Commit ·
cafa4ac
1
Parent(s): 20f88fb
Upload folder using huggingface_hub
Browse files- README.md +21 -0
- adapter_config.json +21 -0
- adapter_model.bin +3 -0
- special_tokens_map.json +24 -0
- tokenizer.json +0 -0
- tokenizer.model +3 -0
- tokenizer_config.json +34 -0
- train_args.json +123 -0
- training_args.bin +3 -0
- training_logs.json +1067 -0
README.md
ADDED
|
@@ -0,0 +1,21 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
library_name: peft
|
| 3 |
+
---
|
| 4 |
+
## Training procedure
|
| 5 |
+
|
| 6 |
+
|
| 7 |
+
The following `bitsandbytes` quantization config was used during training:
|
| 8 |
+
- quant_method: bitsandbytes
|
| 9 |
+
- load_in_8bit: False
|
| 10 |
+
- load_in_4bit: True
|
| 11 |
+
- llm_int8_threshold: 6.0
|
| 12 |
+
- llm_int8_skip_modules: None
|
| 13 |
+
- llm_int8_enable_fp32_cpu_offload: False
|
| 14 |
+
- llm_int8_has_fp16_weight: False
|
| 15 |
+
- bnb_4bit_quant_type: nf4
|
| 16 |
+
- bnb_4bit_use_double_quant: True
|
| 17 |
+
- bnb_4bit_compute_dtype: float16
|
| 18 |
+
### Framework versions
|
| 19 |
+
|
| 20 |
+
|
| 21 |
+
- PEFT 0.5.0
|
adapter_config.json
ADDED
|
@@ -0,0 +1,21 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"auto_mapping": null,
|
| 3 |
+
"base_model_name_or_path": "models/openllama-3b",
|
| 4 |
+
"bias": "none",
|
| 5 |
+
"fan_in_fan_out": false,
|
| 6 |
+
"inference_mode": true,
|
| 7 |
+
"init_lora_weights": true,
|
| 8 |
+
"layers_pattern": null,
|
| 9 |
+
"layers_to_transform": null,
|
| 10 |
+
"lora_alpha": 16,
|
| 11 |
+
"lora_dropout": 0.1,
|
| 12 |
+
"modules_to_save": null,
|
| 13 |
+
"peft_type": "LORA",
|
| 14 |
+
"r": 64,
|
| 15 |
+
"revision": null,
|
| 16 |
+
"target_modules": [
|
| 17 |
+
"q_proj",
|
| 18 |
+
"v_proj"
|
| 19 |
+
],
|
| 20 |
+
"task_type": "SEQ_CLS"
|
| 21 |
+
}
|
adapter_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:97288328bbbc4b8e32d709acfaab036a524b2968124ae4543bc0890b112b19b8
|
| 3 |
+
size 85258545
|
special_tokens_map.json
ADDED
|
@@ -0,0 +1,24 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": {
|
| 3 |
+
"content": "<s>",
|
| 4 |
+
"lstrip": false,
|
| 5 |
+
"normalized": true,
|
| 6 |
+
"rstrip": false,
|
| 7 |
+
"single_word": false
|
| 8 |
+
},
|
| 9 |
+
"eos_token": {
|
| 10 |
+
"content": "</s>",
|
| 11 |
+
"lstrip": false,
|
| 12 |
+
"normalized": true,
|
| 13 |
+
"rstrip": false,
|
| 14 |
+
"single_word": false
|
| 15 |
+
},
|
| 16 |
+
"pad_token": "</s>",
|
| 17 |
+
"unk_token": {
|
| 18 |
+
"content": "<unk>",
|
| 19 |
+
"lstrip": false,
|
| 20 |
+
"normalized": true,
|
| 21 |
+
"rstrip": false,
|
| 22 |
+
"single_word": false
|
| 23 |
+
}
|
| 24 |
+
}
|
tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
tokenizer.model
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ab1b681ec7fc02fed5edd3026687d7a692a918c4dd8e150ca2e3994a6229843b
|
| 3 |
+
size 534194
|
tokenizer_config.json
ADDED
|
@@ -0,0 +1,34 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": {
|
| 3 |
+
"__type": "AddedToken",
|
| 4 |
+
"content": "<s>",
|
| 5 |
+
"lstrip": false,
|
| 6 |
+
"normalized": true,
|
| 7 |
+
"rstrip": false,
|
| 8 |
+
"single_word": false
|
| 9 |
+
},
|
| 10 |
+
"clean_up_tokenization_spaces": false,
|
| 11 |
+
"eos_token": {
|
| 12 |
+
"__type": "AddedToken",
|
| 13 |
+
"content": "</s>",
|
| 14 |
+
"lstrip": false,
|
| 15 |
+
"normalized": true,
|
| 16 |
+
"rstrip": false,
|
| 17 |
+
"single_word": false
|
| 18 |
+
},
|
| 19 |
+
"legacy": null,
|
| 20 |
+
"model_max_length": 2048,
|
| 21 |
+
"pad_token": null,
|
| 22 |
+
"sp_model_kwargs": {},
|
| 23 |
+
"spaces_between_special_tokens": false,
|
| 24 |
+
"tokenizer_class": "LlamaTokenizer",
|
| 25 |
+
"unk_token": {
|
| 26 |
+
"__type": "AddedToken",
|
| 27 |
+
"content": "<unk>",
|
| 28 |
+
"lstrip": false,
|
| 29 |
+
"normalized": true,
|
| 30 |
+
"rstrip": false,
|
| 31 |
+
"single_word": false
|
| 32 |
+
},
|
| 33 |
+
"use_default_system_prompt": true
|
| 34 |
+
}
|
train_args.json
ADDED
|
@@ -0,0 +1,123 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"initial_model_dir": "models/openllama-3b",
|
| 3 |
+
"distribution_id": "wrong_arc",
|
| 4 |
+
"date_trained": "15/10/2023 06:49:10",
|
| 5 |
+
"output_dir": "models/classify_lora/openllama-3b-wrong_arc",
|
| 6 |
+
"overwrite_output_dir": false,
|
| 7 |
+
"do_train": false,
|
| 8 |
+
"do_eval": true,
|
| 9 |
+
"do_predict": false,
|
| 10 |
+
"evaluation_strategy": "steps",
|
| 11 |
+
"prediction_loss_only": false,
|
| 12 |
+
"per_device_train_batch_size": 16,
|
| 13 |
+
"per_device_eval_batch_size": 32,
|
| 14 |
+
"per_gpu_train_batch_size": null,
|
| 15 |
+
"per_gpu_eval_batch_size": null,
|
| 16 |
+
"gradient_accumulation_steps": 1,
|
| 17 |
+
"eval_accumulation_steps": null,
|
| 18 |
+
"eval_delay": 0,
|
| 19 |
+
"learning_rate": 7.2e-05,
|
| 20 |
+
"weight_decay": 0.0,
|
| 21 |
+
"adam_beta1": 0.9,
|
| 22 |
+
