Hugging Face's logo

genies-models
/
llama-13b-creative_writing

PEFT
Model card Files Community
joshuaclymer commited on
Commit
9066c9a
•
1 Parent(s): 77128d6

Upload folder using huggingface_hub

Browse files
Files changed (10) hide show
  1. README.md +21 -0
  2. adapter_config.json +21 -0
  3. adapter_model.bin +3 -0
  4. special_tokens_map.json +24 -0
  5. tokenizer.json +0 -0
  6. tokenizer.model +3 -0
  7. tokenizer_config.json +32 -0
  8. train_args.json +123 -0
  9. training_args.bin +3 -0
  10. training_logs.json +2279 -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/llama-13b",
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": 0.0,
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:adbcd0f989e2e424af22ad5e566396749e5d340da9e3506f616bc162c3292248
3
+ size 209811921
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 sha256:9e556afd44213b6bd1be2b850ebbbd98f5481437a8021afaf58ee7fb1818d347
3
+ size 499723
tokenizer_config.json ADDED
@@ -0,0 +1,32 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
+ "model_max_length": 2048,
20
+ "pad_token": null,
21
+ "sp_model_kwargs": {},
22
+ "tokenizer_class": "LlamaTokenizer",
23
+ "unk_token": {
24
+ "__type": "AddedToken",
25
+ "content": "<unk>",
26
+ "lstrip": false,
27
+ "normalized": true,
28
+ "rstrip": false,
29
+ "single_word": false
30
+ },
31
+ "use_default_system_prompt": true
32
+ }
train_args.json ADDED
@@ -0,0 +1,123 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "initial_model_dir": "models/llama-13b",
3
+ "distribution_id": "creative_writing",
4
+ "date_trained": "10/10/2023 09:12:52",
5
+ "output_dir": "models/classify_lora/llama-13b-creative_writing",
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": 16,
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": 0.0002,
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": 4,
26
+ "max_steps": 150,
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/llama-13b-creative_writing/runs/Oct10_08-06-57_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": 0,
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-llama-13b-creative_writing",
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_creative_writing_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:a1d8b9ae2ab339ef7947e5d5b9b3de710bc194de559d8b441df4de24addb2941
3
+ size 5115
training_logs.json ADDED
@@ -0,0 +1,2279 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ [
2
+ {
3
+ "loss": 0.8734,
4
+ "learning_rate": 0.0002,
5
+ "epoch": 0.03,
6
+ "step": 1
7
+ },
8
+ {
9
+ "loss": 0.7825,
10
+ "learning_rate": 0.0002,
11
+ "epoch": 0.05,
12
+ "step": 2
13
+ },
14
+ {
15
+ "loss": 0.6849,
16
+ "learning_rate": 0.0002,
17
+ "epoch": 0.08,
18
+ "step": 3
19
+ },
20
+ {
21
+ "loss": 0.6394,
22
+ "learning_rate": 0.0002,
23
+ "epoch": 0.11,
24
+ "step": 4
25
+ },
26
+ {
27
+ "loss": 0.5604,
28
+ "learning_rate": 0.0002,
29
+ "epoch": 0.13,
30
+ "step": 5
31
+ },
32
+ {
33
+ "loss": 0.484,
34
+ "learning_rate": 0.0002,
35
+ "epoch": 0.16,
36
+ "step": 6
37
+ },
38
+ {
39
+ "loss": 0.4106,
40
+ "learning_rate": 0.0002,
41
+ "epoch": 0.18,
42
+ "step": 7
43
+ },
44
+ {
45
+ "loss": 0.315,
46
+ "learning_rate": 0.0002,
47
+ "epoch": 0.21,
48
+ "step": 8
49
+ },
50
+ {
51
+ "loss": 0.303,
52
+ "learning_rate": 0.0002,
53
+ "epoch": 0.24,
54
+ "step": 9
55
+ },
56
+ {
57
+ "loss": 0.2428,
58
+ "learning_rate": 0.0002,
59
+ "epoch": 0.26,
60
+ "step": 10
61
+ },
62
+ {
63
+ "loss": 0.1576,
64
+ "learning_rate": 0.0002,
65
+ "epoch": 0.29,
66
+ "step": 11
67
+ },
68
+ {
69
+ "loss": 0.1088,
70
+ "learning_rate": 0.0002,
71
+ "epoch": 0.32,
72
+ "step": 12
73
+ },
74
+ {
75
+ "loss": 0.0539,
76
+ "learning_rate": 0.0002,
77
+ "epoch": 0.34,
78
+ "step": 13
79
+ },
80
+ {
81
+ "loss": 0.0554,
82
+ "learning_rate": 0.0002,
83
+ "epoch": 0.37,
84
+ "step": 14
85
+ },
86
+ {
87
+ "loss": 0.0339,
88
+ "learning_rate": 0.0002,
89
+ "epoch": 0.39,
90
+ "step": 15
91
+ },
92
+ {
93
+ "loss": 0.0978,
94
+ "learning_rate": 0.0002,
95
+ "epoch": 0.42,
96
+ "step": 16
97
+ },
98
+ {
99
+ "loss": 0.0074,
100
+ "learning_rate": 0.0002,
101
+ "epoch": 0.45,
102
+ "step": 17
103
+ },
104
+ {
105
+ "loss": 0.0462,
106
+ "learning_rate": 0.0002,
107
+ "epoch": 0.47,
108
+ "step": 18
109
+ },
110
+ {
111
+ "loss": 0.0314,
112
+ "learning_rate": 0.0002,
113
+ "epoch": 0.5,
114
+ "step": 19
115
+ },
116
+ {
117
+ "loss": 0.0483,
118
+ "learning_rate": 0.0002,
119
+ "epoch": 0.53,
120
+ "step": 20
121
+ },
122
+ {
123
+ "loss": 0.0277,
124
+ "learning_rate": 0.0002,
125
+ "epoch": 0.55,
126
+ "step": 21
127
+ },
128
+ {
129
+ "loss": 0.0407,
130
+ "learning_rate": 0.0002,
131
+ "epoch": 0.58,
132
+ "step": 22
133
+ },
134
+ {
135
+ "loss": 0.055,
136
+ "learning_rate": 0.0002,
137
+ "epoch": 0.61,
138
+ "step": 23
139
+ },
140
+ {
141
+ "loss": 0.0459,
142
+ "learning_rate": 0.0002,
143
+ "epoch": 0.63,
144
+ "step": 24
145
+ },
146
+ {
147
+ "loss": 0.0388,
148
+ "learning_rate": 0.0002,
149
+ "epoch": 0.66,
150
+ "step": 25
151
+ },
152
+ {
153
+ "eval_creative_writing_loss": 0.10124114900827408,
154
+ "eval_creative_writing_score": -0.029021821916103363,
155
+ "eval_creative_writing_brier_score": 0.029021821916103363,
156
+ "eval_creative_writing_average_probability": 0.949838399887085,
157
+ "eval_creative_writing_accuracy": 0.96,
158
+ "eval_creative_writing_probabilities": [
159
+ 0.968286395072937,
160
+ 1.0,
161
+ 0.9910894632339478,
162
+ 1.0,
163
+ 0.7539140582084656,
164
+ 1.0,
165
+ 0.9999864101409912,
166
+ 1.0,
167
+ 0.10593231767416,
168
+ 0.9996770620346069,
169
+ 0.35035908222198486,
170
+ 1.0,
171
+ 0.9999998807907104,
172
+ 1.0,
173
+ 0.995671272277832,
174
+ 0.9999988079071045,
175
+ 0.9999849796295166,
176
+ 0.9999967813491821,
177
+ 0.9999188184738159,
178
+ 1.0,
179
+ 0.9981185793876648,
180
+ 0.9999852180480957,
181
+ 0.999852180480957,
182
+ 0.9999985694885254,
183
+ 0.7055116891860962,
184
+ 0.9584224820137024,
185
+ 0.9999998807907104,
186
+ 1.0,
187
+ 0.0526885986328125,
188
+ 0.9999996423721313,
189
+ 0.9999980926513672,
190
+ 1.0,
191
+ 0.9999997615814209,
192
+ 1.0,
193
+ 0.9999997615814209,
194
+ 1.0,
195
+ 0.996961772441864,
196
+ 1.0,
197
+ 0.9960450530052185,
198
+ 0.9987917542457581,
199
+ 0.9999934434890747,
200
+ 1.0,
201
+ 0.9999978542327881,
202
+ 1.0,
203
+ 0.9987438321113586,
204
+ 0.9999847412109375,
205
+ 0.9999997615814209,
206
+ 1.0,
207
+ 0.999990701675415,
208
+ 0.9999983310699463,
209
+ 0.9999998807907104,
210
+ 1.0,
211
+ 0.968285322189331,
212
+ 1.0,
213
+ 0.988308310508728,
214
+ 0.9999992847442627,
215
+ 0.9514806270599365,
216
+ 0.9999994039535522,
217
+ 0.9962933659553528,
