] for TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[5]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[key]/linear_0 INFO:nncf:Performing signed weight quantization for: TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[5]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[key]/linear_0 INFO:nncf:Set sign: True and scale: [0.1570, ] for TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[5]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[value]/linear_0 INFO:nncf:Performing signed weight quantization for: TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[5]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[value]/linear_0 INFO:nncf:Set sign: True and scale: [0.1515, ] for TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[5]/BertAttention[attention]/BertSelfOutput[output]/NNCFLinear[dense]/linear_0 INFO:nncf:Performing signed weight quantization for: TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[5]/BertAttention[attention]/BertSelfOutput[output]/NNCFLinear[dense]/linear_0 INFO:nncf:Set sign: True and scale: [0.2263, ] for TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[5]/BertIntermediate[intermediate]/NNCFLinear[dense]/linear_0 INFO:nncf:Performing signed weight quantization for: TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[5]/BertIntermediate[intermediate]/NNCFLinear[dense]/linear_0 INFO:nncf:Set sign: True and scale: [0.2835, ] for TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[5]/BertOutput[output]/NNCFLinear[dense]/linear_0 INFO:nncf:Performing signed weight quantization for: TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[5]/BertOutput[output]/NNCFLinear[dense]/linear_0 INFO:nncf:Set sign: True and scale: [0.1689, ] for TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[6]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[query]/linear_0 INFO:nncf:Performing signed weight quantization for: TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[6]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[query]/linear_0 INFO:nncf:Set sign: True and scale: [0.1792, ] for TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[6]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[key]/linear_0 INFO:nncf:Performing signed weight quantization for: TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[6]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[key]/linear_0 INFO:nncf:Set sign: True and scale: [0.1346, ] for TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[6]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[value]/linear_0 INFO:nncf:Performing signed weight quantization for: TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[6]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[value]/linear_0 INFO:nncf:Set sign: True and scale: [0.1372, ] for TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[6]/BertAttention[attention]/BertSelfOutput[output]/NNCFLinear[dense]/linear_0 INFO:nncf:Performing signed weight quantization for: TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[6]/BertAttention[attention]/BertSelfOutput[output]/NNCFLinear[dense]/linear_0 INFO:nncf:Set sign: True and scale: [0.1811, ] for TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[6]/BertIntermediate[intermediate]/NNCFLinear[dense]/linear_0 INFO:nncf:Performing signed weight quantization for: TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[6]/BertIntermediate[intermediate]/NNCFLinear[dense]/linear_0 INFO:nncf:Set sign: True and scale: [0.2911, ] for TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[6]/BertOutput[output]/NNCFLinear[dense]/linear_0 INFO:nncf:Performing signed weight quantization for: TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[6]/BertOutput[output]/NNCFLinear[dense]/linear_0 INFO:nncf:Set sign: True and scale: [0.1743, ] for TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[7]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[query]/linear_0 INFO:nncf:Performing signed weight quantization for: TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[7]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[query]/linear_0 INFO:nncf:Set sign: True and scale: [0.1844, ] for TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[7]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[key]/linear_0 INFO:nncf:Performing signed weight quantization for: TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[7]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[key]/linear_0 INFO:nncf:Set