"adam_beta2": 0.999,
|
| 23 |
+
"adam_epsilon": 1e-08,
|
| 24 |
+
"max_grad_norm": 0.3,
|
| 25 |
+
"num_train_epochs": 6,
|
| 26 |
+
"max_steps": 100,
|
| 27 |
+
"lr_scheduler_type": "constant",
|
| 28 |
+
"warmup_ratio": 0.03,
|
| 29 |
+
"warmup_steps": 0,
|
| 30 |
+
"log_level": "passive",
|
| 31 |
+
"log_level_replica": "warning",
|
| 32 |
+
"log_on_each_node": true,
|
| 33 |
+
"logging_dir": "models/classify_lora/openllama-3b-wrong_arc/runs/Oct15_06-39-53_compute-permanent-node-975",
|
| 34 |
+
"logging_strategy": "steps",
|
| 35 |
+
"logging_first_step": false,
|
| 36 |
+
"logging_steps": 1,
|
| 37 |
+
"logging_nan_inf_filter": true,
|
| 38 |
+
"save_strategy": "steps",
|
| 39 |
+
"save_steps": 25,
|
| 40 |
+
"save_total_limit": 0,
|
| 41 |
+
"save_safetensors": false,
|
| 42 |
+
"save_on_each_node": false,
|
| 43 |
+
"no_cuda": false,
|
| 44 |
+
"use_cpu": false,
|
| 45 |
+
"use_mps_device": false,
|
| 46 |
+
"seed": 42,
|
| 47 |
+
"data_seed": null,
|
| 48 |
+
"jit_mode_eval": false,
|
| 49 |
+
"use_ipex": false,
|
| 50 |
+
"bf16": false,
|
| 51 |
+
"fp16": false,
|
| 52 |
+
"fp16_opt_level": "O1",
|
| 53 |
+
"half_precision_backend": "auto",
|
| 54 |
+
"bf16_full_eval": false,
|
| 55 |
+
"fp16_full_eval": false,
|
| 56 |
+
"tf32": null,
|
| 57 |
+
"local_rank": 1,
|
| 58 |
+
"ddp_backend": null,
|
| 59 |
+
"tpu_num_cores": null,
|
| 60 |
+
"tpu_metrics_debug": false,
|
| 61 |
+
"debug": [],
|
| 62 |
+
"dataloader_drop_last": false,
|
| 63 |
+
"eval_steps": 25,
|
| 64 |
+
"dataloader_num_workers": 0,
|
| 65 |
+
"past_index": -1,
|
| 66 |
+
"run_name": "train|models-classify_lora-openllama-3b-wrong_arc",
|
| 67 |
+
"disable_tqdm": false,
|
| 68 |
+
"remove_unused_columns": false,
|
| 69 |
+
"label_names": null,
|
| 70 |
+
"load_best_model_at_end": false,
|
| 71 |
+
"metric_for_best_model": "eval_wrong_arc_score",
|
| 72 |
+
"greater_is_better": true,
|
| 73 |
+
"ignore_data_skip": false,
|
| 74 |
+
"sharded_ddp": [],
|
| 75 |
+
"fsdp": [],
|
| 76 |
+
"fsdp_min_num_params": 0,
|
| 77 |
+
"fsdp_config": {
|
| 78 |
+
"min_num_params": 0,
|
| 79 |
+
"xla": false,
|
| 80 |
+
"xla_fsdp_grad_ckpt": false
|
| 81 |
+
},
|
| 82 |
+
"fsdp_transformer_layer_cls_to_wrap": null,
|
| 83 |
+
"deepspeed": "configs/ds_zero_1.json",
|
| 84 |
+
"label_smoothing_factor": 0.0,
|
| 85 |
+
"optim": "paged_adamw_32bit",
|
| 86 |
+
"optim_args": null,
|
| 87 |
+
"adafactor": false,
|
| 88 |
+
"group_by_length": false,
|
| 89 |
+
"length_column_name": "length",
|
| 90 |
+
"report_to": [
|
| 91 |
+
"wandb"
|
| 92 |
+
],
|
| 93 |
+
"ddp_find_unused_parameters": false,
|
| 94 |
+
"ddp_bucket_cap_mb": null,
|
| 95 |
+
"ddp_broadcast_buffers": null,
|
| 96 |
+
"dataloader_pin_memory": true,
|
| 97 |
+
"skip_memory_metrics": true,
|
| 98 |
+
"use_legacy_prediction_loop": false,
|
| 99 |
+
"push_to_hub": false,
|
| 100 |
+
"resume_from_checkpoint": null,
|
| 101 |
+
"hub_model_id": null,
|
| 102 |
+
"hub_strategy": "every_save",
|
| 103 |
+
"hub_token": null,
|
| 104 |
+
"hub_private_repo": false,
|
| 105 |
+
"hub_always_push": false,
|
| 106 |
+
"gradient_checkpointing": false,
|
| 107 |
+
"include_inputs_for_metrics": false,
|
| 108 |
+
"fp16_backend": "auto",
|
| 109 |
+
"push_to_hub_model_id": null,
|
| 110 |
+
"push_to_hub_organization": null,
|
| 111 |
+
"push_to_hub_token": null,
|
| 112 |
+
"_n_gpu": 1,
|
| 113 |
+
"mp_parameters": "",
|
| 114 |
+
"auto_find_batch_size": false,
|
| 115 |
+
"full_determinism": false,
|
| 116 |
+
"torchdynamo": null,
|
| 117 |
+
"ray_scope": "last",
|
| 118 |
+
"ddp_timeout": 1800,
|
| 119 |
+
"torch_compile": false,
|
| 120 |
+
"torch_compile_backend": null,
|
| 121 |
+
"torch_compile_mode": null,
|
| 122 |
+
"dispatch_batches": null
|
| 123 |
+
}
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:89f7856deb6b4712e30dcb568c12fb5aebbcf93172eb8916b1fcdfe59dac35f1
|
| 3 |
+
size 5115
|
training_logs.json
ADDED
|
@@ -0,0 +1,1067 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
[
|
| 2 |
+
{
|
| 3 |
+
"loss": 0.7375,
|
| 4 |
+
"learning_rate": 7.2e-05,
|
| 5 |
+
"epoch": 0.05,
|
| 6 |
+
"step": 1
|
| 7 |
+
},
|
| 8 |
+
{
|
| 9 |
+
"loss": 0.7385,
|
| 10 |
+
"learning_rate": 7.2e-05,
|
| 11 |
+
"epoch": 0.11,
|
| 12 |
+
"step": 2
|
| 13 |
+
},
|
| 14 |
+
{
|
| 15 |
+
"loss": 0.7184,
|
| 16 |
+
"learning_rate": 7.2e-05,
|
| 17 |
+
"epoch": 0.16,
|
| 18 |
+
"step": 3
|
| 19 |
+
},
|
| 20 |
+
{
|
| 21 |
+
"loss": 0.6809,
|
| 22 |
+
"learning_rate": 7.2e-05,
|
| 23 |
+
"epoch": 0.21,
|
| 24 |
+
"step": 4
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"loss": 0.6997,
|
| 28 |
+
"learning_rate": 7.2e-05,
|
| 29 |
+
"epoch": 0.26,
|
| 30 |
+
"step": 5
|
| 31 |
+
},
|
| 32 |
+
{
|
| 33 |
+
"loss": 0.7176,
|
| 34 |
+
"learning_rate": 7.2e-05,
|
| 35 |
+
"epoch": 0.32,
|
| 36 |
+
"step": 6
|
| 37 |
+
},
|
| 38 |
+
{
|
| 39 |
+
"loss": 0.6942,
|
