Training in progress, step 488, checkpoint
Browse files- last-checkpoint/chat_template.jinja +1 -0
- last-checkpoint/config.json +62 -0
- last-checkpoint/generation_config.json +15 -0
- last-checkpoint/model.safetensors +3 -0
- last-checkpoint/optimizer.pt +3 -0
- last-checkpoint/rng_state.pth +3 -0
- last-checkpoint/scheduler.pt +3 -0
- last-checkpoint/tokenizer.json +0 -0
- last-checkpoint/tokenizer_config.json +18 -0
- last-checkpoint/trainer_state.json +3458 -0
- last-checkpoint/training_args.bin +3 -0
last-checkpoint/chat_template.jinja
ADDED
|
@@ -0,0 +1 @@
|
|
|
|
|
|
|
| 1 |
+
{% if messages[0]['role'] == 'system' %}{% set loop_messages = messages[1:] %}{% set system_message = messages[0]['content'] %}<｜hy_begin▁of▁sentence｜>{{ system_message }}<｜hy_place▁holder▁no▁3｜>{% else %}{% set loop_messages = messages %}<｜hy_begin▁of▁sentence｜>{% endif %}{% for message in loop_messages %}{% if message['role'] == 'user' %}<｜hy_User｜>{{ message['content'] }}{% elif message['role'] == 'assistant' %}<｜hy_Assistant｜>{{ message['content'] }}<｜hy_place▁holder▁no▁2｜>{% endif %}{% endfor %}{% if add_generation_prompt %}<｜hy_Assistant｜>{% else %}<｜hy_place▁holder▁no▁8｜>{% endif %}
|
last-checkpoint/config.json
ADDED
|
@@ -0,0 +1,62 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"add_classification_head": false,
|
| 3 |
+
"architectures": [
|
| 4 |
+
"HunYuanDenseV1ForCausalLM"
|
| 5 |
+
],
|
| 6 |
+
"attention_bias": false,
|
| 7 |
+
"attention_dropout": 0.0,
|
| 8 |
+
"attention_head_dim": 128,
|
| 9 |
+
"bos_token_id": 120000,
|
| 10 |
+
"cla_share_factor": 2,
|
| 11 |
+
"class_num": 0,
|
| 12 |
+
"dense_list": [
|
| 13 |
+
2048,
|
| 14 |
+
0
|
| 15 |
+
],
|
| 16 |
+
"dtype": "bfloat16",
|
| 17 |
+
"eos_token_id": 120020,
|
| 18 |
+
"head_dim": 128,
|
| 19 |
+
"hidden_act": "silu",
|
| 20 |
+
"hidden_size": 2048,
|
| 21 |
+
"im_end_id": 5,
|
| 22 |
+
"im_newline_id": 11,
|
| 23 |
+
"im_start_id": 4,
|
| 24 |
+
"initializer_range": 0.02,
|
| 25 |
+
"intermediate_size": 6144,
|
| 26 |
+
"mask_init_id": 12,
|
| 27 |
+
"max_position_embeddings": 262144,
|
| 28 |
+
"mlp_bias": false,
|
| 29 |
+
"model_type": "hunyuan_v1_dense",
|
| 30 |
+
"norm_type": "rms",
|
| 31 |
+
"num_attention_heads": 16,
|
| 32 |
+
"num_hidden_layers": 32,
|
| 33 |
+
"num_key_value_heads": 4,
|
| 34 |
+
"org_vocab_size": 120818,
|
| 35 |
+
"pad_id": 120002,
|
| 36 |
+
"pad_token_id": 120002,
|
| 37 |
+
"pool_type": "last",
|
| 38 |
+
"pretraining_tp": 1,
|
| 39 |
+
"rms_norm_eps": 1e-05,
|
| 40 |
+
"rope_parameters": {
|
| 41 |
+
"alpha": 1000.0,
|
| 42 |
+
"beta_fast": 32,
|
| 43 |
+
"beta_slow": 1,
|
| 44 |
+
"factor": 1.0,
|
| 45 |
+
"mscale": 1.0,
|
| 46 |
+
"mscale_all_dim": 1.0,
|
| 47 |
+
"rope_theta": 10000.0,
|
| 48 |
+
"rope_type": "dynamic",
|
| 49 |
+
"type": "dynamic"
|
| 50 |
+
},
|
| 51 |
+
"sep_token_id": 120007,
|
| 52 |
+
"text_end_id": 7,
|
| 53 |
+
"text_start_id": 6,
|
| 54 |
+
"tie_word_embeddings": true,
|
| 55 |
+
"transformers_version": "5.0.0rc0",
|
| 56 |
+
"unsloth_version": "2026.1.3",
|
| 57 |
+
"use_cache": false,