sign: True and scale: [0.1330, ] for TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[7]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[value]/linear_0 INFO:nncf:Performing signed weight quantization for: TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[7]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[value]/linear_0 INFO:nncf:Set sign: True and scale: [0.1359, ] for TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[7]/BertAttention[attention]/BertSelfOutput[output]/NNCFLinear[dense]/linear_0 INFO:nncf:Performing signed weight quantization for: TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[7]/BertAttention[attention]/BertSelfOutput[output]/NNCFLinear[dense]/linear_0 INFO:nncf:Set sign: True and scale: [0.2244, ] for TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[7]/BertIntermediate[intermediate]/NNCFLinear[dense]/linear_0 INFO:nncf:Performing signed weight quantization for: TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[7]/BertIntermediate[intermediate]/NNCFLinear[dense]/linear_0 INFO:nncf:Set sign: True and scale: [0.2513, ] for TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[7]/BertOutput[output]/NNCFLinear[dense]/linear_0 INFO:nncf:Performing signed weight quantization for: TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[7]/BertOutput[output]/NNCFLinear[dense]/linear_0 INFO:nncf:Set sign: True and scale: [0.1815, ] for TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[8]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[query]/linear_0 INFO:nncf:Performing signed weight quantization for: TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[8]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[query]/linear_0 INFO:nncf:Set sign: True and scale: [0.1927, ] for TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[8]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[key]/linear_0 INFO:nncf:Performing signed weight quantization for: TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[8]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[key]/linear_0 INFO:nncf:Set sign: True and scale: [0.1330, ] for TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[8]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[value]/linear_0 INFO:nncf:Performing signed weight quantization for: TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[8]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[value]/linear_0 INFO:nncf:Set sign: True and scale: [0.1297, ] for TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[8]/BertAttention[attention]/BertSelfOutput[output]/NNCFLinear[dense]/linear_0 INFO:nncf:Performing signed weight quantization for: TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[8]/BertAttention[attention]/BertSelfOutput[output]/NNCFLinear[dense]/linear_0 INFO:nncf:Set sign: True and scale: [0.1677, ] for TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[8]/BertIntermediate[intermediate]/NNCFLinear[dense]/linear_0 INFO:nncf:Performing signed weight quantization for: TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[8]/BertIntermediate[intermediate]/NNCFLinear[dense]/linear_0 INFO:nncf:Set sign: True and scale: [0.2830, ] for TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[8]/BertOutput[output]/NNCFLinear[dense]/linear_0 INFO:nncf:Performing signed weight quantization for: TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[8]/BertOutput[output]/NNCFLinear[dense]/linear_0 INFO:nncf:Set sign: True and scale: [0.1797, ] for TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[9]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[query]/linear_0 INFO:nncf:Performing signed weight quantization for: TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[9]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[query]/linear_0 INFO:nncf:Set sign: True and scale: [0.2035, ] for TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[9]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[key]/linear_0 INFO:nncf:Performing signed weight quantization for: TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[9]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[key]/linear_0 INFO:nncf:Set sign: True and scale: [0.1279, ] for TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[9]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[value]/linear_0 INFO:nncf:Performing signed weight quantization for: TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[9]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[value]/linear_0 