| 40 |
+
"learning_rate": 7.2e-05,
|
| 41 |
+
"epoch": 0.37,
|
| 42 |
+
"step": 7
|
| 43 |
+
},
|
| 44 |
+
{
|
| 45 |
+
"loss": 0.7213,
|
| 46 |
+
"learning_rate": 7.2e-05,
|
| 47 |
+
"epoch": 0.42,
|
| 48 |
+
"step": 8
|
| 49 |
+
},
|
| 50 |
+
{
|
| 51 |
+
"loss": 0.6942,
|
| 52 |
+
"learning_rate": 7.2e-05,
|
| 53 |
+
"epoch": 0.47,
|
| 54 |
+
"step": 9
|
| 55 |
+
},
|
| 56 |
+
{
|
| 57 |
+
"loss": 0.7114,
|
| 58 |
+
"learning_rate": 7.2e-05,
|
| 59 |
+
"epoch": 0.53,
|
| 60 |
+
"step": 10
|
| 61 |
+
},
|
| 62 |
+
{
|
| 63 |
+
"loss": 0.7008,
|
| 64 |
+
"learning_rate": 7.2e-05,
|
| 65 |
+
"epoch": 0.58,
|
| 66 |
+
"step": 11
|
| 67 |
+
},
|
| 68 |
+
{
|
| 69 |
+
"loss": 0.681,
|
| 70 |
+
"learning_rate": 7.2e-05,
|
| 71 |
+
"epoch": 0.63,
|
| 72 |
+
"step": 12
|
| 73 |
+
},
|
| 74 |
+
{
|
| 75 |
+
"loss": 0.7248,
|
| 76 |
+
"learning_rate": 7.2e-05,
|
| 77 |
+
"epoch": 0.68,
|
| 78 |
+
"step": 13
|
| 79 |
+
},
|
| 80 |
+
{
|
| 81 |
+
"loss": 0.7351,
|
| 82 |
+
"learning_rate": 7.2e-05,
|
| 83 |
+
"epoch": 0.74,
|
| 84 |
+
"step": 14
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"loss": 0.7095,
|
| 88 |
+
"learning_rate": 7.2e-05,
|
| 89 |
+
"epoch": 0.79,
|
| 90 |
+
"step": 15
|
| 91 |
+
},
|
| 92 |
+
{
|
| 93 |
+
"loss": 0.683,
|
| 94 |
+
"learning_rate": 7.2e-05,
|
| 95 |
+
"epoch": 0.84,
|
| 96 |
+
"step": 16
|
| 97 |
+
},
|
| 98 |
+
{
|
| 99 |
+
"loss": 0.6719,
|
| 100 |
+
"learning_rate": 7.2e-05,
|
| 101 |
+
"epoch": 0.89,
|
| 102 |
+
"step": 17
|
| 103 |
+
},
|
| 104 |
+
{
|
| 105 |
+
"loss": 0.6764,
|
| 106 |
+
"learning_rate": 7.2e-05,
|
| 107 |
+
"epoch": 0.95,
|
| 108 |
+
"step": 18
|
| 109 |
+
},
|
| 110 |
+
{
|
| 111 |
+
"loss": 0.6767,
|
| 112 |
+
"learning_rate": 7.2e-05,
|
| 113 |
+
"epoch": 1.0,
|
| 114 |
+
"step": 19
|
| 115 |
+
},
|
| 116 |
+
{
|
| 117 |
+
"loss": 0.6652,
|
| 118 |
+
"learning_rate": 7.2e-05,
|
| 119 |
+
"epoch": 1.05,
|
| 120 |
+
"step": 20
|
| 121 |
+
},
|
| 122 |
+
{
|
| 123 |
+
"loss": 0.6749,
|
| 124 |
+
"learning_rate": 7.2e-05,
|
| 125 |
+
"epoch": 1.11,
|
| 126 |
+
"step": 21
|
| 127 |
+
},
|
| 128 |
+
{
|
| 129 |
+
"loss": 0.6826,
|
| 130 |
+
"learning_rate": 7.2e-05,
|
| 131 |
+
"epoch": 1.16,
|
| 132 |
+
"step": 22
|
| 133 |
+
},
|
| 134 |
+
{
|
| 135 |
+
"loss": 0.6545,
|
| 136 |
+
"learning_rate": 7.2e-05,
|
| 137 |
+
"epoch": 1.21,
|
| 138 |
+
"step": 23
|
| 139 |
+
},
|
| 140 |
+
{
|
| 141 |
+
"loss": 0.6917,
|
| 142 |
+
"learning_rate": 7.2e-05,
|
| 143 |
+
"epoch": 1.26,
|
| 144 |
+
"step": 24
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"loss": 0.65,
|
| 148 |
+
"learning_rate": 7.2e-05,
|
| 149 |
+
"epoch": 1.32,
|
| 150 |
+
"step": 25
|
| 151 |
+
},
|
| 152 |
+
{
|
| 153 |
+
"eval_wrong_arc_loss": 0.6511574387550354,
|
| 154 |
+
"eval_wrong_arc_score": -0.2280832678079605,
|
| 155 |
+
"eval_wrong_arc_brier_score": 0.2280832678079605,
|
| 156 |
+
"eval_wrong_arc_average_probability": 0.5255815386772156,
|
| 157 |
+
"eval_wrong_arc_accuracy": 0.68,
|
| 158 |
+
"eval_wrong_arc_probabilities": [
|
| 159 |
+
0.5341923236846924,
|
| 160 |
+
0.4215836822986603,
|
| 161 |
+
0.5577516555786133,
|
| 162 |
+
0.4472099840641022,
|
| 163 |
+
0.5210587382316589,
|
| 164 |
+
0.48162582516670227,
|
| 165 |
+
0.5153342485427856,
|
| 166 |
+
0.5227084159851074,
|
| 167 |
+
0.5714558959007263,
|
| 168 |
+
0.5944980978965759,
|
| 169 |
+
0.5958520174026489,
|
| 170 |
+
0.4774866998195648,
|
| 171 |
+
0.5097010135650635,
|
| 172 |
+
0.5408040881156921,
|
| 173 |
+
0.5164729356765747,
|
| 174 |
+
0.4862683117389679,
|
| 175 |
+
0.611793577671051,
|
| 176 |
+
0.5587571263313293,
|
| 177 |
+
0.46050599217414856,
|
| 178 |
+
0.5235235691070557,
|
| 179 |
+
0.5262281894683838,
|
| 180 |
+
0.519157350063324,
|
| 181 |
+
0.4868156611919403,
|
| 182 |
+
0.5412039160728455,
|
| 183 |
+
0.5091567039489746,
|
| 184 |
+
0.5481002330780029,
|
| 185 |
+
0.4881518483161926,
|
| 186 |
+
0.5103722810745239,
|
| 187 |
+
0.5413199067115784,
|
| 188 |
+
0.5389612317085266,
|
| 189 |
+
0.5303735733032227,
|
| 190 |
+
0.46720021963119507,
|
| 191 |
+
0.47469261288642883,
|
| 192 |
+
0.5217738747596741,
|
| 193 |
+
0.4839140772819519,
|
| 194 |
+
0.5033125281333923,
|
| 195 |
+
0.6740046739578247,
|
| 196 |
+
0.4932643473148346,
|
| 197 |
+
0.5131027698516846,
|
| 198 |
+
0.5083686113357544,
|
| 199 |
+
0.6749250888824463,
|
| 200 |
+
0.5694271326065063,
|
| 201 |
+
0.5096608996391296,
|
| 202 |
+
0.5113099217414856,
|
| 203 |
+
0.5202792882919312,
|
| 204 |
+
0.570429265499115,
|
| 205 |
+
0.6612235307693481,
|
| 206 |
+
0.4875631034374237,
|
| 207 |
+
0.4891292154788971,
|
| 208 |
+
0.5386453866958618,
|
| 209 |
+
0.39854174852371216,
|
| 210 |
+
0.5061803460121155,
|
| 211 |
+
0.4420945346355438,
|
| 212 |
+
0.49443987011909485,
|
| 213 |
+
0.483050137758255,
|
| 214 |
+
0.5371103882789612,
|
| 215 |
+
0.4931206703186035,