218
+ 1.0,
219
+ 0.8887924551963806,
220
+ 1.0,
221
+ 1.0,
222
+ 1.0,
223
+ 0.9997608065605164,
224
+ 1.0,
225
+ 0.4080069065093994,
226
+ 0.6264576315879822,
227
+ 0.8959720134735107,
228
+ 1.0,
229
+ 0.7741034626960754,
230
+ 1.0,
231
+ 1.0,
232
+ 1.0,
233
+ 1.0,
234
+ 1.0,
235
+ 0.8713729381561279,
236
+ 1.0,
237
+ 0.9989833235740662,
238
+ 0.9999933242797852,
239
+ 1.0,
240
+ 1.0,
241
+ 0.9996505975723267,
242
+ 1.0,
243
+ 0.9916742444038391,
244
+ 0.9989118576049805,
245
+ 0.9800452589988708,
246
+ 0.9999994039535522,
247
+ 0.9998782873153687,
248
+ 0.9999986886978149,
249
+ 1.0,
250
+ 1.0,
251
+ 1.0,
252
+ 1.0,
253
+ 0.9999996423721313,
254
+ 0.9999991655349731,
255
+ 0.7797707915306091,
256
+ 0.9964134097099304,
257
+ 0.9997926354408264,
258
+ 1.0
259
+ ],
260
+ "eval_creative_writing_runtime": 31.3856,
261
+ "eval_creative_writing_samples_per_second": 3.186,
262
+ "eval_creative_writing_steps_per_second": 0.127,
263
+ "epoch": 0.66,
264
+ "step": 25
265
+ },
266
+ {
267
+ "eval_biology_with_literary_style_loss": 0.9078262448310852,
268
+ "eval_biology_with_literary_style_score": -0.22649109363555908,
269
+ "eval_biology_with_literary_style_brier_score": 0.22649109363555908,
270
+ "eval_biology_with_literary_style_average_probability": 0.7168461680412292,
271
+ "eval_biology_with_literary_style_accuracy": 0.72,
272
+ "eval_biology_with_literary_style_probabilities": [
273
+ 0.890451967716217,
274
+ 0.9884986281394958,
275
+ 0.4627087116241455,
276
+ 0.26461559534072876,
277
+ 0.9440751671791077,
278
+ 0.9983709454536438,
279
+ 0.14105352759361267,
280
+ 0.5504218935966492,
281
+ 0.0717836245894432,
282
+ 0.9877713918685913,
283
+ 0.9968053102493286,
284
+ 0.8904775977134705,
285
+ 0.2346506416797638,
286
+ 0.2469814568758011,
287
+ 0.6628801822662354,
288
+ 0.9975651502609253,
289
+ 0.16669368743896484,
290
+ 0.5848821401596069,
291
+ 0.9999884366989136,
292
+ 0.999921441078186,
293
+ 0.9999263286590576,
294
+ 0.9972250461578369,
295
+ 0.9998313188552856,
296
+ 0.6406869888305664,
297
+ 0.9976915121078491,
298
+ 0.9530547857284546,
299
+ 0.995725154876709,
300
+ 0.10443083941936493,
301
+ 0.023461028933525085,
302
+ 0.0064667826518416405,
303
+ 0.9999899864196777,
304
+ 0.997308611869812,
305
+ 0.9999022483825684,
306
+ 0.9999972581863403,
307
+ 0.9999064207077026,
308
+ 0.9999868869781494,
309
+ 0.5541044473648071,
310
+ 0.9237357378005981,
311
+ 0.9736377596855164,
312
+ 0.00014247097715269774,
313
+ 0.0008552187937311828,
314
+ 0.0020070422906428576,
315
+ 0.009259264916181564,
316
+ 0.09797927737236023,
317
+ 0.988251268863678,
318
+ 0.9928971529006958,
319
+ 0.003707324853166938,
320
+ 0.9992867112159729,
321
+ 0.9720431566238403,
322
+ 0.9831154942512512,
323
+ 0.9243084192276001,
324
+ 0.056156225502491,
325
+ 0.6303812265396118,
326
+ 0.9999059438705444,
327
+ 0.900888204574585,
328
+ 0.9978699684143066,
329
+ 0.91905677318573,
330
+ 0.22306089103221893,
331
+ 0.48004305362701416,
332
+ 0.02644597738981247,
333
+ 0.9815545678138733,
334
+ 0.9921851754188538,
335
+ 0.9861576557159424,
336
+ 0.9997566342353821,
337
+ 0.4922117292881012,
338
+ 0.9966878294944763,
339
+ 0.9991987347602844,
340
+ 0.9990311861038208,
341
+ 0.9266239404678345,
342
+ 0.9989709854125977,
343
+ 0.9725721478462219,
344
+ 0.9685802459716797,
345
+ 0.013590201735496521,
346
+ 0.7605278491973877,
347
+ 0.9985472559928894,
348
+ 0.998909592628479,
349
+ 0.9852558374404907,
350
+ 0.9983707070350647,
351
+ 0.12097364664077759,
352
+ 0.9766552448272705,
353
+ 0.02587122656404972,
354
+ 0.9760261178016663,
355
+ 0.9663914442062378,
356
+ 0.9991564750671387,
357
+ 0.9999828338623047,
358
+ 0.99964439868927,
359
+ 0.9962877035140991,
360
+ 0.9912015199661255,
361
+ 0.9959142804145813,
362
+ 0.004339241422712803,
363
+ 0.9495350122451782,
364
+ 0.9996113181114197,
365
+ 0.9997592568397522,
366
+ 0.9140135645866394,
367
+ 0.9997426867485046,
368
+ 0.9957128763198853,
369
+ 0.2625727951526642,
370
+ 0.3218887746334076,
371
+ 0.010261102579534054,
372
+ 0.655005931854248
373
+ ],
374
+ "eval_biology_with_literary_style_runtime": 34.5421,
375
+ "eval_biology_with_literary_style_samples_per_second": 2.895,
376
+ "eval_biology_with_literary_style_steps_per_second": 0.116,
377
+ "epoch": 0.66,
378
+ "step": 25
379
+ },
380
+ {
381
+ "loss": 0.211,
382
+ "learning_rate": 0.0002,
383
+ "epoch": 0.68,
384
+ "step": 26
385
+ },
386
+ {
387
+ "loss": 0.128,
388
+ "learning_rate": 0.0002,
389
+ "epoch": 0.71,
390
+ "step": 27
391
+ },
392
+ {
393
+ "loss": 0.0735,
394
+ "learning_rate": 0.0002,
395
+ "epoch": 0.74,
396
+ "step": 28
397
+ },
398
+ {
399
+ "loss": 0.0765,
400
+ "learning_rate": 0.0002,
401
+ "epoch": 0.76,
402
+ "step": 29
403
+ },
404
+ {
405
+ "loss": 0.0729,
406
+ "learning_rate": 0.0002,
407
+ "epoch": 0.79,
408
+ "step": 30
409
+ },
410
+ {
411
+ "loss": 0.0643,
412
+ "learning_rate": 0.0002,
413
+ "epoch": 0.82,
414
+ "step": 31
415
+ },
416
+ {
417
+ "loss": 0.052,
418
+ "learning_rate": 0.0002,
419
+ "epoch": 0.84,
420
+ "step": 32
421
+ },
422
+ {
423
+ "loss": 0.0159,
424
+ "learning_rate": 0.0002,
425
+ "epoch": 0.87,
426
+ "step": 33
427
+ },
428
+ {
429
+ "loss": 0.0546,
430
+ "learning_rate": 0.0002,
431
+ "epoch": 0.89,
432
+ "step": 34
433
+ },
434
+ {
435
+ "loss": 0.0484,
436
+ "learning_rate": 0.0002,
437
+ "epoch": 0.92,
438
+ "step": 35
439
+ },
440
+ {
441
+ "loss": 0.1863,
442
+ "learning_rate": 0.0002,
443
+ "epoch": 0.95,
444
+ "step": 36
445
+ },
446
+ {
447
+ "loss": 0.1555,
448
+ "learning_rate": 0.0002,
449
+ "epoch": 0.97,
450
+ "step": 37
451
+ },
452
+ {
453
+ "loss": 0.0444,
454
+ "learning_rate": 0.0002,
455
+ "epoch": 1.0,
456
+ "step": 38
457
+ },
458
+ {
459
+ "loss": 0.0087,
460
+ "learning_rate": 0.0002,
461
+ "epoch": 1.03,
462
+ "step": 39
463
+ },
464
+ {
465
+ "loss": 0.0122,
466
+ "learning_rate": 0.0002,
467
+ "epoch": 1.05,
468
+ "step": 40
469
+ },
470
+ {
471
+ "loss": 0.0664,
472
+ "learning_rate": 0.0002,
473
+ "epoch": 1.08,
474
+ "step": 41
475
+ },
476
+ {
477
+ "loss": 0.0011,
478
+ "learning_rate": 0.0002,
479
+ "epoch": 1.11,
480
+ "step": 42
481
+ },
482
+ {
483
+ "loss": 0.0363,
484
+ "learning_rate": 0.0002,
485
+ "epoch": 1.13,
486
+ "step": 43
487
+ },
488
+ {
489
+ "loss": 0.0011,
490
+ "learning_rate": 0.0002,
491
+ "epoch": 1.16,
492
+ "step": 44
493
+ },
494
+ {
495
+ "loss": 0.0145,
496
+ "learning_rate": 0.0002,
497
+ "epoch": 1.18,
498
+ "step": 45
499
+ },
500
+ {
501
+ "loss": 0.0055,
502
+ "learning_rate": 0.0002,
503
+ "epoch": 1.21,
504
+ "step": 46
505
+ },
506
+ {
507
+ "loss": 0.0542,
508
+ "learning_rate": 0.0002,
509
+ "epoch": 1.24,