INFO:nncf:Set sign: True and scale: [0.1226, ] for TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[9]/BertAttention[attention]/BertSelfOutput[output]/NNCFLinear[dense]/linear_0 INFO:nncf:Performing signed weight quantization for: TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[9]/BertAttention[attention]/BertSelfOutput[output]/NNCFLinear[dense]/linear_0 INFO:nncf:Set sign: True and scale: [0.2134, ] for TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[9]/BertIntermediate[intermediate]/NNCFLinear[dense]/linear_0 INFO:nncf:Performing signed weight quantization for: TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[9]/BertIntermediate[intermediate]/NNCFLinear[dense]/linear_0 INFO:nncf:Set sign: True and scale: [1.1807, ] for TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[9]/BertOutput[output]/NNCFLinear[dense]/linear_0 INFO:nncf:Performing signed weight quantization for: TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[9]/BertOutput[output]/NNCFLinear[dense]/linear_0 INFO:nncf:Set sign: True and scale: [0.1933, ] for TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[10]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[query]/linear_0 INFO:nncf:Performing signed weight quantization for: TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[10]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[query]/linear_0 INFO:nncf:Set sign: True and scale: [0.1842, ] for TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[10]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[key]/linear_0 INFO:nncf:Performing signed weight quantization for: TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[10]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[key]/linear_0 INFO:nncf:Set sign: True and scale: [0.1474, ] for TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[10]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[value]/linear_0 INFO:nncf:Performing signed weight quantization for: TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[10]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[value]/linear_0 INFO:nncf:Set sign: True and scale: [0.1218, ] for TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[10]/BertAttention[attention]/BertSelfOutput[output]/NNCFLinear[dense]/linear_0 INFO:nncf:Performing signed weight quantization for: TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[10]/BertAttention[attention]/BertSelfOutput[output]/NNCFLinear[dense]/linear_0 INFO:nncf:Set sign: True and scale: [0.4426, ] for TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[10]/BertIntermediate[intermediate]/NNCFLinear[dense]/linear_0 INFO:nncf:Performing signed weight quantization for: TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[10]/BertIntermediate[intermediate]/NNCFLinear[dense]/linear_0 INFO:nncf:Set sign: True and scale: [0.3693, ] for TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[10]/BertOutput[output]/NNCFLinear[dense]/linear_0 INFO:nncf:Performing signed weight quantization for: TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[10]/BertOutput[output]/NNCFLinear[dense]/linear_0 INFO:nncf:Set sign: True and scale: [0.2113, ] for TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[11]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[query]/linear_0 INFO:nncf:Performing signed weight quantization for: TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[11]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[query]/linear_0 INFO:nncf:Set sign: True and scale: [0.1714, ] for TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[11]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[key]/linear_0 INFO:nncf:Performing signed weight quantization for: TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[11]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[key]/linear_0 INFO:nncf:Set sign: True and scale: [0.1439, ] for TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[11]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[value]/linear_0 INFO:nncf:Performing signed weight quantization for: TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[11]/BertAttention[attention]/BertSelfAttention[self]/NNCFLinear[value]/linear_0 INFO:nncf:Set sign: True and scale: [0.1319, ] for TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[11]/BertAttention[attention]/BertSelfOutput[output]/NNCFLinear[dense]/linear_0 INFO:nncf:Performing signed weight quantization for: TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[11]/BertAttention[attention]/BertSelfOutput[output]/NNCFLinear[dense]/linear_0 INFO:nncf:Set sign: True and scale: [0.2404, ] for TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[11]/BertIntermediate[intermediate]/NNCFLinear[dense]/linear_0 INFO:nncf:Performing signed weight quantization for: TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[11]/BertIntermediate[intermediate]/NNCFLinear[dense]/linear_0 INFO:nncf:Set sign: True and scale: [0.2133, ] for TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[11]/BertOutput[output]/NNCFLinear[dense]/linear_0 INFO:nncf:Performing signed weight quantization for: TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/BertModel[bert]/BertEncoder[encoder]/ModuleList[layer]/BertLayer[11]/BertOutput[output]/NNCFLinear[dense]/linear_0 INFO:nncf:Set sign: True and scale: [0.1000, ] for TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/NNCFLinear[qa_outputs]/linear_0 INFO:nncf:Performing signed weight quantization for: TargetType.OPERATION_WITH_WEIGHTS BertForQuestionAnswering/NNCFLinear[qa_outputs]/linear_0 WARNING:nncf:Module `BertForQuestionAnswering/BertModel[bert]/BertEmbeddings[embeddings]/NNCFEmbedding[token_type_embeddings]/embedding_0` has quantized weights and no quantized inputs! WARNING:nncf:Module `BertForQuestionAnswering/BertModel[bert]/BertEmbeddings[embeddings]/NNCFEmbedding[word_embeddings]/embedding_0` has quantized weights and no quantized inputs! WARNING:nncf:Module `BertForQuestionAnswering/BertModel[bert]/BertEmbeddings[embeddings]/NNCFEmbedding[position_embeddings]/embedding_0` has quantized weights and no quantized inputs! INFO:nncf:BatchNorm statistics adaptation █ | 1 / 13 INFO:nncf:BatchNorm statistics adaptation ██ | 2 / 13 INFO:nncf:BatchNorm statistics adaptation ███ | 3 / 13 INFO:nncf:BatchNorm statistics adaptation ████ | 4 / 13 INFO:nncf:BatchNorm statistics adaptation ██████ | 5 / 13 INFO:nncf:BatchNorm statistics adaptation ███████ | 6 / 13 INFO:nncf:BatchNorm statistics adaptation ████████ | 7 / 13 INFO:nncf:BatchNorm statistics adaptation █████████ | 8 / 13 INFO:nncf:BatchNorm statistics adaptation ███████████ | 9 / 13 INFO:nncf:BatchNorm statistics adaptation ████████████ | 10 / 13 INFO:nncf:BatchNorm statistics adaptation █████████████ | 11 / 13 INFO:nncf:BatchNorm statistics adaptation ██████████████ | 12 / 13 INFO:nncf:BatchNorm statistics adaptation ████████████████ | 13 / 13 WARNING:nncf:Graphviz is not installed - only the .dot model visualization format will be used. Install pygraphviz into your Python environment and graphviz system-wide to enable PNG rendering. [INFO|trainer.py:434] 2022-01-16 13:21:10,698 >> max_steps is given, it will override any value given in num_train_epochs [INFO|trainer.py:1209] 2022-01-16 13:21:10,729 >> ***** Running training ***** [INFO|trainer.py:1210] 2022-01-16 13:21:10,729 >> Num examples = 88524 [INFO|trainer.py:1211] 2022-01-16 13:21:10,729 >> Num Epochs = 1 [INFO|trainer.py:1212] 2022-01-16 13:21:10,729 >> Instantaneous batch size per device = 16 [INFO|trainer.py:1213] 2022-01-16 13:21:10,729 >> Total train batch size (w. parallel, distributed & accumulation) = 16 [INFO|trainer.py:1214] 2022-01-16 13:21:10,729 >> Gradient Accumulation steps = 1 [INFO|trainer.py:1215] 2022-01-16 13:21:10,729 >> Total optimization steps = 25 [WARNING|integrations.py:650] 2022-01-16 13:21:11,179 >> Trainer is attempting to log a value of "{0: [0, 2, 4, 5, 6, 7, 11], 1: [0, 2, 3, 5, 6, 7, 8], 2: [8, 4, 7], 3: [2, 4, 6], 4: [1, 2, 11], 5: [1, 2, 5, 6, 7, 11], 6: [0, 2, 3, 7, 10], 7: [1, 3, 6, 7, 11], 8: [0, 3, 4, 5, 8], 9: [1, 3, 4, 5, 7, 9, 10], 10: [1, 4, 5, 6, 7, 8], 11: [4, 5, 7, 8, 10, 11]}" for key "pruned_heads" as a parameter. MLflow's log_param() only accepts values no longer than 250 characters so we dropped this attribute. [INFO|integrations.py:445] 2022-01-16 13:21:11,205 >> Automatic Weights & Biases logging enabled, to disable set os.environ["WANDB_DISABLED"] = "true" wandb: Currently logged in as: vchua (use `wandb login --relogin` to force relogin) wandb: Tracking run with wandb version 0.12.9 wandb: Syncing run run10-bert-squad-cropped-qat-customkd-lt-5eph wandb: View project at https://wandb.ai/vchua/pruneofa-tl%20%28csr-dgx1-03%29 wandb: View run at https://wandb.ai/vchua/pruneofa-tl%20%28csr-dgx1-03%29/runs/3ald1d3c wandb: Run data is saved locally in /home/vchua/tld-poc/transformers/examples/pytorch/question-answering/wandb/run-20220116_132111-3ald1d3c wandb: Run `wandb offline` to turn off syncing. 