|
| 216 |
+
0.5382444262504578,
|
| 217 |
+
0.47580087184906006,
|
| 218 |
+
0.4962416887283325,
|
| 219 |
+
0.5580680966377258,
|
| 220 |
+
0.5028905868530273,
|
| 221 |
+
0.5332112908363342,
|
| 222 |
+
0.5813270211219788,
|
| 223 |
+
0.5666810274124146,
|
| 224 |
+
0.6028505563735962,
|
| 225 |
+
0.505656361579895,
|
| 226 |
+
0.5044450163841248,
|
| 227 |
+
0.547569215297699,
|
| 228 |
+
0.41055724024772644,
|
| 229 |
+
0.4800181984901428,
|
| 230 |
+
0.5832869410514832,
|
| 231 |
+
0.49764689803123474,
|
| 232 |
+
0.5161015391349792,
|
| 233 |
+
0.598942756652832,
|
| 234 |
+
0.48662498593330383,
|
| 235 |
+
0.5584536194801331,
|
| 236 |
+
0.5778775215148926,
|
| 237 |
+
0.5650496482849121,
|
| 238 |
+
0.4185655415058136,
|
| 239 |
+
0.512199878692627,
|
| 240 |
+
0.4503766596317291,
|
| 241 |
+
0.5686261057853699,
|
| 242 |
+
0.5340624451637268,
|
| 243 |
+
0.4932253956794739,
|
| 244 |
+
0.6768128275871277,
|
| 245 |
+
0.5700348019599915,
|
| 246 |
+
0.575413167476654,
|
| 247 |
+
0.47910740971565247,
|
| 248 |
+
0.5817662477493286,
|
| 249 |
+
0.594521164894104,
|
| 250 |
+
0.43558111786842346,
|
| 251 |
+
0.4789729714393616,
|
| 252 |
+
0.5581913590431213,
|
| 253 |
+
0.544609546661377,
|
| 254 |
+
0.6364085674285889,
|
| 255 |
+
0.5175650119781494,
|
| 256 |
+
0.5233790874481201,
|
| 257 |
+
0.5672008395195007,
|
| 258 |
+
0.43879184126853943
|
| 259 |
+
],
|
| 260 |
+
"eval_wrong_arc_runtime": 7.1003,
|
| 261 |
+
"eval_wrong_arc_samples_per_second": 14.084,
|
| 262 |
+
"eval_wrong_arc_steps_per_second": 0.282,
|
| 263 |
+
"epoch": 1.32,
|
| 264 |
+
"step": 25
|
| 265 |
+
},
|
| 266 |
+
{
|
| 267 |
+
"loss": 0.6138,
|
| 268 |
+
"learning_rate": 7.2e-05,
|
| 269 |
+
"epoch": 1.37,
|
| 270 |
+
"step": 26
|
| 271 |
+
},
|
| 272 |
+
{
|
| 273 |
+
"loss": 0.6443,
|
| 274 |
+
"learning_rate": 7.2e-05,
|
| 275 |
+
"epoch": 1.42,
|
| 276 |
+
"step": 27
|
| 277 |
+
},
|
| 278 |
+
{
|
| 279 |
+
"loss": 0.6672,
|
| 280 |
+
"learning_rate": 7.2e-05,
|
| 281 |
+
"epoch": 1.47,
|
| 282 |
+
"step": 28
|
| 283 |
+
},
|
| 284 |
+
{
|
| 285 |
+
"loss": 0.6312,
|
| 286 |
+
"learning_rate": 7.2e-05,
|
| 287 |
+
"epoch": 1.53,
|
| 288 |
+
"step": 29
|
| 289 |
+
},
|
| 290 |
+
{
|
| 291 |
+
"loss": 0.6603,
|
| 292 |
+
"learning_rate": 7.2e-05,
|
| 293 |
+
"epoch": 1.58,
|
| 294 |
+
"step": 30
|
| 295 |
+
},
|
| 296 |
+
{
|
| 297 |
+
"loss": 0.6145,
|
| 298 |
+
"learning_rate": 7.2e-05,
|
| 299 |
+
"epoch": 1.63,
|
| 300 |
+
"step": 31
|
| 301 |
+
},
|
| 302 |
+
{
|
| 303 |
+
"loss": 0.6683,
|
| 304 |
+
"learning_rate": 7.2e-05,
|
| 305 |
+
"epoch": 1.68,
|
| 306 |
+
"step": 32
|
| 307 |
+
},
|
| 308 |
+
{
|
| 309 |
+
"loss": 0.6027,
|
| 310 |
+
"learning_rate": 7.2e-05,
|
| 311 |
+
"epoch": 1.74,
|
| 312 |
+
"step": 33
|
| 313 |
+
},
|
| 314 |
+
{
|
| 315 |
+
"loss": 0.6043,
|
| 316 |
+
"learning_rate": 7.2e-05,
|
| 317 |
+
"epoch": 1.79,
|
| 318 |
+
"step": 34
|
| 319 |
+
},
|
| 320 |
+
{
|
| 321 |
+
"loss": 0.694,
|
| 322 |
+
"learning_rate": 7.2e-05,
|
| 323 |
+
"epoch": 1.84,
|
| 324 |
+
"step": 35
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"loss": 0.6568,
|
| 328 |
+
"learning_rate": 7.2e-05,
|
| 329 |
+
"epoch": 1.89,
|
| 330 |
+
"step": 36
|
| 331 |
+
},
|
| 332 |
+
{
|
| 333 |
+
"loss": 0.6639,
|
| 334 |
+
"learning_rate": 7.2e-05,
|
| 335 |
+
"epoch": 1.95,
|
| 336 |
+
"step": 37
|
| 337 |
+
},
|
| 338 |
+
{
|
| 339 |
+
"loss": 0.5993,
|
| 340 |
+
"learning_rate": 7.2e-05,
|
| 341 |
+
"epoch": 2.0,
|
| 342 |
+
"step": 38
|
| 343 |
+
},
|
| 344 |
+
{
|
| 345 |
+
"loss": 0.5717,
|
| 346 |
+
"learning_rate": 7.2e-05,
|
| 347 |
+
"epoch": 2.05,
|
| 348 |
+
"step": 39
|
| 349 |
+
},
|
| 350 |
+
{
|
| 351 |
+
"loss": 0.6159,
|
| 352 |
+
"learning_rate": 7.2e-05,
|
| 353 |
+
"epoch": 2.11,
|
| 354 |
+
"step": 40
|
| 355 |
+
},
|
| 356 |
+
{
|
| 357 |
+
"loss": 0.591,
|
| 358 |
+
"learning_rate": 7.2e-05,
|
| 359 |
+
"epoch": 2.16,
|
| 360 |
+
"step": 41
|
| 361 |
+
},
|
| 362 |
+
{
|
| 363 |
+
"loss": 0.5329,
|
| 364 |
+
"learning_rate": 7.2e-05,
|
| 365 |
+
"epoch": 2.21,
|
| 366 |
+
"step": 42
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"loss": 0.5347,
|
| 370 |
+
"learning_rate": 7.2e-05,
|
| 371 |
+
"epoch": 2.26,
|
| 372 |
+
"step": 43
|
| 373 |
+
},
|
| 374 |
+
{
|
| 375 |
+
"loss": 0.5587,
|
| 376 |
+
"learning_rate": 7.2e-05,
|
| 377 |
+
"epoch": 2.32,
|
| 378 |
+
"step": 44
|
| 379 |
+
},
|
| 380 |
+
{
|
| 381 |
+
"loss": 0.5215,
|
| 382 |
+
"learning_rate": 7.2e-05,
|
| 383 |
+
"epoch": 2.37,
|
| 384 |
+
"step": 45
|
| 385 |
+
},
|
| 386 |
+
{
|
| 387 |
+
"loss": 0.5184,
|
| 388 |
+
"learning_rate": 7.2e-05,
|
| 389 |
+
"epoch": 2.42,
|
| 390 |
+
"step": 46
|
| 391 |
+
},
|
| 392 |
+
{
|
| 393 |
+
"loss": 0.5284,
|
| 394 |
+
"learning_rate": 7.2e-05,
|
| 395 |
+
"epoch": 2.47,
|
| 396 |
+
"step": 47
|
| 397 |
+
},
|
| 398 |
+
{
|