510
+ "step": 47
511
+ },
512
+ {
513
+ "loss": 0.0092,
514
+ "learning_rate": 0.0002,
515
+ "epoch": 1.26,
516
+ "step": 48
517
+ },
518
+ {
519
+ "loss": 0.0047,
520
+ "learning_rate": 0.0002,
521
+ "epoch": 1.29,
522
+ "step": 49
523
+ },
524
+ {
525
+ "loss": 0.0913,
526
+ "learning_rate": 0.0002,
527
+ "epoch": 1.32,
528
+ "step": 50
529
+ },
530
+ {
531
+ "eval_creative_writing_loss": 0.0620737262070179,
532
+ "eval_creative_writing_score": -0.018390759825706482,
533
+ "eval_creative_writing_brier_score": 0.018390759825706482,
534
+ "eval_creative_writing_average_probability": 0.9659796953201294,
535
+ "eval_creative_writing_accuracy": 0.97,
536
+ "eval_creative_writing_probabilities": [
537
+ 0.9853099584579468,
538
+ 1.0,
539
+ 0.9972212314605713,
540
+ 1.0,
541
+ 0.9858209490776062,
542
+ 1.0,
543
+ 0.9999940395355225,
544
+ 1.0,
545
+ 0.26181724667549133,
546
+ 1.0,
547
+ 0.4172997772693634,
548
+ 1.0,
549
+ 1.0,
550
+ 1.0,
551
+ 0.998902440071106,
552
+ 1.0,
553
+ 1.0,
554
+ 1.0,
555
+ 0.9997408986091614,
556
+ 1.0,
557
+ 0.9994969367980957,
558
+ 0.9999998807907104,
559
+ 0.9999983310699463,
560
+ 1.0,
561
+ 0.8943321108818054,
562
+ 0.9996473789215088,
563
+ 1.0,
564
+ 1.0,
565
+ 0.1366727352142334,
566
+ 1.0,
567
+ 0.9999995231628418,
568
+ 1.0,
569
+ 1.0,
570
+ 1.0,
571
+ 1.0,
572
+ 1.0,
573
+ 0.9997522234916687,
574
+ 1.0,
575
+ 0.9999735355377197,
576
+ 0.9999978542327881,
577
+ 0.9999889135360718,
578
+ 1.0,
579
+ 0.9999996423721313,
580
+ 1.0,
581
+ 0.9998144507408142,
582
+ 0.9999997615814209,
583
+ 1.0,
584
+ 1.0,
585
+ 1.0,
586
+ 1.0,
587
+ 1.0,
588
+ 1.0,
589
+ 0.976751983165741,
590
+ 1.0,
591
+ 0.997978150844574,
592
+ 1.0,
593
+ 0.9721508622169495,
594
+ 1.0,
595
+ 0.9091917872428894,
596
+ 1.0,
597
+ 0.8016390800476074,
598
+ 1.0,
599
+ 1.0,
600
+ 1.0,
601
+ 0.9940642714500427,
602
+ 1.0,
603
+ 0.7791704535484314,
604
+ 0.9723653793334961,
605
+ 0.8304190635681152,
606
+ 1.0,
607
+ 0.738132655620575,
608
+ 1.0,
609
+ 1.0,
610
+ 1.0,
611
+ 1.0,
612
+ 1.0,
613
+ 0.997934103012085,
614
+ 1.0,
615
+ 0.9999746084213257,
616
+ 1.0,
617
+ 1.0,
618
+ 1.0,
619
+ 0.9992375373840332,
620
+ 1.0,
621
+ 0.9998910427093506,
622
+ 0.9999988079071045,
623
+ 0.9967170357704163,
624
+ 1.0,
625
+ 0.9999998807907104,
626
+ 1.0,
627
+ 1.0,
628
+ 1.0,
629
+ 1.0,
630
+ 1.0,
631
+ 1.0,
632
+ 1.0,
633
+ 0.9567482471466064,
634
+ 0.9999606609344482,
635
+ 0.9998536109924316,
636
+ 1.0
637
+ ],
638
+ "eval_creative_writing_runtime": 31.4284,
639
+ "eval_creative_writing_samples_per_second": 3.182,
640
+ "eval_creative_writing_steps_per_second": 0.127,
641
+ "epoch": 1.32,
642
+ "step": 50
643
+ },
644
+ {
645
+ "eval_biology_with_literary_style_loss": 1.3244109153747559,
646
+ "eval_biology_with_literary_style_score": -0.2915953993797302,
647
+ "eval_biology_with_literary_style_brier_score": 0.2915953993797302,
648
+ "eval_biology_with_literary_style_average_probability": 0.6370630860328674,
649
+ "eval_biology_with_literary_style_accuracy": 0.66,
650
+ "eval_biology_with_literary_style_probabilities": [
651
+ 0.8422331213951111,
652
+ 0.3563534617424011,
653
+ 0.20590150356292725,
654
+ 0.34388160705566406,
655
+ 0.9738777279853821,
656
+ 0.9933098554611206,
657
+ 0.013515126891434193,
658
+ 0.03492158278822899,
659
+ 0.0005772275617346168,
660
+ 0.9946145415306091,
661
+ 0.9994958639144897,
662
+ 0.9732575416564941,
663
+ 0.4653439223766327,
664
+ 0.2902985215187073,
665
+ 0.9848002195358276,
666
+ 0.874828040599823,
667
+ 0.007957144640386105,
668
+ 0.10880450904369354,
669
+ 0.9999992847442627,
670
+ 0.9999992847442627,
671
+ 0.9997832179069519,
672
+ 0.9021045565605164,
673
+ 0.9993135929107666,
674
+ 0.09312967211008072,
675
+ 0.9835945963859558,
676
+ 0.9038721919059753,
677
+ 0.9953224062919617,
678
+ 0.005929848644882441,
679
+ 0.004466882906854153,
680
+ 0.034653790295124054,
681
+ 0.9999308586120605,
682
+ 0.9367615580558777,
683
+ 0.9977996945381165,
684
+ 0.9999979734420776,
685
+ 0.996374785900116,
686
+ 0.9997103810310364,
687
+ 0.19381558895111084,
688
+ 0.6907597780227661,
689
+ 0.875636100769043,
690
+ 8.129484740493353e-06,
691
+ 6.22305233264342e-05,
692
+ 0.0003827095788437873,
693
+ 0.0011861087987199426,
694
+ 0.013507246039807796,
695
+ 0.9973799586296082,
696
+ 0.6868321299552917,
697
+ 0.04164070263504982,
698
+ 0.9997265934944153,
699
+ 0.9688616991043091,
700
+ 0.9772934317588806,
701
+ 0.9902143478393555,
702
+ 0.05315324291586876,
703
+ 0.7507219910621643,
704
+ 0.997478187084198,
705
+ 0.8234730362892151,
706
+ 0.9983914494514465,
707
+ 0.5037214159965515,
708
+ 0.1830606311559677,
709
+ 0.6088641285896301,
710
+ 0.056403663009405136,
711
+ 0.9997255206108093,
712
+ 0.997403085231781,
713
+ 0.9995600581169128,
714
+ 0.9978277087211609,
715
+ 0.1915532499551773,
716
+ 0.886441171169281,
717
+ 0.9874991774559021,
718
+ 0.9855878353118896,
719
+ 0.7539471387863159,
720
+ 0.9771825671195984,
721
+ 0.5847952961921692,
722
+ 0.7823383212089539,
723
+ 0.010838981717824936,
724
+ 0.7440186142921448,
725
+ 0.8710861206054688,
726
+ 0.9924225211143494,
727
+ 0.41601699590682983,
728
+ 0.9564679861068726,
729
+ 0.5351160168647766,
730
+ 0.9124112725257874,
731
+ 0.00990450568497181,
732
+ 0.8436535000801086,
733
+ 0.45744913816452026,
734
+ 0.9863516092300415,
735
+ 0.9999998807907104,
736
+ 0.99796462059021,
737
+ 0.984410285949707,
738
+ 0.8628932237625122,
739
+ 0.7051854133605957,
740
+ 0.004238440655171871,
741
+ 0.6792111396789551,
742
+ 0.9969097971916199,
743
+ 0.9998319149017334,
744
+ 0.12430453300476074,
745
+ 0.9992156028747559,
746
+ 0.8144650459289551,
747
+ 0.07938272505998611,
748
+ 0.013606518507003784,
749
+ 0.003284846432507038,
750
+ 0.8325172066688538
751
+ ],
752
+ "eval_biology_with_literary_style_runtime": 34.5464,
753
+ "eval_biology_with_literary_style_samples_per_second": 2.895,
754
+ "eval_biology_with_literary_style_steps_per_second": 0.116,
755
+ "epoch": 1.32,
756
+ "step": 50
757
+ },
758
+ {
759
+ "loss": 0.0092,
760
+ "learning_rate": 0.0002,
761
+ "epoch": 1.34,
762
+ "step": 51
763
+ },
764
+ {
765
+ "loss": 0.0012,
766
+ "learning_rate": 0.0002,
767
+ "epoch": 1.37,
768
+ "step": 52
769
+ },
770
+ {
771
+ "loss": 0.0741,
772
+ "learning_rate": 0.0002,
773
+ "epoch": 1.39,
774
+ "step": 53
775
+ },
776
+ {
777
+ "loss": 0.0191,
778
+ "learning_rate": 0.0002,
779
+ "epoch": 1.42,
780
+ "step": 54
781
+ },
782
+ {
783
+ "loss": 0.0289,
784
+ "learning_rate": 0.0002,
785
+ "epoch": 1.45,
786
+ "step": 55
787
+ },
788
+ {
789
+ "loss": 0.008,
790
+ "learning_rate": 0.0002,
791
+ "epoch": 1.47,
792
+ "step": 56
793
+ },
794
+ {
795
+ "loss": 0.0091,