|
| 58 |
+
"use_cla": false,
|
| 59 |
+
"use_qk_norm": true,
|
| 60 |
+
"use_rotary_pos_emb": true,
|
| 61 |
+
"vocab_size": 120821
|
| 62 |
+
}
|
last-checkpoint/generation_config.json
ADDED
|
@@ -0,0 +1,15 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token_id": 120000,
|
| 3 |
+
"do_sample": true,
|
| 4 |
+
"eos_token_id": [
|
| 5 |
+
120020,
|
| 6 |
+
120001
|
| 7 |
+
],
|
| 8 |
+
"max_length": 262144,
|
| 9 |
+
"pad_token_id": 120002,
|
| 10 |
+
"repetition_penalty": 1.05,
|
| 11 |
+
"temperature": 0.7,
|
| 12 |
+
"top_k": 20,
|
| 13 |
+
"top_p": 0.8,
|
| 14 |
+
"transformers_version": "5.0.0rc0"
|
| 15 |
+
}
|
last-checkpoint/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:4db3a62c1819bacbea34efe1f5ee87aa9e2b7d38b490e4bb8c67fdc8c28b4c90
|
| 3 |
+
size 3582214344
|
last-checkpoint/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:01d6a75837e5b5d94701ceefd4961b2071c0c032d269457ab8ef69bfafc7ee43
|
| 3 |
+
size 5116251049
|
last-checkpoint/rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:9ccb8eeb935749fc43744e0a5eeacdf6f0f10253be15266a497cbca0ffaa2573
|
| 3 |
+
size 14645
|
last-checkpoint/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c59ef4318d3748fb59ca0a03e06e1c7022c3904581b4aac2132ae0cad48d0bbe
|
| 3 |
+
size 1465
|
last-checkpoint/tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
last-checkpoint/tokenizer_config.json
ADDED
|
@@ -0,0 +1,18 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"additional_special_tokens": null,
|
| 3 |
+
"backend": "tokenizers",
|
| 4 |
+
"bos_token": "<｜hy_begin▁of▁sentence｜>",
|
| 5 |
+
"clean_up_tokenization_spaces": true,
|
| 6 |
+
"eos_token": "<｜hy_place▁holder▁no▁2｜>",
|
| 7 |
+
"extra_special_tokens": [
|
| 8 |
+
"<|source|>",
|
| 9 |
+
"<|translation|>",
|
| 10 |
+
"<|im_end|>"
|
| 11 |
+
],
|
| 12 |
+
"is_local": false,
|
| 13 |
+
"model_max_length": 1000000000000000019884624838656,
|
| 14 |
+
"model_specific_special_tokens": {},
|
| 15 |
+
"pad_token": "<｜hy_▁pad▁｜>",
|
| 16 |
+
"padding_side": "right",
|
| 17 |
+
"tokenizer_class": "TokenizersBackend"
|
| 18 |
+
}
|
last-checkpoint/trainer_state.json
ADDED
|
@@ -0,0 +1,3458 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": null,
|
| 3 |
+
"best_metric": null,
|
| 4 |
+
"best_model_checkpoint": null,
|
| 5 |
+
"epoch": 0.2002051282051282,
|
| 6 |
+
"eval_steps": 488,
|
| 7 |
+
"global_step": 488,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.00041025641025641023,
|
| 14 |
+
"grad_norm": 11.5625,
|
| 15 |
+
"learning_rate": 0.0,
|
| 16 |
+
"loss": 1.5559172630310059,
|
| 17 |
+
"step": 1
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.0008205128205128205,
|
| 21 |
+
"grad_norm": 10.9375,
|
| 22 |
+
"learning_rate": 1.0204081632653061e-07,
|
| 23 |
+
"loss": 1.5650700330734253,
|
| 24 |
+
"step": 2
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.0012307692307692308,
|
| 28 |
+
"grad_norm": 9.875,
|
| 29 |
+
"learning_rate": 2.0408163265306121e-07,
|
| 30 |
+
"loss": 1.5351619720458984,
|
| 31 |
+
"step": 3
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.001641025641025641,