0% 0/25 [00:00> Training completed. Do not forget to share your model on huggingface.co/models =) 100% 25/25 [00:19<00:00, 1.32it/s, loss=0.289] 100% 25/25 [00:19<00:00, 1.25it/s, loss=0.289] [INFO|trainer.py:1982] 2022-01-16 13:21:35,237 >> Saving model checkpoint to /data1/vchua/tld-poc-csr-dgx1-03//run10-bert-squad-cropped-qat-customkd-lt-5eph [INFO|configuration_utils.py:379] 2022-01-16 13:21:35,244 >> Configuration saved in /data1/vchua/tld-poc-csr-dgx1-03//run10-bert-squad-cropped-qat-customkd-lt-5eph/config.json [INFO|modeling_utils.py:1004] 2022-01-16 13:21:35,643 >> Model weights saved in /data1/vchua/tld-poc-csr-dgx1-03//run10-bert-squad-cropped-qat-customkd-lt-5eph/pytorch_model.bin [INFO|tokenization_utils_base.py:2006] 2022-01-16 13:21:35,643 >> tokenizer config file saved in /data1/vchua/tld-poc-csr-dgx1-03//run10-bert-squad-cropped-qat-customkd-lt-5eph/tokenizer_config.json [INFO|tokenization_utils_base.py:2012] 2022-01-16 13:21:35,644 >> Special tokens file saved in /data1/vchua/tld-poc-csr-dgx1-03//run10-bert-squad-cropped-qat-customkd-lt-5eph/special_tokens_map.json Statistics of the quantization algorithm: +--------------------------------+-------+ | Statistic's name | Value | +================================+=======+ | Ratio of enabled quantizations | 100 | +--------------------------------+-------+ Statistics of the quantization share: +----------------------------------+----------------------+ | Statistic's name | Value | +==================================+======================+ | Symmetric WQs / All placed WQs | 100.00 % (76 / 76) | +----------------------------------+----------------------+ | Asymmetric WQs / All placed WQs | 0.00 % (0 / 76) | +----------------------------------+----------------------+ | Signed WQs / All placed WQs | 100.00 % (76 / 76) | +----------------------------------+----------------------+ | Unsigned WQs / All placed WQs | 0.00 % (0 / 76) | +----------------------------------+----------------------+ | Per-tensor WQs / All placed WQs | 100.00 % (76 / 76) | +----------------------------------+----------------------+ | Per-channel WQs / All placed WQs | 0.00 % (0 / 76) | +----------------------------------+----------------------+ | Placed WQs / Potential WQs | 100.00 % (76 / 76) | +----------------------------------+----------------------+ | Symmetric AQs / All placed AQs | 100.00 % (162 / 162) | +----------------------------------+----------------------+ | Asymmetric AQs / All placed AQs | 0.00 % (0 / 162) | +----------------------------------+----------------------+ | Signed AQs / All placed AQs | 92.59 % (150 / 162) | +----------------------------------+----------------------+ | Unsigned AQs / All placed AQs | 7.41 % (12 / 162) | +----------------------------------+----------------------+ | Per-tensor AQs / All placed AQs | 100.00 % (162 / 162) | +----------------------------------+----------------------+ | Per-channel AQs / All placed AQs | 0.00 % (0 / 162) | +----------------------------------+----------------------+ Statistics of the bitwidth distribution: +--------------+---------------------+--------------------+--------------------+ | Num bits (N) | N-bits WQs / Placed | N-bits AQs / | N-bits Qs / Placed | | | WQs | Placed AQs | Qs | +==============+=====================+====================+====================+ | 8 | 100.00 % (76 / 76) | 100.00 % (162 / | 100.00 % (238 / | | | | 162) | 238) | +--------------+---------------------+--------------------+--------------------+ {'loss': 0.2891, 'learning_rate': 0.0, 'compression_loss': 0.0, 'label_loss': 0.30624544620513916, 'teacher_loss': 0.2872086465358734, 'epoch': 0.0} {'train_runtime': 24.5019, 'train_samples_per_second': 16.325, 'train_steps_per_second': 1.02, 'train_loss': 0.3293510854244232, 'epoch': 0.0} ***** train metrics ***** epoch = 0.0 train_loss = 0.3294 train_runtime = 0:00:24.50 train_samples = 88524 train_samples_per_second = 16.325 train_steps_per_second = 1.02 01/16/2022 13:21:35 - INFO - __main__ - *** Evaluate *** [INFO|trainer.py:554] 2022-01-16 13:21:35,765 >> The following columns in the evaluation set don't have a corresponding argument in `NNCFNetwork.forward` and have been ignored: offset_mapping, example_id. [INFO|trainer.py:2244] 2022-01-16 13:21:35,767 >> ***** Running Evaluation ***** [INFO|trainer.py:2246] 2022-01-16 13:21:35,767 >> Num examples = 10784 [INFO|trainer.py:2249] 2022-01-16 13:21:35,767 >> Batch size = 128 0% 0/85 [00:00