796
+ "learning_rate": 0.0002,
797
+ "epoch": 1.5,
798
+ "step": 57
799
+ },
800
+ {
801
+ "loss": 0.001,
802
+ "learning_rate": 0.0002,
803
+ "epoch": 1.53,
804
+ "step": 58
805
+ },
806
+ {
807
+ "loss": 0.0589,
808
+ "learning_rate": 0.0002,
809
+ "epoch": 1.55,
810
+ "step": 59
811
+ },
812
+ {
813
+ "loss": 0.0014,
814
+ "learning_rate": 0.0002,
815
+ "epoch": 1.58,
816
+ "step": 60
817
+ },
818
+ {
819
+ "loss": 0.0009,
820
+ "learning_rate": 0.0002,
821
+ "epoch": 1.61,
822
+ "step": 61
823
+ },
824
+ {
825
+ "loss": 0.0032,
826
+ "learning_rate": 0.0002,
827
+ "epoch": 1.63,
828
+ "step": 62
829
+ },
830
+ {
831
+ "loss": 0.0001,
832
+ "learning_rate": 0.0002,
833
+ "epoch": 1.66,
834
+ "step": 63
835
+ },
836
+ {
837
+ "loss": 0.0163,
838
+ "learning_rate": 0.0002,
839
+ "epoch": 1.68,
840
+ "step": 64
841
+ },
842
+ {
843
+ "loss": 0.0007,
844
+ "learning_rate": 0.0002,
845
+ "epoch": 1.71,
846
+ "step": 65
847
+ },
848
+ {
849
+ "loss": 0.0153,
850
+ "learning_rate": 0.0002,
851
+ "epoch": 1.74,
852
+ "step": 66
853
+ },
854
+ {
855
+ "loss": 0.0266,
856
+ "learning_rate": 0.0002,
857
+ "epoch": 1.76,
858
+ "step": 67
859
+ },
860
+ {
861
+ "loss": 0.0547,
862
+ "learning_rate": 0.0002,
863
+ "epoch": 1.79,
864
+ "step": 68
865
+ },
866
+ {
867
+ "loss": 0.0024,
868
+ "learning_rate": 0.0002,
869
+ "epoch": 1.82,
870
+ "step": 69
871
+ },
872
+ {
873
+ "loss": 0.0118,
874
+ "learning_rate": 0.0002,
875
+ "epoch": 1.84,
876
+ "step": 70
877
+ },
878
+ {
879
+ "loss": 0.0031,
880
+ "learning_rate": 0.0002,
881
+ "epoch": 1.87,
882
+ "step": 71
883
+ },
884
+ {
885
+ "loss": 0.0076,
886
+ "learning_rate": 0.0002,
887
+ "epoch": 1.89,
888
+ "step": 72
889
+ },
890
+ {
891
+ "loss": 0.0201,
892
+ "learning_rate": 0.0002,
893
+ "epoch": 1.92,
894
+ "step": 73
895
+ },
896
+ {
897
+ "loss": 0.0105,
898
+ "learning_rate": 0.0002,
899
+ "epoch": 1.95,
900
+ "step": 74
901
+ },
902
+ {
903
+ "loss": 0.0089,
904
+ "learning_rate": 0.0002,
905
+ "epoch": 1.97,
906
+ "step": 75
907
+ },
908
+ {
909
+ "eval_creative_writing_loss": 0.09592155367136002,
910
+ "eval_creative_writing_score": -0.020631961524486542,
911
+ "eval_creative_writing_brier_score": 0.020631961524486542,
912
+ "eval_creative_writing_average_probability": 0.971525251865387,
913
+ "eval_creative_writing_accuracy": 0.97,
914
+ "eval_creative_writing_probabilities": [
915
+ 0.9997228980064392,
916
+ 1.0,
917
+ 0.9999998807907104,
918
+ 1.0,
919
+ 0.9996482133865356,
920
+ 1.0,
921
+ 0.9999998807907104,
922
+ 1.0,
923
+ 0.2378961592912674,
924
+ 0.9999994039535522,
925
+ 0.42051035165786743,
926
+ 1.0,
927
+ 1.0,
928
+ 1.0,
929
+ 0.9999306201934814,
930
+ 1.0,
931
+ 0.9999998807907104,
932
+ 1.0,
933
+ 0.9999996423721313,
934
+ 1.0,
935
+ 0.999981164932251,
936
+ 0.9999998807907104,
937
+ 0.9999912977218628,
938
+ 1.0,
939
+ 0.9809040427207947,
940
+ 0.9995905756950378,
941
+ 1.0,
942
+ 1.0,
943
+ 0.0023363870568573475,
944
+ 1.0,
945
+ 1.0,
946
+ 1.0,
947
+ 1.0,
948
+ 1.0,
949
+ 1.0,
950
+ 1.0,
951
+ 0.9999964237213135,
952
+ 1.0,
953
+ 0.9986351132392883,
954
+ 0.9997923970222473,
955
+ 0.9999994039535522,
956
+ 1.0,
957
+ 1.0,
958
+ 1.0,
959
+ 0.9998650550842285,
960
+ 0.9999997615814209,
961
+ 1.0,
962
+ 1.0,
963
+ 1.0,
964
+ 1.0,
965
+ 1.0,
966
+ 1.0,
967
+ 0.9987547397613525,
968
+ 1.0,
969
+ 0.9995219707489014,
970
+ 1.0,
971
+ 0.9831072092056274,
972
+ 1.0,
973
+ 0.9974205493927002,
974
+ 1.0,
975
+ 0.9997766613960266,
976
+ 1.0,
977
+ 1.0,
978
+ 1.0,
979
+ 0.9999995231628418,
980
+ 1.0,
981
+ 0.6153692007064819,
982
+ 0.9825475811958313,
983
+ 0.9969502091407776,
984
+ 1.0,
985
+ 0.9921615719795227,
986
+ 1.0,
987
+ 1.0,
988
+ 1.0,
989
+ 1.0,
990
+ 1.0,
991
+ 0.9998723268508911,
992
+ 1.0,
993
+ 0.9999754428863525,
994
+ 0.9999998807907104,
995
+ 1.0,
996
+ 1.0,
997
+ 0.9998829364776611,
998
+ 1.0,
999
+ 0.9971277117729187,
1000
+ 0.9999291896820068,
1001
+ 0.9999527931213379,
1002
+ 1.0,
1003
+ 1.0,
1004
+ 1.0,
1005
+ 1.0,
1006
+ 1.0,
1007
+ 1.0,
1008
+ 1.0,
1009
+ 1.0,
1010
+ 1.0,
1011
+ 0.952387273311615,
1012
+ 0.9989917874336243,
1013
+ 0.9999982118606567,
1014
+ 1.0
1015
+ ],
1016
+ "eval_creative_writing_runtime": 31.3865,
1017
+ "eval_creative_writing_samples_per_second": 3.186,
1018
+ "eval_creative_writing_steps_per_second": 0.127,
1019
+ "epoch": 1.97,
1020
+ "step": 75
1021
+ },
1022
+ {
1023
+ "eval_biology_with_literary_style_loss": 1.4261603355407715,
1024
+ "eval_biology_with_literary_style_score": -0.23246526718139648,
1025
+ "eval_biology_with_literary_style_brier_score": 0.23246526718139648,
1026
+ "eval_biology_with_literary_style_average_probability": 0.739600419998169,
1027
+ "eval_biology_with_literary_style_accuracy": 0.75,
1028
+ "eval_biology_with_literary_style_probabilities": [
1029
+ 0.9998239874839783,
1030
+ 0.9999821186065674,
1031
+ 0.997286319732666,
1032
+ 0.7309690117835999,
1033
+ 0.9989686012268066,
1034
+ 0.9998588562011719,
1035
+ 0.1100221648812294,
1036
+ 0.013960055075585842,
1037
+ 0.0014234904665499926,
1038
+ 0.9999901056289673,
1039
+ 0.9999978542327881,
1040
+ 0.9998490810394287,
1041
+ 0.5011758804321289,
1042
+ 0.24851495027542114,
1043
+ 0.9544274806976318,
1044
+ 0.996453046798706,
1045
+ 0.0021988353691995144,
1046
+ 0.05229213833808899,
1047
+ 1.0,
1048
+ 1.0,
1049
+ 0.9999995231628418,
1050
+ 0.9999839067459106,
1051
+ 0.999997615814209,
1052
+ 0.9901941418647766,
1053
+ 0.9993062019348145,
1054
+ 0.9996980428695679,
1055
+ 0.9998052716255188,
1056
+ 7.921013457234949e-05,
1057
+ 0.00241833901964128,
1058
+ 0.00010733860108302906,
1059
+ 1.0,
1060
+ 0.9999939203262329,
1061
+ 1.0,
1062
+ 1.0,
1063
+ 0.9999995231628418,
1064
+ 1.0,
1065
+ 0.37197402119636536,
1066
+ 0.906427264213562,
1067
+ 0.9985352754592896,
1068
+ 9.472542927824179e-08,
1069
+ 1.8663062292034738e-05,
1070
+ 0.0007226847810670733,
1071
+ 0.0009153155260719359,
1072
+ 0.027196291834115982,
1073
+ 0.996335506439209,
1074
+ 0.9934919476509094,
1075
+ 0.009263264015316963,
1076
+ 0.9999862909317017,
1077
+ 0.9986127614974976,
1078
+ 0.999937891960144,
1079
+ 0.9997109770774841,
1080
+ 0.5540578961372375,
1081
+ 0.9694981575012207,
1082
+ 0.9999996423721313,
1083
+ 0.9991618394851685,
1084
+ 0.9999830722808838,
1085
+ 0.9839572906494141,
1086
+ 0.3787292242050171,
1087
+ 0.7562795877456665,
1088
+ 0.024417594075202942,
1089
+ 0.9996551275253296,
1090
+ 0.9969584941864014,
1091
+ 0.998749852180481,
1092
+ 0.9999865293502808,
1093
+ 0.7677027583122253,
1094
+ 0.9995812773704529,
1095
+ 0.9999984502792358,
1096
+ 0.9999995231628418,
1097
+ 0.9999768733978271,
1098
+ 0.9979830980300903,
1099