|
| 35 |
+
"grad_norm": 12.5625,
|
| 36 |
+
"learning_rate": 3.0612244897959183e-07,
|
| 37 |
+
"loss": 1.5397462844848633,
|
| 38 |
+
"step": 4
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.0020512820512820513,
|
| 42 |
+
"grad_norm": 11.4375,
|
| 43 |
+
"learning_rate": 4.0816326530612243e-07,
|
| 44 |
+
"loss": 1.5788651704788208,
|
| 45 |
+
"step": 5
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.0024615384615384616,
|
| 49 |
+
"grad_norm": 10.3125,
|
| 50 |
+
"learning_rate": 5.102040816326531e-07,
|
| 51 |
+
"loss": 1.51237154006958,
|
| 52 |
+
"step": 6
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"epoch": 0.002871794871794872,
|
| 56 |
+
"grad_norm": 10.875,
|
| 57 |
+
"learning_rate": 6.122448979591837e-07,
|
| 58 |
+
"loss": 1.5072423219680786,
|
| 59 |
+
"step": 7
|
| 60 |
+
},
|
| 61 |
+
{
|
| 62 |
+
"epoch": 0.003282051282051282,
|
| 63 |
+
"grad_norm": 11.3125,
|
| 64 |
+
"learning_rate": 7.142857142857143e-07,
|
| 65 |
+
"loss": 1.5596964359283447,
|
| 66 |
+
"step": 8
|
| 67 |
+
},
|
| 68 |
+
{
|
| 69 |
+
"epoch": 0.0036923076923076922,
|
| 70 |
+
"grad_norm": 12.0625,
|
| 71 |
+
"learning_rate": 8.163265306122449e-07,
|
| 72 |
+
"loss": 1.5828701257705688,
|
| 73 |
+
"step": 9
|
| 74 |
+
},
|
| 75 |
+
{
|
| 76 |
+
"epoch": 0.0041025641025641026,
|
| 77 |
+
"grad_norm": 12.5625,
|
| 78 |
+
"learning_rate": 9.183673469387756e-07,
|
| 79 |
+
"loss": 1.5764329433441162,
|
| 80 |
+
"step": 10
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 0.0045128205128205125,
|
| 84 |
+
"grad_norm": 9.25,
|
| 85 |
+
"learning_rate": 1.0204081632653063e-06,
|
| 86 |
+
"loss": 1.461578369140625,
|
| 87 |
+
"step": 11
|
| 88 |
+
},
|
| 89 |
+
{
|
| 90 |
+
"epoch": 0.004923076923076923,
|
| 91 |
+
"grad_norm": 11.25,
|
| 92 |
+
"learning_rate": 1.122448979591837e-06,
|
| 93 |
+
"loss": 1.545849323272705,
|
| 94 |
+
"step": 12
|
| 95 |
+
},
|
| 96 |
+
{
|
| 97 |
+
"epoch": 0.005333333333333333,
|
| 98 |
+
"grad_norm": 10.875,
|
| 99 |
+
"learning_rate": 1.2244897959183673e-06,
|
| 100 |
+
"loss": 1.4574928283691406,
|
| 101 |
+
"step": 13
|
| 102 |
+
},
|
| 103 |
+
{
|
| 104 |
+
"epoch": 0.005743589743589744,
|
| 105 |
+
"grad_norm": 10.75,
|
| 106 |
+
"learning_rate": 1.3265306122448982e-06,
|
| 107 |
+
"loss": 1.5373510122299194,
|
| 108 |
+
"step": 14
|
| 109 |
+
},
|
| 110 |
+
{
|
| 111 |
+
"epoch": 0.006153846153846154,
|
| 112 |
+
"grad_norm": 12.9375,
|
| 113 |
+
"learning_rate": 1.4285714285714286e-06,
|
| 114 |
+
"loss": 1.4119579792022705,
|
| 115 |
+
"step": 15
|
| 116 |
+
},
|
| 117 |
+
{
|
| 118 |
+
"epoch": 0.006564102564102564,
|
| 119 |
+
"grad_norm": 10.125,
|
| 120 |
+
"learning_rate": 1.5306122448979593e-06,
|
| 121 |
+
"loss": 1.549330711364746,
|
| 122 |
+
"step": 16
|
| 123 |
+
},
|
| 124 |
+
{
|
| 125 |
+
"epoch": 0.0069743589743589745,
|
| 126 |
+
"grad_norm": 9.5625,
|
| 127 |
+
"learning_rate": 1.6326530612244897e-06,
|
| 128 |
+
"loss": 1.5673086643218994,
|
| 129 |
+
"step": 17
|
| 130 |
+
},
|
| 131 |
+
{
|
| 132 |
+