| 399 |
+
"loss": 0.491,
|
| 400 |
+
"learning_rate": 7.2e-05,
|
| 401 |
+
"epoch": 2.53,
|
| 402 |
+
"step": 48
|
| 403 |
+
},
|
| 404 |
+
{
|
| 405 |
+
"loss": 0.482,
|
| 406 |
+
"learning_rate": 7.2e-05,
|
| 407 |
+
"epoch": 2.58,
|
| 408 |
+
"step": 49
|
| 409 |
+
},
|
| 410 |
+
{
|
| 411 |
+
"loss": 0.486,
|
| 412 |
+
"learning_rate": 7.2e-05,
|
| 413 |
+
"epoch": 2.63,
|
| 414 |
+
"step": 50
|
| 415 |
+
},
|
| 416 |
+
{
|
| 417 |
+
"eval_wrong_arc_loss": 0.48614898324012756,
|
| 418 |
+
"eval_wrong_arc_score": -0.15555404126644135,
|
| 419 |
+
"eval_wrong_arc_brier_score": 0.15555404126644135,
|
| 420 |
+
"eval_wrong_arc_average_probability": 0.6473783850669861,
|
| 421 |
+
"eval_wrong_arc_accuracy": 0.79,
|
| 422 |
+
"eval_wrong_arc_probabilities": [
|
| 423 |
+
0.6561150550842285,
|
| 424 |
+
0.4626675546169281,
|
| 425 |
+
0.6256658434867859,
|
| 426 |
+
0.8533488512039185,
|
| 427 |
+
0.5858703851699829,
|
| 428 |
+
0.6902235150337219,
|
| 429 |
+
0.6159930229187012,
|
| 430 |
+
0.7639041543006897,
|
| 431 |
+
0.8640973567962646,
|
| 432 |
+
0.772244930267334,
|
| 433 |
+
0.8565906882286072,
|
| 434 |
+
0.29203683137893677,
|
| 435 |
+
0.5727916955947876,
|
| 436 |
+
0.8369653820991516,
|
| 437 |
+
0.7363625764846802,
|
| 438 |
+
0.44277945160865784,
|
| 439 |
+
0.8875776529312134,
|
| 440 |
+
0.40449512004852295,
|
| 441 |
+
0.592871904373169,
|
| 442 |
+
0.7689143419265747,
|
| 443 |
+
0.5688448548316956,
|
| 444 |
+
0.8267202377319336,
|
| 445 |
+
0.5138616561889648,
|
| 446 |
+
0.806187093257904,
|
| 447 |
+
0.6455444097518921,
|
| 448 |
+
0.8391571640968323,
|
| 449 |
+
0.36543235182762146,
|
| 450 |
+
0.5734689235687256,
|
| 451 |
+
0.747251033782959,
|
| 452 |
+
0.6627150774002075,
|
| 453 |
+
0.6402449607849121,
|
| 454 |
+
0.6396260857582092,
|
| 455 |
+
0.6404427886009216,
|
| 456 |
+
0.6774603724479675,
|
| 457 |
+
0.3316854238510132,
|
| 458 |
+
0.6492359042167664,
|
| 459 |
+
0.9437365531921387,
|
| 460 |
+
0.5006998181343079,
|
| 461 |
+
0.6551035642623901,
|
| 462 |
+
0.5634311437606812,
|
| 463 |
+
0.9296560287475586,
|
| 464 |
+
0.8223538398742676,
|
| 465 |
+
0.6447325944900513,
|
| 466 |
+
0.4997992515563965,
|
| 467 |
+
0.8213691115379333,
|
| 468 |
+
0.3131840229034424,
|
| 469 |
+
0.9470311999320984,
|
| 470 |
+
0.30666762590408325,
|
| 471 |
+
0.4539206027984619,
|
| 472 |
+
0.5038973689079285,
|
| 473 |
+
0.23626333475112915,
|
| 474 |
+
0.468450129032135,
|
| 475 |
+
0.62051922082901,
|
| 476 |
+
0.754213809967041,
|
| 477 |
+
0.6286126375198364,
|
| 478 |
+
0.6793380975723267,
|
| 479 |
+
0.4252707064151764,
|
| 480 |
+
0.8759824633598328,
|
| 481 |
+
0.28135573863983154,
|
| 482 |
+
0.5500273108482361,
|
| 483 |
+
0.8892378807067871,
|
| 484 |
+
0.4467567503452301,
|
| 485 |
+
0.6596640944480896,
|
| 486 |
+
0.816720724105835,
|
| 487 |
+
0.7590842247009277,
|
| 488 |
+
0.9068907499313354,
|
| 489 |
+
0.7511820197105408,
|
| 490 |
+
0.640140175819397,
|
| 491 |
+
0.7239473462104797,
|
| 492 |
+
0.26979541778564453,
|
| 493 |
+
0.8412765264511108,
|
| 494 |
+
0.7638338208198547,
|
| 495 |
+
0.7051005363464355,
|
| 496 |
+
0.5634142160415649,
|
| 497 |
+
0.6077967286109924,
|
| 498 |
+
0.44273415207862854,
|
| 499 |
+
0.5198932886123657,
|
| 500 |
+
0.8605012893676758,
|
| 501 |
+
0.8117179274559021,
|
| 502 |
+
0.3314230144023895,
|
| 503 |
+
0.5430377125740051,
|
| 504 |
+
0.6297375559806824,
|
| 505 |
+
0.8171581625938416,
|
| 506 |
+
0.6848751306533813,
|
| 507 |
+
0.5711643099784851,
|
| 508 |
+
0.8888359665870667,
|
| 509 |
+
0.7509227991104126,
|
| 510 |
+
0.6692346930503845,
|
| 511 |
+
0.48311302065849304,
|
| 512 |
+
0.7981626987457275,
|
| 513 |
+
0.7328034043312073,
|
| 514 |
+
0.342570424079895,
|
| 515 |
+
0.638855516910553,
|
| 516 |
+
0.8649654388427734,
|
| 517 |
+
0.7085446715354919,
|
| 518 |
+
0.7994809150695801,
|
| 519 |
+
0.6453239917755127,
|
| 520 |
+
0.3845190703868866,
|
| 521 |
+
0.8838135004043579,
|
| 522 |
+
0.7545994520187378
|
| 523 |
+
],
|
| 524 |
+
"eval_wrong_arc_runtime": 7.0912,
|
| 525 |
+
"eval_wrong_arc_samples_per_second": 14.102,
|
| 526 |
+
"eval_wrong_arc_steps_per_second": 0.282,
|
| 527 |
+
"epoch": 2.63,
|
| 528 |
+
"step": 50
|
| 529 |
+
},
|
| 530 |
+
{
|
| 531 |
+
"loss": 0.5629,
|
| 532 |
+
"learning_rate": 7.2e-05,
|
| 533 |
+
"epoch": 2.68,
|
| 534 |
+
"step": 51
|
| 535 |
+
},
|
| 536 |
+
{
|
| 537 |
+
"loss": 0.4267,
|
| 538 |
+
"learning_rate": 7.2e-05,
|
| 539 |
+
"epoch": 2.74,
|
| 540 |
+
"step": 52
|
| 541 |
+
},
|
| 542 |
+
{
|
| 543 |
+
"loss": 0.4321,
|
| 544 |
+
"learning_rate": 7.2e-05,
|
| 545 |
+
"epoch": 2.79,
|
| 546 |
+
"step": 53
|
| 547 |
+
},
|
| 548 |
+
{
|
| 549 |
+
"loss": 0.4345,
|
| 550 |
+
"learning_rate": 7.2e-05,
|
| 551 |
+
"epoch": 2.84,
|
| 552 |
+
"step": 54
|
| 553 |
+
},
|
| 554 |
+
{
|
| 555 |
+
"loss": 0.5663,
|
| 556 |
+