+ 0.9924507737159729,
1100
+ 0.9991115927696228,
1101
+ 7.307936175493523e-05,
1102
+ 0.7401599884033203,
1103
+ 0.929680347442627,
1104
+ 0.9999669790267944,
1105
+ 0.8817222118377686,
1106
+ 0.9998908042907715,
1107
+ 0.1284332275390625,
1108
+ 0.928813636302948,
1109
+ 0.0032492727041244507,
1110
+ 0.9993767142295837,
1111
+ 0.9767991900444031,
1112
+ 0.9966356158256531,
1113
+ 1.0,
1114
+ 0.9999995231628418,
1115
+ 0.9999977350234985,
1116
+ 0.9992828965187073,
1117
+ 0.999262273311615,
1118
+ 1.0285020834999159e-05,
1119
+ 0.9999904632568359,
1120
+ 1.0,
1121
+ 1.0,
1122
+ 0.9980078339576721,
1123
+ 0.9999998807907104,
1124
+ 0.999974250793457,
1125
+ 0.034345593303442,
1126
+ 0.024079279974102974,
1127
+ 0.0003281444078311324,
1128
+ 0.9998185038566589
1129
+ ],
1130
+ "eval_biology_with_literary_style_runtime": 34.5167,
1131
+ "eval_biology_with_literary_style_samples_per_second": 2.897,
1132
+ "eval_biology_with_literary_style_steps_per_second": 0.116,
1133
+ "epoch": 1.97,
1134
+ "step": 75
1135
+ },
1136
+ {
1137
+ "loss": 0.0031,
1138
+ "learning_rate": 0.0002,
1139
+ "epoch": 2.0,
1140
+ "step": 76
1141
+ },
1142
+ {
1143
+ "loss": 0.0008,
1144
+ "learning_rate": 0.0002,
1145
+ "epoch": 2.03,
1146
+ "step": 77
1147
+ },
1148
+ {
1149
+ "loss": 0.0019,
1150
+ "learning_rate": 0.0002,
1151
+ "epoch": 2.05,
1152
+ "step": 78
1153
+ },
1154
+ {
1155
+ "loss": 0.0,
1156
+ "learning_rate": 0.0002,
1157
+ "epoch": 2.08,
1158
+ "step": 79
1159
+ },
1160
+ {
1161
+ "loss": 0.0009,
1162
+ "learning_rate": 0.0002,
1163
+ "epoch": 2.11,
1164
+ "step": 80
1165
+ },
1166
+ {
1167
+ "loss": 0.0005,
1168
+ "learning_rate": 0.0002,
1169
+ "epoch": 2.13,
1170
+ "step": 81
1171
+ },
1172
+ {
1173
+ "loss": 0.0006,
1174
+ "learning_rate": 0.0002,
1175
+ "epoch": 2.16,
1176
+ "step": 82
1177
+ },
1178
+ {
1179
+ "loss": 0.0001,
1180
+ "learning_rate": 0.0002,
1181
+ "epoch": 2.18,
1182
+ "step": 83
1183
+ },
1184
+ {
1185
+ "loss": 0.0002,
1186
+ "learning_rate": 0.0002,
1187
+ "epoch": 2.21,
1188
+ "step": 84
1189
+ },
1190
+ {
1191
+ "loss": 0.0002,
1192
+ "learning_rate": 0.0002,
1193
+ "epoch": 2.24,
1194
+ "step": 85
1195
+ },
1196
+ {
1197
+ "loss": 0.0002,
1198
+ "learning_rate": 0.0002,
1199
+ "epoch": 2.26,
1200
+ "step": 86
1201
+ },
1202
+ {
1203
+ "loss": 0.0002,
1204
+ "learning_rate": 0.0002,
1205
+ "epoch": 2.29,
1206
+ "step": 87
1207
+ },
1208
+ {
1209
+ "loss": 0.0013,
1210
+ "learning_rate": 0.0002,
1211
+ "epoch": 2.32,
1212
+ "step": 88
1213
+ },
1214
+ {
1215
+ "loss": 0.0,
1216
+ "learning_rate": 0.0002,
1217
+ "epoch": 2.34,
1218
+ "step": 89
1219
+ },
1220
+ {
1221
+ "loss": 0.0018,
1222
+ "learning_rate": 0.0002,
1223
+ "epoch": 2.37,
1224
+ "step": 90
1225
+ },
1226
+ {
1227
+ "loss": 0.0012,
1228
+ "learning_rate": 0.0002,
1229
+ "epoch": 2.39,
1230
+ "step": 91
1231
+ },
1232
+ {
1233
+ "loss": 0.0004,
1234
+ "learning_rate": 0.0002,
1235
+ "epoch": 2.42,
1236
+ "step": 92
1237
+ },
1238
+ {
1239
+ "loss": 0.0038,
1240
+ "learning_rate": 0.0002,
1241
+ "epoch": 2.45,
1242
+ "step": 93
1243
+ },
1244
+ {
1245
+ "loss": 0.0001,
1246
+ "learning_rate": 0.0002,
1247
+ "epoch": 2.47,
1248
+ "step": 94
1249
+ },
1250
+ {
1251
+ "loss": 0.0007,
1252
+ "learning_rate": 0.0002,
1253
+ "epoch": 2.5,
1254
+ "step": 95
1255
+ },
1256
+ {
1257
+ "loss": 0.028,
1258
+ "learning_rate": 0.0002,
1259
+ "epoch": 2.53,
1260
+ "step": 96
1261
+ },
1262
+ {
1263
+ "loss": 0.003,
1264
+ "learning_rate": 0.0002,
1265
+ "epoch": 2.55,
1266
+ "step": 97
1267
+ },
1268
+ {
1269
+ "loss": 0.0082,
1270
+ "learning_rate": 0.0002,
1271
+ "epoch": 2.58,
1272
+ "step": 98
1273
+ },
1274
+ {
1275
+ "loss": 0.0001,
1276
+ "learning_rate": 0.0002,
1277
+ "epoch": 2.61,
1278
+ "step": 99
1279
+ },
1280
+ {
1281
+ "loss": 0.0002,
1282
+ "learning_rate": 0.0002,
1283
+ "epoch": 2.63,
1284
+ "step": 100
1285
+ },
1286
+ {
1287
+ "eval_creative_writing_loss": 0.12560071051120758,
1288
+ "eval_creative_writing_score": -0.02528984844684601,
1289
+ "eval_creative_writing_brier_score": 0.02528984844684601,
1290
+ "eval_creative_writing_average_probability": 0.9707810282707214,
1291
+ "eval_creative_writing_accuracy": 0.97,
1292
+ "eval_creative_writing_probabilities": [
1293
+ 0.9999761581420898,
1294
+ 1.0,
1295
+ 0.9999992847442627,
1296
+ 1.0,
1297
+ 0.9999547004699707,
1298
+ 1.0,
1299
+ 1.0,
1300
+ 1.0,
1301
+ 0.01668044924736023,
1302
+ 1.0,
1303
+ 0.2576431930065155,
1304
+ 1.0,
1305
+ 1.0,
1306
+ 1.0,
1307
+ 0.9999969005584717,
1308
+ 1.0,
1309
+ 1.0,
1310
+ 1.0,
1311
+ 1.0,
1312
+ 1.0,
1313
+ 0.9999998807907104,
1314
+ 1.0,
1315
+ 1.0,
1316
+ 1.0,
1317
+ 0.9979181885719299,
1318
+ 0.9999982118606567,
1319
+ 1.0,
1320
+ 1.0,
1321
+ 0.004256995394825935,
1322
+ 1.0,
1323
+ 1.0,
1324
+ 1.0,
1325
+ 1.0,
1326
+ 1.0,
1327
+ 1.0,
1328
+ 1.0,
1329
+ 0.9999716281890869,
1330
+ 1.0,
1331
+ 0.9999717473983765,
1332
+ 0.9999995231628418,
1333
+ 0.9999998807907104,
1334
+ 1.0,
1335
+ 1.0,
1336
+ 1.0,
1337
+ 0.999849796295166,
1338
+ 1.0,
1339
+ 1.0,
1340
+ 1.0,
1341
+ 1.0,
1342
+ 1.0,
1343
+ 1.0,
1344
+ 1.0,
1345
+ 0.9995261430740356,
1346
+ 1.0,
1347
+ 0.9999653100967407,
1348
+ 1.0,
1349
+ 0.9834521412849426,
1350
+ 1.0,
1351
+ 0.9987533092498779,
1352
+ 1.0,
1353
+ 0.9978016018867493,
1354
+ 1.0,
1355
+ 1.0,
1356
+ 1.0,
1357
+ 0.9999966621398926,
1358
+ 1.0,
1359
+ 0.8659273386001587,
1360
+ 0.9995256662368774,
1361
+ 0.9950549602508545,
1362
+ 1.0,
1363
+ 0.9657167792320251,
1364
+ 1.0,
1365
+ 1.0,
1366
+ 1.0,
1367
+ 1.0,
1368
+ 1.0,
1369
+ 0.9999123811721802,
1370
+ 1.0,
1371
+ 0.9999997615814209,
1372
+ 1.0,
1373
+ 1.0,
1374
+ 1.0,
1375
+ 0.9999539852142334,
1376
+ 1.0,
1377
+ 0.9996681213378906,
1378
+ 0.9999990463256836,
1379
+ 0.9996721744537354,
1380
+ 1.0,
1381
+ 1.0,
1382
+ 1.0,
1383
+ 1.0,
1384
+ 1.0,
1385
+ 1.0,
1386
+ 1.0,
1387
+ 1.0,
1388
+ 1.0,
1389
+ 0.9969596862792969,
1390
+ 0.9999977350234985,
1391
+ 0.9999960660934448,
1392
+ 1.0
1393
+ ],
1394
+ "eval_creative_writing_runtime": 31.387,
1395
+ "eval_creative_writing_samples_per_second": 3.186,
1396
+ "eval_creative_writing_steps_per_second": 0.127,
1397
+ "epoch": 2.63,
1398
+ "step": 100
1399
+ },
1400
+ {
1401
+ "eval_biology_with_literary_style_loss": 1.4950416088104248,
1402
+ "eval_biology_with_literary_style_score": -0.239924356341362,
1403
+ "eval_biology_with_literary_style_brier_score": 0.239924356341362,
1404
+ "eval_biology_with_literary_style_average_probability": 0.7326297760009766,
1405
+ "eval_biology_with_literary_style_accuracy": 0.75,
1406
+ "eval_biology_with_literary_style_probabilities": [
1407
+ 0.9994196891784668,
1408
+ 0.9992738366127014,