"epoch": 0.0073846153846153844,
|
| 133 |
+
"grad_norm": 9.1875,
|
| 134 |
+
"learning_rate": 1.7346938775510206e-06,
|
| 135 |
+
"loss": 1.357297420501709,
|
| 136 |
+
"step": 18
|
| 137 |
+
},
|
| 138 |
+
{
|
| 139 |
+
"epoch": 0.007794871794871795,
|
| 140 |
+
"grad_norm": 10.0625,
|
| 141 |
+
"learning_rate": 1.8367346938775512e-06,
|
| 142 |
+
"loss": 1.4897398948669434,
|
| 143 |
+
"step": 19
|
| 144 |
+
},
|
| 145 |
+
{
|
| 146 |
+
"epoch": 0.008205128205128205,
|
| 147 |
+
"grad_norm": 10.75,
|
| 148 |
+
"learning_rate": 1.938775510204082e-06,
|
| 149 |
+
"loss": 1.496138095855713,
|
| 150 |
+
"step": 20
|
| 151 |
+
},
|
| 152 |
+
{
|
| 153 |
+
"epoch": 0.008615384615384615,
|
| 154 |
+
"grad_norm": 9.8125,
|
| 155 |
+
"learning_rate": 2.0408163265306125e-06,
|
| 156 |
+
"loss": 1.3633430004119873,
|
| 157 |
+
"step": 21
|
| 158 |
+
},
|
| 159 |
+
{
|
| 160 |
+
"epoch": 0.009025641025641025,
|
| 161 |
+
"grad_norm": 10.125,
|
| 162 |
+
"learning_rate": 2.1428571428571427e-06,
|
| 163 |
+
"loss": 1.5695936679840088,
|
| 164 |
+
"step": 22
|
| 165 |
+
},
|
| 166 |
+
{
|
| 167 |
+
"epoch": 0.009435897435897437,
|
| 168 |
+
"grad_norm": 8.875,
|
| 169 |
+
"learning_rate": 2.244897959183674e-06,
|
| 170 |
+
"loss": 1.4395489692687988,
|
| 171 |
+
"step": 23
|
| 172 |
+
},
|
| 173 |
+
{
|
| 174 |
+
"epoch": 0.009846153846153846,
|
| 175 |
+
"grad_norm": 9.125,
|
| 176 |
+
"learning_rate": 2.3469387755102044e-06,
|
| 177 |
+
"loss": 1.5492465496063232,
|
| 178 |
+
"step": 24
|
| 179 |
+
},
|
| 180 |
+
{
|
| 181 |
+
"epoch": 0.010256410256410256,
|
| 182 |
+
"grad_norm": 10.625,
|
| 183 |
+
"learning_rate": 2.4489795918367347e-06,
|
| 184 |
+
"loss": 1.3773704767227173,
|
| 185 |
+
"step": 25
|
| 186 |
+
},
|
| 187 |
+
{
|
| 188 |
+
"epoch": 0.010666666666666666,
|
| 189 |
+
"grad_norm": 8.9375,
|
| 190 |
+
"learning_rate": 2.5510204081632657e-06,
|
| 191 |
+
"loss": 1.4431257247924805,
|
| 192 |
+
"step": 26
|
| 193 |
+
},
|
| 194 |
+
{
|
| 195 |
+
"epoch": 0.011076923076923076,
|
| 196 |
+
"grad_norm": 8.75,
|
| 197 |
+
"learning_rate": 2.6530612244897964e-06,
|
| 198 |
+
"loss": 1.4262057542800903,
|
| 199 |
+
"step": 27
|
| 200 |
+
},
|
| 201 |
+
{
|
| 202 |
+
"epoch": 0.011487179487179488,
|
| 203 |
+
"grad_norm": 9.5625,
|
| 204 |
+
"learning_rate": 2.7551020408163266e-06,
|
| 205 |
+
"loss": 1.4837062358856201,
|
| 206 |
+
"step": 28
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 0.011897435897435898,
|
| 210 |
+
"grad_norm": 8.8125,
|
| 211 |
+
"learning_rate": 2.8571428571428573e-06,
|
| 212 |
+
"loss": 1.3567763566970825,
|
| 213 |
+
"step": 29
|
| 214 |
+
},
|
| 215 |
+
{
|
| 216 |
+
"epoch": 0.012307692307692308,
|
| 217 |
+
"grad_norm": 8.75,
|
| 218 |
+
"learning_rate": 2.959183673469388e-06,
|
| 219 |
+
"loss": 1.4458659887313843,
|
| 220 |
+
"step": 30
|
| 221 |
+
},
|
| 222 |
+
{
|
| 223 |
+
"epoch": 0.012717948717948718,
|
| 224 |
+
"grad_norm": 9.4375,
|
| 225 |
+
"learning_rate": 3.0612244897959185e-06,
|
| 226 |
+
"loss": 1.6264022588729858,
|
| 227 |
+
"step": 31
|
| 228 |
+
},
|
| 229 |
+
{