"learning_rate": 7.2e-05,
|
| 557 |
+
"epoch": 2.89,
|
| 558 |
+
"step": 55
|
| 559 |
+
},
|
| 560 |
+
{
|
| 561 |
+
"loss": 0.3682,
|
| 562 |
+
"learning_rate": 7.2e-05,
|
| 563 |
+
"epoch": 2.95,
|
| 564 |
+
"step": 56
|
| 565 |
+
},
|
| 566 |
+
{
|
| 567 |
+
"loss": 0.2817,
|
| 568 |
+
"learning_rate": 7.2e-05,
|
| 569 |
+
"epoch": 3.0,
|
| 570 |
+
"step": 57
|
| 571 |
+
},
|
| 572 |
+
{
|
| 573 |
+
"loss": 0.2573,
|
| 574 |
+
"learning_rate": 7.2e-05,
|
| 575 |
+
"epoch": 3.05,
|
| 576 |
+
"step": 58
|
| 577 |
+
},
|
| 578 |
+
{
|
| 579 |
+
"loss": 0.3374,
|
| 580 |
+
"learning_rate": 7.2e-05,
|
| 581 |
+
"epoch": 3.11,
|
| 582 |
+
"step": 59
|
| 583 |
+
},
|
| 584 |
+
{
|
| 585 |
+
"loss": 0.3706,
|
| 586 |
+
"learning_rate": 7.2e-05,
|
| 587 |
+
"epoch": 3.16,
|
| 588 |
+
"step": 60
|
| 589 |
+
},
|
| 590 |
+
{
|
| 591 |
+
"loss": 0.4201,
|
| 592 |
+
"learning_rate": 7.2e-05,
|
| 593 |
+
"epoch": 3.21,
|
| 594 |
+
"step": 61
|
| 595 |
+
},
|
| 596 |
+
{
|
| 597 |
+
"loss": 0.3508,
|
| 598 |
+
"learning_rate": 7.2e-05,
|
| 599 |
+
"epoch": 3.26,
|
| 600 |
+
"step": 62
|
| 601 |
+
},
|
| 602 |
+
{
|
| 603 |
+
"loss": 0.3724,
|
| 604 |
+
"learning_rate": 7.2e-05,
|
| 605 |
+
"epoch": 3.32,
|
| 606 |
+
"step": 63
|
| 607 |
+
},
|
| 608 |
+
{
|
| 609 |
+
"loss": 0.2794,
|
| 610 |
+
"learning_rate": 7.2e-05,
|
| 611 |
+
"epoch": 3.37,
|
| 612 |
+
"step": 64
|
| 613 |
+
},
|
| 614 |
+
{
|
| 615 |
+
"loss": 0.2064,
|
| 616 |
+
"learning_rate": 7.2e-05,
|
| 617 |
+
"epoch": 3.42,
|
| 618 |
+
"step": 65
|
| 619 |
+
},
|
| 620 |
+
{
|
| 621 |
+
"loss": 0.202,
|
| 622 |
+
"learning_rate": 7.2e-05,
|
| 623 |
+
"epoch": 3.47,
|
| 624 |
+
"step": 66
|
| 625 |
+
},
|
| 626 |
+
{
|
| 627 |
+
"loss": 0.2556,
|
| 628 |
+
"learning_rate": 7.2e-05,
|
| 629 |
+
"epoch": 3.53,
|
| 630 |
+
"step": 67
|
| 631 |
+
},
|
| 632 |
+
{
|
| 633 |
+
"loss": 0.191,
|
| 634 |
+
"learning_rate": 7.2e-05,
|
| 635 |
+
"epoch": 3.58,
|
| 636 |
+
"step": 68
|
| 637 |
+
},
|
| 638 |
+
{
|
| 639 |
+
"loss": 0.5198,
|
| 640 |
+
"learning_rate": 7.2e-05,
|
| 641 |
+
"epoch": 3.63,
|
| 642 |
+
"step": 69
|
| 643 |
+
},
|
| 644 |
+
{
|
| 645 |
+
"loss": 0.1593,
|
| 646 |
+
"learning_rate": 7.2e-05,
|
| 647 |
+
"epoch": 3.68,
|
| 648 |
+
"step": 70
|
| 649 |
+
},
|
| 650 |
+
{
|
| 651 |
+
"loss": 0.2818,
|
| 652 |
+
"learning_rate": 7.2e-05,
|
| 653 |
+
"epoch": 3.74,
|
| 654 |
+
"step": 71
|
| 655 |
+
},
|
| 656 |
+
{
|
| 657 |
+
"loss": 0.1736,
|
| 658 |
+
"learning_rate": 7.2e-05,
|
| 659 |
+
"epoch": 3.79,
|
| 660 |
+
"step": 72
|
| 661 |
+
},
|
| 662 |
+
{
|
| 663 |
+
"loss": 0.2594,
|
| 664 |
+
"learning_rate": 7.2e-05,
|
| 665 |
+
"epoch": 3.84,
|
| 666 |
+
"step": 73
|
| 667 |
+
},
|
| 668 |
+
{
|
| 669 |
+
"loss": 0.3317,
|
| 670 |
+
"learning_rate": 7.2e-05,
|
| 671 |
+
"epoch": 3.89,
|
| 672 |
+
"step": 74
|
| 673 |
+
},
|
| 674 |
+
{
|
| 675 |
+
"loss": 0.1084,
|
| 676 |
+
"learning_rate": 7.2e-05,
|
| 677 |
+
"epoch": 3.95,
|
| 678 |
+
"step": 75
|
| 679 |
+
},
|
| 680 |
+
{
|
| 681 |
+
"eval_wrong_arc_loss": 0.48962050676345825,
|
| 682 |
+
"eval_wrong_arc_score": -0.1449090540409088,
|
| 683 |
+
"eval_wrong_arc_brier_score": 0.1449090540409088,
|
| 684 |
+
"eval_wrong_arc_average_probability": 0.7831840515136719,
|
| 685 |
+
"eval_wrong_arc_accuracy": 0.8,
|
| 686 |
+
"eval_wrong_arc_probabilities": [
|
| 687 |
+
0.9962900876998901,
|
| 688 |
+
0.7803617715835571,
|
| 689 |
+
0.5364444851875305,
|
| 690 |
+
0.9998961687088013,
|
| 691 |
+
0.8887804746627808,
|
| 692 |
+
0.9976217150688171,
|
| 693 |
+
0.8334865570068359,
|
| 694 |
+
0.9998819828033447,
|
| 695 |
+
0.9999058246612549,
|
| 696 |
+
0.9978801012039185,
|
| 697 |
+
0.9933562278747559,
|
| 698 |
+
0.1421491801738739,
|
| 699 |
+
0.8300302624702454,
|
| 700 |
+
0.9992189407348633,
|
| 701 |
+
0.9899032115936279,
|
| 702 |
+
0.21271571516990662,
|
| 703 |
+
0.9991642236709595,
|
| 704 |
+
0.014693806879222393,
|
| 705 |
+
0.9953280687332153,
|
| 706 |
+
0.9968209266662598,
|
| 707 |
+
0.199941948056221,
|
| 708 |
+
0.9912959337234497,
|
| 709 |
+
0.9270946979522705,
|
| 710 |
+
0.9991445541381836,
|
| 711 |
+
0.8223531246185303,
|
| 712 |
+
0.9992208480834961,
|
| 713 |
+
0.2905525267124176,
|
| 714 |
+
0.7823147773742676,
|
| 715 |
+
0.9986461997032166,
|
| 716 |
+
0.9961987137794495,
|
| 717 |
+
0.973800778388977,
|
| 718 |
+
0.896918535232544,
|
| 719 |
+
0.9690271019935608,
|
| 720 |
+
0.5849995613098145,
|
| 721 |
+
0.07284952700138092,
|
| 722 |
+
0.9723120927810669,
|
| 723 |
+
0.999985933303833,
|
| 724 |
+
0.9009866118431091,
|
| 725 |
+
0.9817610383033752,
|
| 726 |
+
0.5555633306503296,
|
| 727 |
+
0.9998956918716431,
|
| 728 |
+
0.9909346699714661,
|
| 729 |
+
0.9936639070510864,
|
| 730 |
+
0.4922553598880768,
|
| 731 |
+
0.9994981288909912,
|
| 732 |
+
0.004467986058443785,
|
| 733 |
+
0.9999967813491821,
|