1409
+ 0.983923614025116,
1410
+ 0.9044151902198792,
1411
+ 0.9997815489768982,
1412
+ 0.9998750686645508,
1413
+ 0.006262939423322678,
1414
+ 0.00128739804495126,
1415
+ 8.855006308294833e-05,
1416
+ 0.9999986886978149,
1417
+ 0.9999992847442627,
1418
+ 0.9998844861984253,
1419
+ 0.684658944606781,
1420
+ 0.4036870300769806,
1421
+ 0.9940760135650635,
1422
+ 0.9792241454124451,
1423
+ 0.0024490836076438427,
1424
+ 0.030131032690405846,
1425
+ 1.0,
1426
+ 1.0,
1427
+ 0.9999988079071045,
1428
+ 0.9999747276306152,
1429
+ 0.9999988079071045,
1430
+ 0.9765191674232483,
1431
+ 0.9992048144340515,
1432
+ 0.9988371729850769,
1433
+ 0.9996092915534973,
1434
+ 8.090558549156412e-05,
1435
+ 0.004592935089021921,
1436
+ 0.002051251009106636,
1437
+ 1.0,
1438
+ 0.9999823570251465,
1439
+ 0.9999997615814209,
1440
+ 1.0,
1441
+ 0.9999997615814209,
1442
+ 0.9999996423721313,
1443
+ 0.523539662361145,
1444
+ 0.9182876348495483,
1445
+ 0.9991507530212402,
1446
+ 6.639822203169388e-08,
1447
+ 3.326304067741148e-05,
1448
+ 0.00180066900793463,
1449
+ 0.00022590388834942132,
1450
+ 0.0030361979734152555,
1451
+ 0.9938976168632507,
1452
+ 0.999334990978241,
1453
+ 0.014086072333157063,
1454
+ 0.9999929666519165,
1455
+ 0.9996891021728516,
1456
+ 0.9999194145202637,
1457
+ 0.9998146891593933,
1458
+ 0.5219235420227051,
1459
+ 0.9443055391311646,
1460
+ 0.999993085861206,
1461
+ 0.9992896318435669,
1462
+ 0.9999892711639404,
1463
+ 0.8569433093070984,
1464
+ 0.11361022293567657,
1465
+ 0.6792670488357544,
1466
+ 0.07281597703695297,
1467
+ 0.9995861649513245,
1468
+ 0.9714840054512024,
1469
+ 0.9986534118652344,
1470
+ 0.9999567270278931,
1471
+ 0.7200058102607727,
1472
+ 0.9917659163475037,
1473
+ 0.9999903440475464,
1474
+ 0.9999982118606567,
1475
+ 0.999862551689148,
1476
+ 0.9478113055229187,
1477
+ 0.9273961782455444,
1478
+ 0.9989776611328125,
1479
+ 9.58814489422366e-05,
1480
+ 0.9144469499588013,
1481
+ 0.8570235371589661,
1482
+ 0.999974250793457,
1483
+ 0.27402758598327637,
1484
+ 0.9973011612892151,
1485
+ 0.07933907210826874,
1486
+ 0.9523033499717712,
1487
+ 0.005996616557240486,
1488
+ 0.9999831914901733,
1489
+ 0.9931454658508301,
1490
+ 0.9989646673202515,
1491
+ 1.0,
1492
+ 0.9999984502792358,
1493
+ 0.999990701675415,
1494
+ 0.9977810978889465,
1495
+ 0.9995942711830139,
1496
+ 0.00018957872816827148,
1497
+ 0.9999784231185913,
1498
+ 0.9999998807907104,
1499
+ 1.0,
1500
+ 0.9992635846138,
1501
+ 1.0,
1502
+ 0.9999830722808838,
1503
+ 0.024764718487858772,
1504
+ 0.0030579320155084133,
1505
+ 0.0002520766283851117,
1506
+ 0.9998053908348083
1507
+ ],
1508
+ "eval_biology_with_literary_style_runtime": 34.5448,
1509
+ "eval_biology_with_literary_style_samples_per_second": 2.895,
1510
+ "eval_biology_with_literary_style_steps_per_second": 0.116,
1511
+ "epoch": 2.63,
1512
+ "step": 100
1513
+ },
1514
+ {
1515
+ "loss": 0.0,
1516
+ "learning_rate": 0.0002,
1517
+ "epoch": 2.66,
1518
+ "step": 101
1519
+ },
1520
+ {
1521
+ "loss": 0.0,
1522
+ "learning_rate": 0.0002,
1523
+ "epoch": 2.68,
1524
+ "step": 102
1525
+ },
1526
+ {
1527
+ "loss": 0.0008,
1528
+ "learning_rate": 0.0002,
1529
+ "epoch": 2.71,
1530
+ "step": 103
1531
+ },
1532
+ {
1533
+ "loss": 0.0,
1534
+ "learning_rate": 0.0002,
1535
+ "epoch": 2.74,
1536
+ "step": 104
1537
+ },
1538
+ {
1539
+ "loss": 0.0002,
1540
+ "learning_rate": 0.0002,
1541
+ "epoch": 2.76,
1542
+ "step": 105
1543
+ },
1544
+ {
1545
+ "loss": 0.0,
1546
+ "learning_rate": 0.0002,
1547
+ "epoch": 2.79,
1548
+ "step": 106
1549
+ },
1550
+ {
1551
+ "loss": 0.0012,
1552
+ "learning_rate": 0.0002,
1553
+ "epoch": 2.82,
1554
+ "step": 107
1555
+ },
1556
+ {
1557
+ "loss": 0.0002,
1558
+ "learning_rate": 0.0002,
1559
+ "epoch": 2.84,
1560
+ "step": 108
1561
+ },
1562
+ {
1563
+ "loss": 0.0003,
1564
+ "learning_rate": 0.0002,
1565
+ "epoch": 2.87,
1566
+ "step": 109
1567
+ },
1568
+ {
1569
+ "loss": 0.0004,
1570
+ "learning_rate": 0.0002,
1571
+ "epoch": 2.89,
1572
+ "step": 110
1573
+ },
1574
+ {
1575
+ "loss": 0.0001,
1576
+ "learning_rate": 0.0002,
1577
+ "epoch": 2.92,
1578
+ "step": 111
1579
+ },
1580
+ {
1581
+ "loss": 0.0001,
1582
+ "learning_rate": 0.0002,
1583
+ "epoch": 2.95,
1584
+ "step": 112
1585
+ },
1586
+ {
1587
+ "loss": 0.0,
1588
+ "learning_rate": 0.0002,
1589
+ "epoch": 2.97,
1590
+ "step": 113
1591
+ },
1592
+ {
1593
+ "loss": 0.0005,
1594
+ "learning_rate": 0.0002,
1595
+ "epoch": 3.0,
1596
+ "step": 114
1597
+ },
1598
+ {
1599
+ "loss": 0.0002,
1600
+ "learning_rate": 0.0002,
1601
+ "epoch": 3.03,
1602
+ "step": 115
1603
+ },
1604
+ {
1605
+ "loss": 0.0001,
1606
+ "learning_rate": 0.0002,
1607
+ "epoch": 3.05,
1608
+ "step": 116
1609
+ },
1610
+ {
1611
+ "loss": 0.0,
1612
+ "learning_rate": 0.0002,
1613
+ "epoch": 3.08,
1614
+ "step": 117
1615
+ },
1616
+ {
1617
+ "loss": 0.0001,
1618
+ "learning_rate": 0.0002,
1619
+ "epoch": 3.11,
1620
+ "step": 118
1621
+ },
1622
+ {
1623
+ "loss": 0.0,
1624
+ "learning_rate": 0.0002,
1625
+ "epoch": 3.13,
1626
+ "step": 119
1627
+ },
1628
+ {
1629
+ "loss": 0.0001,
1630
+ "learning_rate": 0.0002,
1631
+ "epoch": 3.16,
1632
+ "step": 120
1633
+ },
1634
+ {
1635
+ "loss": 0.0001,
1636
+ "learning_rate": 0.0002,
1637
+ "epoch": 3.18,
1638
+ "step": 121
1639
+ },
1640
+ {
1641
+ "loss": 0.0001,
1642
+ "learning_rate": 0.0002,
1643
+ "epoch": 3.21,
1644
+ "step": 122
1645
+ },
1646
+ {
1647
+ "loss": 0.0001,
1648
+ "learning_rate": 0.0002,
1649
+ "epoch": 3.24,
1650
+ "step": 123
1651
+ },
1652
+ {
1653
+ "loss": 0.0001,
1654
+ "learning_rate": 0.0002,
1655
+ "epoch": 3.26,
1656
+ "step": 124
1657
+ },
1658
+ {
1659
+ "loss": 0.0002,
1660
+ "learning_rate": 0.0002,
1661
+ "epoch": 3.29,
1662
+ "step": 125
1663
+ },
1664
+ {
1665
+ "eval_creative_writing_loss": 0.12550988793373108,
1666
+ "eval_creative_writing_score": -0.023599309846758842,
1667
+ "eval_creative_writing_brier_score": 0.023599309846758842,
1668
+ "eval_creative_writing_average_probability": 0.9722127318382263,
1669
+ "eval_creative_writing_accuracy": 0.97,
1670
+ "eval_creative_writing_probabilities": [
1671
+ 0.9999924898147583,
1672
+ 1.0,
1673
+ 0.9999974966049194,
1674
+ 1.0,
1675
+ 0.9999760389328003,
1676
+ 1.0,
1677
+ 1.0,
1678
+ 1.0,
1679
+ 0.004820940550416708,
1680
+ 1.0,
1681
+ 0.38430270552635193,
1682
+ 1.0,
1683
+ 1.0,
1684
+ 1.0,
1685
+ 0.9999996423721313,
1686
+ 1.0,
1687
+ 1.0,
1688
+ 1.0,
1689
+ 1.0,
1690
+ 1.0,
1691
+ 1.0,
1692
+ 1.0,
1693
+ 1.0,
1694
+ 1.0,
1695
+ 0.9995892643928528,
1696
+ 0.9999997615814209,
1697
+ 1.0,
1698
+ 1.0,
1699
+ 0.012191555462777615,
1700
+ 1.0,
1701
+ 1.0,
1702
+ 1.0,
1703
+ 1.0,
1704
+ 1.0,
1705
+ 1.0,