|
| 230 |
+
"epoch": 0.013128205128205127,
|
| 231 |
+
"grad_norm": 8.125,
|
| 232 |
+
"learning_rate": 3.1632653061224496e-06,
|
| 233 |
+
"loss": 1.4201624393463135,
|
| 234 |
+
"step": 32
|
| 235 |
+
},
|
| 236 |
+
{
|
| 237 |
+
"epoch": 0.01353846153846154,
|
| 238 |
+
"grad_norm": 7.5,
|
| 239 |
+
"learning_rate": 3.2653061224489794e-06,
|
| 240 |
+
"loss": 1.3213391304016113,
|
| 241 |
+
"step": 33
|
| 242 |
+
},
|
| 243 |
+
{
|
| 244 |
+
"epoch": 0.013948717948717949,
|
| 245 |
+
"grad_norm": 6.625,
|
| 246 |
+
"learning_rate": 3.3673469387755105e-06,
|
| 247 |
+
"loss": 1.314402461051941,
|
| 248 |
+
"step": 34
|
| 249 |
+
},
|
| 250 |
+
{
|
| 251 |
+
"epoch": 0.014358974358974359,
|
| 252 |
+
"grad_norm": 6.53125,
|
| 253 |
+
"learning_rate": 3.469387755102041e-06,
|
| 254 |
+
"loss": 1.3699743747711182,
|
| 255 |
+
"step": 35
|
| 256 |
+
},
|
| 257 |
+
{
|
| 258 |
+
"epoch": 0.014769230769230769,
|
| 259 |
+
"grad_norm": 7.4375,
|
| 260 |
+
"learning_rate": 3.5714285714285718e-06,
|
| 261 |
+
"loss": 1.4822998046875,
|
| 262 |
+
"step": 36
|
| 263 |
+
},
|
| 264 |
+
{
|
| 265 |
+
"epoch": 0.015179487179487179,
|
| 266 |
+
"grad_norm": 6.4375,
|
| 267 |
+
"learning_rate": 3.6734693877551024e-06,
|
| 268 |
+
"loss": 1.363675594329834,
|
| 269 |
+
"step": 37
|
| 270 |
+
},
|
| 271 |
+
{
|
| 272 |
+
"epoch": 0.01558974358974359,
|
| 273 |
+
"grad_norm": 6.8125,
|
| 274 |
+
"learning_rate": 3.7755102040816327e-06,
|
| 275 |
+
"loss": 1.3546408414840698,
|
| 276 |
+
"step": 38
|
| 277 |
+
},
|
| 278 |
+
{
|
| 279 |
+
"epoch": 0.016,
|
| 280 |
+
"grad_norm": 6.0,
|
| 281 |
+
"learning_rate": 3.877551020408164e-06,
|
| 282 |
+
"loss": 1.476768970489502,
|
| 283 |
+
"step": 39
|
| 284 |
+
},
|
| 285 |
+
{
|
| 286 |
+
"epoch": 0.01641025641025641,
|
| 287 |
+
"grad_norm": 6.34375,
|
| 288 |
+
"learning_rate": 3.979591836734694e-06,
|
| 289 |
+
"loss": 1.3525794744491577,
|
| 290 |
+
"step": 40
|
| 291 |
+
},
|
| 292 |
+
{
|
| 293 |
+
"epoch": 0.01682051282051282,
|
| 294 |
+
"grad_norm": 6.53125,
|
| 295 |
+
"learning_rate": 4.081632653061225e-06,
|
| 296 |
+
"loss": 1.4766862392425537,
|
| 297 |
+
"step": 41
|
| 298 |
+
},
|
| 299 |
+
{
|
| 300 |
+
"epoch": 0.01723076923076923,
|
| 301 |
+
"grad_norm": 5.9375,
|
| 302 |
+
"learning_rate": 4.183673469387755e-06,
|
| 303 |
+
"loss": 1.3278095722198486,
|
| 304 |
+
"step": 42
|
| 305 |
+
},
|
| 306 |
+
{
|
| 307 |
+
"epoch": 0.01764102564102564,
|
| 308 |
+
"grad_norm": 6.03125,
|
| 309 |
+
"learning_rate": 4.2857142857142855e-06,
|
| 310 |
+
"loss": 1.3326995372772217,
|
| 311 |
+
"step": 43
|
| 312 |
+
},
|
| 313 |
+
{
|
| 314 |
+
"epoch": 0.01805128205128205,
|
| 315 |
+
"grad_norm": 6.0625,
|
| 316 |
+
"learning_rate": 4.3877551020408165e-06,
|
| 317 |
+
"loss": 1.5898628234863281,
|
| 318 |
+
"step": 44
|
| 319 |
+
},
|
| 320 |
+
{
|
| 321 |
+
"epoch": 0.018461538461538463,
|
| 322 |
+
"grad_norm": 6.15625,
|
| 323 |
+
"learning_rate": 4.489795918367348e-06,
|
| 324 |
+
"loss": 1.4401524066925049,
|
| 325 |
+
"step": 45
|
| 326 |
+
},
|
| 327 |
+
{
|
| 328 |