| 734 |
+
0.013173140585422516,
|
| 735 |
+
0.3301887512207031,
|
| 736 |
+
0.4937175214290619,
|
| 737 |
+
0.15117959678173065,
|
| 738 |
+
0.6426385641098022,
|
| 739 |
+
0.9926222562789917,
|
| 740 |
+
0.9969263672828674,
|
| 741 |
+
0.9407752752304077,
|
| 742 |
+
0.7943993210792542,
|
| 743 |
+
0.28421616554260254,
|
| 744 |
+
0.9991441965103149,
|
| 745 |
+
0.015714088454842567,
|
| 746 |
+
0.9596678018569946,
|
| 747 |
+
0.9966815114021301,
|
| 748 |
+
0.4579419493675232,
|
| 749 |
+
0.9928516745567322,
|
| 750 |
+
0.9990314245223999,
|
| 751 |
+
0.9978936314582825,
|
| 752 |
+
0.9997265934944153,
|
| 753 |
+
0.9977813363075256,
|
| 754 |
+
0.8815402388572693,
|
| 755 |
+
0.9994369149208069,
|
| 756 |
+
0.40593746304512024,
|
| 757 |
+
0.9994971752166748,
|
| 758 |
+
0.9803721308708191,
|
| 759 |
+
0.9809514284133911,
|
| 760 |
+
0.7131913304328918,
|
| 761 |
+
0.802761435508728,
|
| 762 |
+
0.5336641669273376,
|
| 763 |
+
0.6622551083564758,
|
| 764 |
+
0.9991828799247742,
|
| 765 |
+
0.9978545308113098,
|
| 766 |
+
0.07150473445653915,
|
| 767 |
+
0.7831286191940308,
|
| 768 |
+
0.9591096639633179,
|
| 769 |
+
0.9961942434310913,
|
| 770 |
+
0.9644544720649719,
|
| 771 |
+
0.5869894623756409,
|
| 772 |
+
0.9993104934692383,
|
| 773 |
+
0.9948341846466064,
|
| 774 |
+
0.8777645826339722,
|
| 775 |
+
0.4935871660709381,
|
| 776 |
+
0.996717631816864,
|
| 777 |
+
0.9972440004348755,
|
| 778 |
+
0.19612468779087067,
|
| 779 |
+
0.9670541882514954,
|
| 780 |
+
0.9996756315231323,
|
| 781 |
+
0.9910316467285156,
|
| 782 |
+
0.941999077796936,
|
| 783 |
+
0.839278519153595,
|
| 784 |
+
0.05020187422633171,
|
| 785 |
+
0.9999673366546631,
|
| 786 |
+
0.9996931552886963
|
| 787 |
+
],
|
| 788 |
+
"eval_wrong_arc_runtime": 7.0918,
|
| 789 |
+
"eval_wrong_arc_samples_per_second": 14.101,
|
| 790 |
+
"eval_wrong_arc_steps_per_second": 0.282,
|
| 791 |
+
"epoch": 3.95,
|
| 792 |
+
"step": 75
|
| 793 |
+
},
|
| 794 |
+
{
|
| 795 |
+
"loss": 0.1467,
|
| 796 |
+
"learning_rate": 7.2e-05,
|
| 797 |
+
"epoch": 4.0,
|
| 798 |
+
"step": 76
|
| 799 |
+
},
|
| 800 |
+
{
|
| 801 |
+
"loss": 0.1119,
|
| 802 |
+
"learning_rate": 7.2e-05,
|
| 803 |
+
"epoch": 4.05,
|
| 804 |
+
"step": 77
|
| 805 |
+
},
|
| 806 |
+
{
|
| 807 |
+
"loss": 0.1514,
|
| 808 |
+
"learning_rate": 7.2e-05,
|
| 809 |
+
"epoch": 4.11,
|
| 810 |
+
"step": 78
|
| 811 |
+
},
|
| 812 |
+
{
|
| 813 |
+
"loss": 0.1085,
|
| 814 |
+
"learning_rate": 7.2e-05,
|
| 815 |
+
"epoch": 4.16,
|
| 816 |
+
"step": 79
|
| 817 |
+
},
|
| 818 |
+
{
|
| 819 |
+
"loss": 0.1115,
|
| 820 |
+
"learning_rate": 7.2e-05,
|
| 821 |
+
"epoch": 4.21,
|
| 822 |
+
"step": 80
|
| 823 |
+
},
|
| 824 |
+
{
|
| 825 |
+
"loss": 0.0831,
|
| 826 |
+
"learning_rate": 7.2e-05,
|
| 827 |
+
"epoch": 4.26,
|
| 828 |
+
"step": 81
|
| 829 |
+
},
|
| 830 |
+
{
|
| 831 |
+
"loss": 0.0891,
|
| 832 |
+
"learning_rate": 7.2e-05,
|
| 833 |
+
"epoch": 4.32,
|
| 834 |
+
"step": 82
|
| 835 |
+
},
|
| 836 |
+
{
|
| 837 |
+
"loss": 0.0702,
|
| 838 |
+
"learning_rate": 7.2e-05,
|
| 839 |
+
"epoch": 4.37,
|
| 840 |
+
"step": 83
|
| 841 |
+
},
|
| 842 |
+
{
|
| 843 |
+
"loss": 0.0847,
|
| 844 |
+
"learning_rate": 7.2e-05,
|
| 845 |
+
"epoch": 4.42,
|
| 846 |
+
"step": 84
|
| 847 |
+
},
|
| 848 |
+
{
|
| 849 |
+
"loss": 0.1381,
|
| 850 |
+
"learning_rate": 7.2e-05,
|
| 851 |
+
"epoch": 4.47,
|
| 852 |
+
"step": 85
|
| 853 |
+
},
|
| 854 |
+
{
|
| 855 |
+
"loss": 0.1193,
|
| 856 |
+
"learning_rate": 7.2e-05,
|
| 857 |
+
"epoch": 4.53,
|
| 858 |
+
"step": 86
|
| 859 |
+
},
|
| 860 |
+
{
|
| 861 |
+
"loss": 0.1743,
|
| 862 |
+
"learning_rate": 7.2e-05,
|
| 863 |
+
"epoch": 4.58,
|
| 864 |
+
"step": 87
|
| 865 |
+
},
|
| 866 |
+
{
|
| 867 |
+
"loss": 0.1424,
|
| 868 |
+
"learning_rate": 7.2e-05,
|
| 869 |
+
"epoch": 4.63,
|
| 870 |
+
"step": 88
|
| 871 |
+
},
|
| 872 |
+
{
|
| 873 |
+
"loss": 0.1921,
|
| 874 |
+
"learning_rate": 7.2e-05,
|
| 875 |
+
"epoch": 4.68,
|
| 876 |
+
"step": 89
|
| 877 |
+
},
|
| 878 |
+
{
|
| 879 |
+
"loss": 0.291,
|
| 880 |
+
"learning_rate": 7.2e-05,
|
| 881 |
+
"epoch": 4.74,
|
| 882 |
+
"step": 90
|
| 883 |
+
},
|
| 884 |
+
{
|
| 885 |
+
"loss": 0.0726,
|
| 886 |
+
"learning_rate": 7.2e-05,
|
| 887 |
+
"epoch": 4.79,
|
| 888 |
+
"step": 91
|
| 889 |
+
},
|
| 890 |
+
{
|
| 891 |
+
"loss": 0.1394,
|
| 892 |
+
"learning_rate": 7.2e-05,
|
| 893 |
+
"epoch": 4.84,
|
| 894 |
+
"step": 92
|
| 895 |
+
},
|
| 896 |
+
{
|
| 897 |
+
"loss": 0.0854,
|
| 898 |
+
"learning_rate": 7.2e-05,
|
| 899 |
+
"epoch": 4.89,
|
| 900 |
+
"step": 93
|
| 901 |
+
},
|
| 902 |
+
{
|
| 903 |
+
"loss": 0.03,
|
| 904 |
+
"learning_rate": 7.2e-05,
|
| 905 |
+
"epoch": 4.95,
|
| 906 |
+
"step": 94
|
| 907 |
+
},
|
| 908 |
+
{
|
| 909 |
+
"loss": 0.0593,
|
| 910 |
+
"learning_rate": 7.2e-05,
|
| 911 |
+
"epoch": 5.0,
|
| 912 |
+
"step": 95
|
| 913 |
+
},
|
| 914 |
+
{