1706
+ 1.0,
1707
+ 0.9999853372573853,
1708
+ 1.0,
1709
+ 0.999993085861206,
1710
+ 0.9999998807907104,
1711
+ 1.0,
1712
+ 1.0,
1713
+ 1.0,
1714
+ 1.0,
1715
+ 0.9999395608901978,
1716
+ 1.0,
1717
+ 1.0,
1718
+ 1.0,
1719
+ 1.0,
1720
+ 1.0,
1721
+ 1.0,
1722
+ 1.0,
1723
+ 0.9998890161514282,
1724
+ 1.0,
1725
+ 0.9999903440475464,
1726
+ 1.0,
1727
+ 0.9879515767097473,
1728
+ 1.0,
1729
+ 0.9996850490570068,
1730
+ 1.0,
1731
+ 0.9710411429405212,
1732
+ 1.0,
1733
+ 1.0,
1734
+ 1.0,
1735
+ 0.9999980926513672,
1736
+ 1.0,
1737
+ 0.8841173648834229,
1738
+ 0.999834418296814,
1739
+ 0.9965130686759949,
1740
+ 1.0,
1741
+ 0.9834631085395813,
1742
+ 1.0,
1743
+ 1.0,
1744
+ 1.0,
1745
+ 1.0,
1746
+ 1.0,
1747
+ 0.9999334812164307,
1748
+ 1.0,
1749
+ 0.9999998807907104,
1750
+ 1.0,
1751
+ 1.0,
1752
+ 1.0,
1753
+ 0.9999955892562866,
1754
+ 1.0,
1755
+ 0.9997425675392151,
1756
+ 0.9999978542327881,
1757
+ 0.9998661279678345,
1758
+ 1.0,
1759
+ 1.0,
1760
+ 1.0,
1761
+ 1.0,
1762
+ 1.0,
1763
+ 1.0,
1764
+ 1.0,
1765
+ 1.0,
1766
+ 1.0,
1767
+ 0.9984672665596008,
1768
+ 0.9999994039535522,
1769
+ 0.9999991655349731,
1770
+ 1.0
1771
+ ],
1772
+ "eval_creative_writing_runtime": 31.3796,
1773
+ "eval_creative_writing_samples_per_second": 3.187,
1774
+ "eval_creative_writing_steps_per_second": 0.127,
1775
+ "epoch": 3.29,
1776
+ "step": 125
1777
+ },
1778
+ {
1779
+ "eval_biology_with_literary_style_loss": 1.4181084632873535,
1780
+ "eval_biology_with_literary_style_score": -0.22596853971481323,
1781
+ "eval_biology_with_literary_style_brier_score": 0.22596853971481323,
1782
+ "eval_biology_with_literary_style_average_probability": 0.7482656836509705,
1783
+ "eval_biology_with_literary_style_accuracy": 0.75,
1784
+ "eval_biology_with_literary_style_probabilities": [
1785
+ 0.9995549321174622,
1786
+ 0.9985436201095581,
1787
+ 0.9808439612388611,
1788
+ 0.9283376932144165,
1789
+ 0.9999039173126221,
1790
+ 0.9999103546142578,
1791
+ 0.0073650372214615345,
1792
+ 0.0004983382532373071,
1793
+ 3.531386028043926e-05,
1794
+ 0.9999995231628418,
1795
+ 0.9999997615814209,
1796
+ 0.9999420642852783,
1797
+ 0.7384650707244873,
1798
+ 0.46473434567451477,
1799
+ 0.9987931251525879,
1800
+ 0.9886658191680908,
1801
+ 0.008153733797371387,
1802
+ 0.09072274714708328,
1803
+ 1.0,
1804
+ 1.0,
1805
+ 0.9999998807907104,
1806
+ 0.9999436140060425,
1807
+ 0.9999998807907104,
1808
+ 0.9902008175849915,
1809
+ 0.9996306896209717,
1810
+ 0.9992589354515076,
1811
+ 0.9999074935913086,
1812
+ 0.00022981569054536521,
1813
+ 0.020188162103295326,
1814
+ 0.02801305055618286,
1815
+ 1.0,
1816
+ 0.9999865293502808,
1817
+ 0.9999995231628418,
1818
+ 1.0,
1819
+ 0.9999997615814209,
1820
+ 0.9999998807907104,
1821
+ 0.6005551218986511,
1822
+ 0.9702953696250916,
1823
+ 0.9998131394386292,
1824
+ 1.041131092449632e-08,
1825
+ 9.587202839611564e-06,
1826
+ 0.0011823305394500494,
1827
+ 0.0001448177790734917,
1828
+ 0.009138714522123337,
1829
+ 0.9979574680328369,
1830
+ 0.9991582632064819,
1831
+ 0.016484873369336128,
1832
+ 0.9999951124191284,
1833
+ 0.999769389629364,
1834
+ 0.9998552799224854,
1835
+ 0.9998192191123962,
1836
+ 0.8466882705688477,
1837
+ 0.9816719889640808,
1838
+ 0.9999973773956299,
1839
+ 0.9993175268173218,
1840
+ 0.9999885559082031,
1841
+ 0.9391415119171143,
1842
+ 0.1099478080868721,
1843
+ 0.7412080764770508,
1844
+ 0.18274378776550293,
1845
+ 0.999987006187439,
1846
+ 0.9898987412452698,
1847
+ 0.9998290538787842,
1848
+ 0.9999748468399048,
1849
+ 0.776476263999939,
1850
+ 0.9796139001846313,
1851
+ 0.9999961853027344,
1852
+ 0.9999972581863403,
1853
+ 0.9996709823608398,
1854
+ 0.9429153800010681,
1855
+ 0.8567234873771667,
1856
+ 0.9939687252044678,
1857
+ 0.00026437846827320755,
1858
+ 0.9818727374076843,
1859
+ 0.9595638513565063,
1860
+ 0.9999904632568359,
1861
+ 0.45074892044067383,
1862
+ 0.9988915324211121,
1863
+ 0.27005892992019653,
1864
+ 0.9760595560073853,
1865
+ 0.003961615264415741,
1866
+ 0.999995231628418,
1867
+ 0.9971864819526672,
1868
+ 0.9999188184738159,
1869
+ 1.0,
1870
+ 0.9999996423721313,
1871
+ 0.9999929666519165,
1872
+ 0.9993189573287964,
1873
+ 0.9999390840530396,
1874
+ 0.0004752865352202207,
1875
+ 0.9999654293060303,
1876
+ 1.0,
1877
+ 1.0,
1878
+ 0.9996334314346313,
1879
+ 1.0,
1880
+ 0.9999845027923584,
1881
+ 0.012308338657021523,
1882
+ 0.0005812794552184641,
1883
+ 0.0003732819459401071,
1884
+ 0.999724805355072
1885
+ ],
1886
+ "eval_biology_with_literary_style_runtime": 34.5294,
1887
+ "eval_biology_with_literary_style_samples_per_second": 2.896,
1888
+ "eval_biology_with_literary_style_steps_per_second": 0.116,
1889
+ "epoch": 3.29,
1890
+ "step": 125
1891
+ },
1892
+ {
1893
+ "loss": 0.0002,
1894
+ "learning_rate": 0.0002,
1895
+ "epoch": 3.32,
1896
+ "step": 126
1897
+ },
1898
+ {
1899
+ "loss": 0.0,
1900
+ "learning_rate": 0.0002,
1901
+ "epoch": 3.34,
1902
+ "step": 127
1903
+ },
1904
+ {
1905
+ "loss": 0.0001,
1906
+ "learning_rate": 0.0002,
1907
+ "epoch": 3.37,
1908
+ "step": 128
1909
+ },
1910
+ {
1911
+ "loss": 0.0008,
1912
+ "learning_rate": 0.0002,
1913
+ "epoch": 3.39,
1914
+ "step": 129
1915
+ },
1916
+ {
1917
+ "loss": 0.0002,
1918
+ "learning_rate": 0.0002,
1919
+ "epoch": 3.42,
1920
+ "step": 130
1921
+ },
1922
+ {
1923
+ "loss": 0.0,
1924
+ "learning_rate": 0.0002,
1925
+ "epoch": 3.45,
1926
+ "step": 131
1927
+ },
1928
+ {
1929
+ "loss": 0.0001,
1930
+ "learning_rate": 0.0002,
1931
+ "epoch": 3.47,
1932
+ "step": 132
1933
+ },
1934
+ {
1935
+ "loss": 0.0001,
1936
+ "learning_rate": 0.0002,
1937
+ "epoch": 3.5,
1938
+ "step": 133
1939
+ },
1940
+ {
1941
+ "loss": 0.0,
1942
+ "learning_rate": 0.0002,
1943
+ "epoch": 3.53,
1944
+ "step": 134
1945
+ },
1946
+ {
1947
+ "loss": 0.0,
1948
+ "learning_rate": 0.0002,
1949
+ "epoch": 3.55,
1950
+ "step": 135
1951
+ },
1952
+ {
1953
+ "loss": 0.0,
1954
+ "learning_rate": 0.0002,
1955
+ "epoch": 3.58,
1956
+ "step": 136
1957
+ },
1958
+ {
1959
+ "loss": 0.0,
1960
+ "learning_rate": 0.0002,
1961
+ "epoch": 3.61,
1962
+ "step": 137
1963
+ },
1964
+ {
1965
+ "loss": 0.0,
1966
+ "learning_rate": 0.0002,
1967
+ "epoch": 3.63,
1968
+ "step": 138
1969
+ },
1970
+ {
1971
+ "loss": 0.0001,
1972
+ "learning_rate": 0.0002,
1973
+ "epoch": 3.66,
1974
+ "step": 139
1975
+ },
1976
+ {
1977
+ "loss": 0.0,
1978
+ "learning_rate": 0.0002,
1979
+ "epoch": 3.68,
1980
+ "step": 140
1981
+ },
1982
+ {
1983
+ "loss": 0.0001,
1984
+ "learning_rate": 0.0002,
1985
+ "epoch": 3.71,
1986
+ "step": 141
1987
+ },
1988
+ {
1989
+ "loss": 0.0002,
1990
+ "learning_rate": 0.0002,
1991
+ "epoch": 3.74,
1992
+ "step": 142
1993
+ },
1994
+ {
1995
+ "loss": 0.0,