+
"epoch": 0.018871794871794873,
|
| 329 |
+
"grad_norm": 5.9375,
|
| 330 |
+
"learning_rate": 4.591836734693878e-06,
|
| 331 |
+
"loss": 1.5021121501922607,
|
| 332 |
+
"step": 46
|
| 333 |
+
},
|
| 334 |
+
{
|
| 335 |
+
"epoch": 0.019282051282051283,
|
| 336 |
+
"grad_norm": 5.6875,
|
| 337 |
+
"learning_rate": 4.693877551020409e-06,
|
| 338 |
+
"loss": 1.3678241968154907,
|
| 339 |
+
"step": 47
|
| 340 |
+
},
|
| 341 |
+
{
|
| 342 |
+
"epoch": 0.019692307692307693,
|
| 343 |
+
"grad_norm": 5.21875,
|
| 344 |
+
"learning_rate": 4.795918367346939e-06,
|
| 345 |
+
"loss": 1.3304778337478638,
|
| 346 |
+
"step": 48
|
| 347 |
+
},
|
| 348 |
+
{
|
| 349 |
+
"epoch": 0.020102564102564103,
|
| 350 |
+
"grad_norm": 6.15625,
|
| 351 |
+
"learning_rate": 4.897959183673469e-06,
|
| 352 |
+
"loss": 1.4725492000579834,
|
| 353 |
+
"step": 49
|
| 354 |
+
},
|
| 355 |
+
{
|
| 356 |
+
"epoch": 0.020512820512820513,
|
| 357 |
+
"grad_norm": 4.84375,
|
| 358 |
+
"learning_rate": 5e-06,
|
| 359 |
+
"loss": 1.236943006515503,
|
| 360 |
+
"step": 50
|
| 361 |
+
},
|
| 362 |
+
{
|
| 363 |
+
"epoch": 0.020923076923076923,
|
| 364 |
+
"grad_norm": 5.125,
|
| 365 |
+
"learning_rate": 4.997907074089577e-06,
|
| 366 |
+
"loss": 1.3986563682556152,
|
| 367 |
+
"step": 51
|
| 368 |
+
},
|
| 369 |
+
{
|
| 370 |
+
"epoch": 0.021333333333333333,
|
| 371 |
+
"grad_norm": 4.8125,
|
| 372 |
+
"learning_rate": 4.995814148179155e-06,
|
| 373 |
+
"loss": 1.3600258827209473,
|
| 374 |
+
"step": 52
|
| 375 |
+
},
|
| 376 |
+
{
|
| 377 |
+
"epoch": 0.021743589743589743,
|
| 378 |
+
"grad_norm": 5.71875,
|
| 379 |
+
"learning_rate": 4.993721222268732e-06,
|
| 380 |
+
"loss": 1.3663980960845947,
|
| 381 |
+
"step": 53
|
| 382 |
+
},
|
| 383 |
+
{
|
| 384 |
+
"epoch": 0.022153846153846152,
|
| 385 |
+
"grad_norm": 4.75,
|
| 386 |
+
"learning_rate": 4.991628296358309e-06,
|
| 387 |
+
"loss": 1.4747812747955322,
|
| 388 |
+
"step": 54
|
| 389 |
+
},
|
| 390 |
+
{
|
| 391 |
+
"epoch": 0.022564102564102566,
|
| 392 |
+
"grad_norm": 5.0,
|
| 393 |
+
"learning_rate": 4.989535370447887e-06,
|
| 394 |
+
"loss": 1.3586950302124023,
|
| 395 |
+
"step": 55
|
| 396 |
+
},
|
| 397 |
+
{
|
| 398 |
+
"epoch": 0.022974358974358976,
|
| 399 |
+
"grad_norm": 4.59375,
|
| 400 |
+
"learning_rate": 4.987442444537464e-06,
|
| 401 |
+
"loss": 1.301880121231079,
|
| 402 |
+
"step": 56
|
| 403 |
+
},
|
| 404 |
+
{
|
| 405 |
+
"epoch": 0.023384615384615386,
|
| 406 |
+
"grad_norm": 4.46875,
|
| 407 |
+
"learning_rate": 4.985349518627041e-06,
|
| 408 |
+
"loss": 1.2627747058868408,
|
| 409 |
+
"step": 57
|
| 410 |
+
},
|
| 411 |
+
{
|
| 412 |
+
"epoch": 0.023794871794871796,
|
| 413 |
+
"grad_norm": 4.6875,
|
| 414 |
+
"learning_rate": 4.983256592716618e-06,
|
| 415 |
+
"loss": 1.3262357711791992,
|
| 416 |
+
"step": 58
|
| 417 |
+
},
|
| 418 |
+
{
|
| 419 |
+
"epoch": 0.024205128205128205,
|
| 420 |
+
"grad_norm": 4.90625,
|
| 421 |
+
"learning_rate": 4.981163666806195e-06,
|
| 422 |
+
"loss": 1.3234881162643433,
|
| 423 |
+
"step": 59
|
| 424 |
+
},
|
| 425 |
+
{
|
| 426 |