|
| 915 |
+
"loss": 0.0627,
|
| 916 |
+
"learning_rate": 7.2e-05,
|
| 917 |
+
"epoch": 5.05,
|
| 918 |
+
"step": 96
|
| 919 |
+
},
|
| 920 |
+
{
|
| 921 |
+
"loss": 0.0182,
|
| 922 |
+
"learning_rate": 7.2e-05,
|
| 923 |
+
"epoch": 5.11,
|
| 924 |
+
"step": 97
|
| 925 |
+
},
|
| 926 |
+
{
|
| 927 |
+
"loss": 0.0373,
|
| 928 |
+
"learning_rate": 7.2e-05,
|
| 929 |
+
"epoch": 5.16,
|
| 930 |
+
"step": 98
|
| 931 |
+
},
|
| 932 |
+
{
|
| 933 |
+
"loss": 0.0205,
|
| 934 |
+
"learning_rate": 7.2e-05,
|
| 935 |
+
"epoch": 5.21,
|
| 936 |
+
"step": 99
|
| 937 |
+
},
|
| 938 |
+
{
|
| 939 |
+
"loss": 0.0142,
|
| 940 |
+
"learning_rate": 7.2e-05,
|
| 941 |
+
"epoch": 5.26,
|
| 942 |
+
"step": 100
|
| 943 |
+
},
|
| 944 |
+
{
|
| 945 |
+
"eval_wrong_arc_loss": 0.7816460132598877,
|
| 946 |
+
"eval_wrong_arc_score": -0.16116684675216675,
|
| 947 |
+
"eval_wrong_arc_brier_score": 0.16116684675216675,
|
| 948 |
+
"eval_wrong_arc_average_probability": 0.7979229688644409,
|
| 949 |
+
"eval_wrong_arc_accuracy": 0.81,
|
| 950 |
+
"eval_wrong_arc_probabilities": [
|
| 951 |
+
0.999994158744812,
|
| 952 |
+
0.7483964562416077,
|
| 953 |
+
0.3130388855934143,
|
| 954 |
+
0.9999998807907104,
|
| 955 |
+
0.9902498126029968,
|
| 956 |
+
0.9998146891593933,
|
| 957 |
+
0.9767554402351379,
|
| 958 |
+
1.0,
|
| 959 |
+
0.9999997615814209,
|
| 960 |
+
0.9999923706054688,
|
| 961 |
+
0.999737560749054,
|
| 962 |
+
0.027735406532883644,
|
| 963 |
+
0.9908527731895447,
|
| 964 |
+
0.9999980926513672,
|
| 965 |
+
0.99986732006073,
|
| 966 |
+
0.025411054491996765,
|
| 967 |
+
0.9999935626983643,
|
| 968 |
+
3.376677705091424e-05,
|
| 969 |
+
0.9994077682495117,
|
| 970 |
+
0.9985150694847107,
|
| 971 |
+
0.021305063739418983,
|
| 972 |
+
0.9998865127563477,
|
| 973 |
+
0.9992591738700867,
|
| 974 |
+
0.9999784231185913,
|
| 975 |
+
0.9170340895652771,
|
| 976 |
+
0.9999704360961914,
|
| 977 |
+
0.07415499538183212,
|
| 978 |
+
0.8610795736312866,
|
| 979 |
+
0.9999791383743286,
|
| 980 |
+
0.9999957084655762,
|
| 981 |
+
0.999575674533844,
|
| 982 |
+
0.9272408485412598,
|
| 983 |
+
0.9736331701278687,
|
| 984 |
+
0.8866708278656006,
|
| 985 |
+
0.027791421860456467,
|
| 986 |
+
0.999729573726654,
|
| 987 |
+
1.0,
|
| 988 |
+
0.9970165491104126,
|
| 989 |
+
0.9990842342376709,
|
| 990 |
+
0.5740329027175903,
|
| 991 |
+
0.9999998807907104,
|
| 992 |
+
0.9968683123588562,
|
| 993 |
+
0.999996542930603,
|
| 994 |
+
0.5142654776573181,
|
| 995 |
+
0.9999994039535522,
|
| 996 |
+
4.652805728255771e-05,
|
| 997 |
+
1.0,
|
| 998 |
+
0.0004388771194498986,
|
| 999 |
+
0.23682957887649536,
|
| 1000 |
+
0.42817360162734985,
|
| 1001 |
+
0.21864689886569977,
|
| 1002 |
+
0.9052879214286804,
|
| 1003 |
+
0.999488353729248,
|
| 1004 |
+
0.9999974966049194,
|
| 1005 |
+
0.9822047352790833,
|
| 1006 |
+
0.9483726024627686,
|
| 1007 |
+
0.3384818732738495,
|
| 1008 |
+
0.999998927116394,
|
| 1009 |
+
0.0001653187209740281,
|
| 1010 |
+
0.9996745586395264,
|
| 1011 |
+
0.9999986886978149,
|
| 1012 |
+
0.5172927379608154,
|
| 1013 |
+
0.9980478286743164,
|
| 1014 |
+
0.9999990463256836,
|
| 1015 |
+
0.9999717473983765,
|
| 1016 |
+
0.9999998807907104,
|
| 1017 |
+
0.9999914169311523,
|
| 1018 |
+
0.9598847031593323,
|
| 1019 |
+
0.9999996423721313,
|
| 1020 |
+
0.16434389352798462,
|
| 1021 |
+
0.9999890327453613,
|
| 1022 |
+
0.9985514283180237,
|
| 1023 |
+
0.9981107711791992,
|
| 1024 |
+
0.6783352494239807,
|
| 1025 |
+
0.9952026605606079,
|
| 1026 |
+
0.37246665358543396,
|
| 1027 |
+
0.847752034664154,
|
| 1028 |
+
0.9999322891235352,
|
| 1029 |
+
0.9999262094497681,
|
| 1030 |
+
0.020832795649766922,
|
| 1031 |
+
0.8913257718086243,
|
| 1032 |
+
0.9978765249252319,
|
| 1033 |
+
0.9999139308929443,
|
| 1034 |
+
0.9987577199935913,
|
| 1035 |
+
0.881140947341919,
|
| 1036 |
+
0.9999871253967285,
|
| 1037 |
+
0.9999966621398926,
|
| 1038 |
+
0.997133731842041,
|
| 1039 |
+
0.529548168182373,
|
| 1040 |
+
0.9999359846115112,
|
| 1041 |
+
0.9999946355819702,
|
| 1042 |
+
0.08537785708904266,
|
| 1043 |
+
0.9985488057136536,
|
| 1044 |
+
0.9999998807907104,
|
| 1045 |
+
0.9999691247940063,
|
| 1046 |
+
0.9985161423683167,
|
| 1047 |
+
0.9663605690002441,
|
| 1048 |
+
0.0011356659233570099,
|
| 1049 |
+
0.9999998807907104,
|
| 1050 |
+
0.9999998807907104
|
| 1051 |
+
],
|
| 1052 |
+
"eval_wrong_arc_runtime": 7.0875,
|
| 1053 |
+
"eval_wrong_arc_samples_per_second": 14.109,
|
| 1054 |
+
"eval_wrong_arc_steps_per_second": 0.282,
|
| 1055 |
+
"epoch": 5.26,
|
| 1056 |
+
"step": 100
|
| 1057 |
+
},
|
| 1058 |
+
{
|
| 1059 |
+
"train_runtime": 554.4056,
|
| 1060 |
+
"train_samples_per_second": 5.772,
|
| 1061 |
+
"train_steps_per_second": 0.18,
|
| 1062 |
+
"total_flos": 0.0,
|
| 1063 |
+
"train_loss": 0.4284818640910089,
|
| 1064 |
+
"epoch": 5.26,
|
| 1065 |
+
"step": 100
|
| 1066 |
+
}
|
| 1067 |
+
]
|