1996
+ "learning_rate": 0.0002,
1997
+ "epoch": 3.76,
1998
+ "step": 143
1999
+ },
2000
+ {
2001
+ "loss": 0.0,
2002
+ "learning_rate": 0.0002,
2003
+ "epoch": 3.79,
2004
+ "step": 144
2005
+ },
2006
+ {
2007
+ "loss": 0.0001,
2008
+ "learning_rate": 0.0002,
2009
+ "epoch": 3.82,
2010
+ "step": 145
2011
+ },
2012
+ {
2013
+ "loss": 0.0,
2014
+ "learning_rate": 0.0002,
2015
+ "epoch": 3.84,
2016
+ "step": 146
2017
+ },
2018
+ {
2019
+ "loss": 0.0,
2020
+ "learning_rate": 0.0002,
2021
+ "epoch": 3.87,
2022
+ "step": 147
2023
+ },
2024
+ {
2025
+ "loss": 0.0001,
2026
+ "learning_rate": 0.0002,
2027
+ "epoch": 3.89,
2028
+ "step": 148
2029
+ },
2030
+ {
2031
+ "loss": 0.0,
2032
+ "learning_rate": 0.0002,
2033
+ "epoch": 3.92,
2034
+ "step": 149
2035
+ },
2036
+ {
2037
+ "loss": 0.0,
2038
+ "learning_rate": 0.0002,
2039
+ "epoch": 3.95,
2040
+ "step": 150
2041
+ },
2042
+ {
2043
+ "eval_creative_writing_loss": 0.12696059048175812,
2044
+ "eval_creative_writing_score": -0.023474015295505524,
2045
+ "eval_creative_writing_brier_score": 0.023474015295505524,
2046
+ "eval_creative_writing_average_probability": 0.9725755453109741,
2047
+ "eval_creative_writing_accuracy": 0.97,
2048
+ "eval_creative_writing_probabilities": [
2049
+ 0.9999969005584717,
2050
+ 1.0,
2051
+ 0.9999990463256836,
2052
+ 1.0,
2053
+ 0.9999603033065796,
2054
+ 1.0,
2055
+ 1.0,
2056
+ 1.0,
2057
+ 0.0047756098210811615,
2058
+ 1.0,
2059
+ 0.3958245515823364,
2060
+ 1.0,
2061
+ 1.0,
2062
+ 1.0,
2063
+ 0.9999998807907104,
2064
+ 1.0,
2065
+ 1.0,
2066
+ 1.0,
2067
+ 1.0,
2068
+ 1.0,
2069
+ 1.0,
2070
+ 1.0,
2071
+ 1.0,
2072
+ 1.0,
2073
+ 0.9996107220649719,
2074
+ 0.9999997615814209,
2075
+ 1.0,
2076
+ 1.0,
2077
+ 0.009845593012869358,
2078
+ 1.0,
2079
+ 1.0,
2080
+ 1.0,
2081
+ 1.0,
2082
+ 1.0,
2083
+ 1.0,
2084
+ 1.0,
2085
+ 0.9999935626983643,
2086
+ 1.0,
2087
+ 0.999993085861206,
2088
+ 0.9999998807907104,
2089
+ 1.0,
2090
+ 1.0,
2091
+ 1.0,
2092
+ 1.0,
2093
+ 0.9999793767929077,
2094
+ 1.0,
2095
+ 1.0,
2096
+ 1.0,
2097
+ 1.0,
2098
+ 1.0,
2099
+ 1.0,
2100
+ 1.0,
2101
+ 0.9999181032180786,
2102
+ 1.0,
2103
+ 0.9999945163726807,
2104
+ 1.0,
2105
+ 0.9935240745544434,
2106
+ 1.0,
2107
+ 0.9997866749763489,
2108
+ 1.0,
2109
+ 0.9811118245124817,
2110
+ 1.0,
2111
+ 1.0,
2112
+ 1.0,
2113
+ 0.9999995231628418,
2114
+ 1.0,
2115
+ 0.8962162137031555,
2116
+ 0.9998764991760254,
2117
+ 0.9970778226852417,
2118
+ 1.0,
2119
+ 0.9822664856910706,
2120
+ 1.0,
2121
+ 1.0,
2122
+ 1.0,
2123
+ 1.0,
2124
+ 1.0,
2125
+ 0.9999604225158691,
2126
+ 1.0,
2127
+ 0.9999998807907104,
2128
+ 1.0,
2129
+ 1.0,
2130
+ 1.0,
2131
+ 0.9999983310699463,
2132
+ 1.0,
2133
+ 0.9996516704559326,
2134
+ 0.9999957084655762,
2135
+ 0.9999384880065918,
2136
+ 1.0,
2137
+ 1.0,
2138
+ 1.0,
2139
+ 1.0,
2140
+ 1.0,
2141
+ 1.0,
2142
+ 1.0,
2143
+ 1.0,
2144
+ 1.0,
2145
+ 0.9982588887214661,
2146
+ 0.9999991655349731,
2147
+ 0.9999997615814209,
2148
+ 1.0
2149
+ ],
2150
+ "eval_creative_writing_runtime": 31.3872,
2151
+ "eval_creative_writing_samples_per_second": 3.186,
2152
+ "eval_creative_writing_steps_per_second": 0.127,
2153
+ "epoch": 3.95,
2154
+ "step": 150
2155
+ },
2156
+ {
2157
+ "eval_biology_with_literary_style_loss": 1.508569598197937,
2158
+ "eval_biology_with_literary_style_score": -0.223800390958786,
2159
+ "eval_biology_with_literary_style_brier_score": 0.223800390958786,
2160
+ "eval_biology_with_literary_style_average_probability": 0.753645658493042,
2161
+ "eval_biology_with_literary_style_accuracy": 0.77,
2162
+ "eval_biology_with_literary_style_probabilities": [
2163
+ 0.9997276663780212,
2164
+ 0.9992485642433167,
2165
+ 0.9895316362380981,
2166
+ 0.9514719843864441,
2167
+ 0.9999614953994751,
2168
+ 0.9999442100524902,
2169
+ 0.004651335533708334,
2170
+ 0.0001869379193522036,
2171
+ 1.4444511180045083e-05,
2172
+ 0.9999998807907104,
2173
+ 1.0,
2174
+ 0.9999759197235107,
2175
+ 0.79854816198349,
2176
+ 0.5046471357345581,
2177
+ 0.9990099668502808,
2178
+ 0.99369215965271,
2179
+ 0.0043740225955843925,
2180
+ 0.05449388176202774,
2181
+ 1.0,
2182
+ 1.0,
2183
+ 1.0,
2184
+ 0.9999809265136719,
2185
+ 1.0,
2186
+ 0.99128657579422,
2187
+ 0.9998859167098999,
2188
+ 0.999728262424469,
2189
+ 0.9999839067459106,
2190
+ 0.00016595126362517476,
2191
+ 0.02103896625339985,
2192
+ 0.024976782500743866,
2193
+ 1.0,
2194
+ 0.9999954700469971,
2195
+ 0.9999998807907104,
2196
+ 1.0,
2197
+ 1.0,
2198
+ 1.0,
2199
+ 0.6514801383018494,
2200
+ 0.9780262112617493,
2201
+ 0.9999254941940308,
2202
+ 2.189738612656811e-09,
2203
+ 3.2669267966412008e-06,
2204
+ 0.0008265993092209101,
2205
+ 9.858178964350373e-05,
2206
+ 0.007948733866214752,
2207
+ 0.9988512992858887,
2208
+ 0.9993112087249756,
2209
+ 0.007991177029907703,
2210
+ 0.9999977350234985,
2211
+ 0.9998981952667236,
2212
+ 0.9999642372131348,
2213
+ 0.9999281167984009,
2214
+ 0.8989755511283875,
2215
+ 0.9909796714782715,
2216
+ 0.9999992847442627,
2217
+ 0.9996218681335449,
2218
+ 0.9999951124191284,
2219
+ 0.9505088925361633,
2220
+ 0.1275900900363922,
2221
+ 0.8378230929374695,
2222
+ 0.1916273832321167,
2223
+ 0.9999955892562866,
2224
+ 0.9941998720169067,
2225
+ 0.9999337196350098,
2226
+ 0.9999921321868896,
2227
+ 0.7730574607849121,
2228
+ 0.9855798482894897,
2229
+ 0.9999992847442627,
2230
+ 0.9999995231628418,
2231
+ 0.9999063014984131,
2232
+ 0.9654306769371033,
2233
+ 0.9086850881576538,
2234
+ 0.9970625042915344,
2235
+ 0.00011868889123434201,
2236
+ 0.9871053695678711,
2237
+ 0.962754487991333,
2238
+ 0.999996542930603,
2239
+ 0.5310143828392029,
2240
+ 0.9996671676635742,
2241
+ 0.28350409865379333,
2242
+ 0.9845860600471497,
2243
+ 0.002773666987195611,
2244
+ 0.9999983310699463,
2245
+ 0.9990894794464111,
2246
+ 0.999972939491272,
2247
+ 1.0,
2248
+ 0.9999998807907104,
2249
+ 0.9999982118606567,
2250
+ 0.9997745156288147,
2251
+ 0.9999819993972778,
2252
+ 0.00044673527008853853,
2253
+ 0.9999911785125732,
2254
+ 1.0,
2255
+ 1.0,
2256
+ 0.9998553991317749,
2257
+ 1.0,
2258
+ 0.9999963045120239,
2259
+ 0.01165656466037035,
2260
+ 0.0004307126218918711,
2261
+ 0.00019968363631051034,
2262
+ 0.9999244213104248
2263
+ ],
2264
+ "eval_biology_with_literary_style_runtime": 34.5257,
2265
+ "eval_biology_with_literary_style_samples_per_second": 2.896,
2266
+ "eval_biology_with_literary_style_steps_per_second": 0.116,
2267
+ "epoch": 3.95,
2268
+ "step": 150
2269
+ },
2270
+ {
2271
+ "train_runtime": 3937.8753,
2272
+ "train_samples_per_second": 1.219,
2273
+ "train_steps_per_second": 0.038,
2274
+ "total_flos": 0.0,
2275
+ "train_loss": 0.053941556412501084,
2276
+ "epoch": 3.95,
2277
+ "step": 150
2278
+ }
2279
+ ]