+
"epoch": 0.024615384615384615,
|
| 427 |
+
"grad_norm": 5.0625,
|
| 428 |
+
"learning_rate": 4.9790707408957726e-06,
|
| 429 |
+
"loss": 1.2444298267364502,
|
| 430 |
+
"step": 60
|
| 431 |
+
},
|
| 432 |
+
{
|
| 433 |
+
"epoch": 0.025025641025641025,
|
| 434 |
+
"grad_norm": 5.0625,
|
| 435 |
+
"learning_rate": 4.97697781498535e-06,
|
| 436 |
+
"loss": 1.3685760498046875,
|
| 437 |
+
"step": 61
|
| 438 |
+
},
|
| 439 |
+
{
|
| 440 |
+
"epoch": 0.025435897435897435,
|
| 441 |
+
"grad_norm": 4.75,
|
| 442 |
+
"learning_rate": 4.974884889074927e-06,
|
| 443 |
+
"loss": 1.196195363998413,
|
| 444 |
+
"step": 62
|
| 445 |
+
},
|
| 446 |
+
{
|
| 447 |
+
"epoch": 0.025846153846153845,
|
| 448 |
+
"grad_norm": 4.6875,
|
| 449 |
+
"learning_rate": 4.972791963164505e-06,
|
| 450 |
+
"loss": 1.314781904220581,
|
| 451 |
+
"step": 63
|
| 452 |
+
},
|
| 453 |
+
{
|
| 454 |
+
"epoch": 0.026256410256410255,
|
| 455 |
+
"grad_norm": 4.375,
|
| 456 |
+
"learning_rate": 4.970699037254081e-06,
|
| 457 |
+
"loss": 1.1586852073669434,
|
| 458 |
+
"step": 64
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 0.02666666666666667,
|
| 462 |
+
"grad_norm": 4.3125,
|
| 463 |
+
"learning_rate": 4.968606111343659e-06,
|
| 464 |
+
"loss": 1.2497059106826782,
|
| 465 |
+
"step": 65
|
| 466 |
+
},
|
| 467 |
+
{
|
| 468 |
+
"epoch": 0.02707692307692308,
|
| 469 |
+
"grad_norm": 5.15625,
|
| 470 |
+
"learning_rate": 4.966513185433236e-06,
|
| 471 |
+
"loss": 1.27781081199646,
|
| 472 |
+
"step": 66
|
| 473 |
+
},
|
| 474 |
+
{
|
| 475 |
+
"epoch": 0.027487179487179488,
|
| 476 |
+
"grad_norm": 4.4375,
|
| 477 |
+
"learning_rate": 4.964420259522813e-06,
|
| 478 |
+
"loss": 1.2858244180679321,
|
| 479 |
+
"step": 67
|
| 480 |
+
},
|
| 481 |
+
{
|
| 482 |
+
"epoch": 0.027897435897435898,
|
| 483 |
+
"grad_norm": 4.53125,
|
| 484 |
+
"learning_rate": 4.96232733361239e-06,
|
| 485 |
+
"loss": 1.330633282661438,
|
| 486 |
+
"step": 68
|
| 487 |
+
},
|
| 488 |
+
{
|
| 489 |
+
"epoch": 0.028307692307692308,
|
| 490 |
+
"grad_norm": 5.15625,
|
| 491 |
+
"learning_rate": 4.960234407701968e-06,
|
| 492 |
+
"loss": 1.3277004957199097,
|
| 493 |
+
"step": 69
|
| 494 |
+
},
|
| 495 |
+
{
|
| 496 |
+
"epoch": 0.028717948717948718,
|
| 497 |
+
"grad_norm": 4.5,
|
| 498 |
+
"learning_rate": 4.9581414817915455e-06,
|
| 499 |
+
"loss": 1.32395339012146,
|
| 500 |
+
"step": 70
|
| 501 |
+
},
|
| 502 |
+
{
|
| 503 |
+
"epoch": 0.029128205128205128,
|
| 504 |
+
"grad_norm": 4.75,
|
| 505 |
+
"learning_rate": 4.956048555881122e-06,
|
| 506 |
+
"loss": 1.4005342721939087,
|
| 507 |
+
"step": 71
|
| 508 |
+
},
|
| 509 |
+
{
|
| 510 |
+
"epoch": 0.029538461538461538,
|
| 511 |
+
"grad_norm": 4.5,
|
| 512 |
+
"learning_rate": 4.953955629970699e-06,
|
| 513 |
+
"loss": 1.1521869897842407,
|
| 514 |
+
"step": 72
|
| 515 |
+
},
|
| 516 |
+
{
|
| 517 |
+
"epoch": 0.029948717948717948,
|
| 518 |
+
"grad_norm": 4.15625,
|
| 519 |
+
"learning_rate": 4.951862704060277e-06,
|
| 520 |
+
"loss": 1.1352651119232178,
|
| 521 |
+
"step": 73
|
| 522 |
+
},
|
| 523 |
+
{
|
| 524 |
+
"epoch": 0.030358974358974358,
|
| 525 |
+
"grad_norm": 5.03125,
|
| 526 |
+
"learning_rate": 4.9497697781498536e-06,
|
| 527 |
+
"loss": 1.3799540996551514,
|
| 528 |
+
"step": 74
|
| 529 |
+
},
|
| 530 |
+
{
|
| 531 |
+
"epoch": 0.03076923076923077,
|
| 532 |
+
"grad_norm": 4.28125,
|
| 533 |
+
"learning_rate": 4.947676852239431e-06,
|
| 534 |
+
"loss": 1.397932767868042,
|
| 535 |
+
"step": 75
|
| 536 |
+
},
|
| 537 |
+
{
|
| 538 |
+
"epoch": 0.03117948717948718,
|
| 539 |
+
"grad_norm": 5.3125,
|
| 540 |
+
"learning_rate": 4.945583926329009e-06,
|
| 541 |
+
"loss": 1.3446513414382935,
|
| 542 |
+
"step": 76
|
| 543 |
+
},
|
| 544 |
+
{
|
| 545 |
+
"epoch": 0.03158974358974359,
|
| 546 |
+
"grad_norm": 4.53125,
|
| 547 |
+
"learning_rate": 4.943491000418586e-06,
|
| 548 |
+
"loss": 1.3176562786102295,
|
| 549 |
+
"step": 77
|
| 550 |
+
},
|
| 551 |
+
{
|
| 552 |
+
"epoch": 0.032,
|
| 553 |
+
"grad_norm": 4.9375,
|
| 554 |
+
"learning_rate": 4.941398074508163e-06,
|
| 555 |
+
"loss": 1.2723119258880615,
|
| 556 |
+
"step": 78
|
| 557 |
+
},
|
| 558 |
+
{
|
| 559 |
+
"epoch": 0.03241025641025641,
|
| 560 |
+
"grad_norm": 3.96875,
|
| 561 |
+
"learning_rate": 4.93930514859774e-06,
|
| 562 |
+
"loss": 1.2040520906448364,
|
| 563 |
+
"step": 79
|
| 564 |
+
},
|
| 565 |
+
{
|
| 566 |
+
"epoch": 0.03282051282051282,
|
| 567 |
+
"grad_norm": 4.5,
|
| 568 |
+
"learning_rate": 4.937212222687317e-06,
|
| 569 |
+
"loss": 1.2306987047195435,
|
| 570 |
+
"step": 80
|
| 571 |
+
},
|
| 572 |
+
{
|
| 573 |
+
"epoch": 0.03323076923076923,
|
| 574 |
+
"grad_norm": 4.625,
|
| 575 |
+
"learning_rate": 4.9351192967768945e-06,
|
| 576 |
+
"loss": 1.29660964012146,
|
| 577 |
+
"step": 81
|
| 578 |
+
},
|
| 579 |
+
{
|
| 580 |
+
"epoch": 0.03364102564102564,
|
| 581 |
+
"grad_norm": 4.8125,
|
| 582 |
+
"learning_rate": 4.933026370866471e-06,
|
| 583 |
+
"loss": 1.5290207862854004,
|
| 584 |
+
"step": 82
|
| 585 |
+
},
|
| 586 |
+
{
|
| 587 |
+
"epoch": 0.03405128205128205,
|
| 588 |
+
"grad_norm": 4.4375,
|
| 589 |
+
"learning_rate": 4.930933444956049e-06,
|
| 590 |
+
"loss": 1.3405927419662476,
|
| 591 |
+
"step": 83
|
| 592 |
+
},
|
| 593 |
+
{
|
| 594 |
+
"epoch": 0.03446153846153846,
|
| 595 |
+
"grad_norm": 4.65625,
|
| 596 |
+
"learning_rate": 4.9288405190456265e-06,
|
| 597 |
+
"loss": 1.2912219762802124,
|
| 598 |
+
"step": 84
|
| 599 |
+
},
|
| 600 |
+
{
|
| 601 |
+
"epoch": 0.03487179487179487,
|
| 602 |
+
"grad_norm": 4.5625,
|
| 603 |
+
"learning_rate": 4.926747593135203e-06,
|
| 604 |
+
"loss": 1.4224706888198853,
|
| 605 |
+
"step": 85
|
| 606 |
+
},
|
| 607 |
+
{
|
| 608 |
+
"epoch": 0.03528205128205128,
|
| 609 |
+
"grad_norm": 5.28125,
|
| 610 |
+
"learning_rate": 4.924654667224781e-06,
|
| 611 |
+
"loss": 1.4317601919174194,
|
| 612 |
+
"step": 86
|
| 613 |
+
},
|
| 614 |
+
{
|
| 615 |
+
"epoch": 0.03569230769230769,
|
| 616 |
+
"grad_norm": 4.75,
|
| 617 |
+
"learning_rate": 4.922561741314358e-06,
|
| 618 |
+
"loss": 1.3233680725097656,
|
| 619 |
+
"step": 87
|
| 620 |
+
},
|
| 621 |
+
{
|
| 622 |
+
"epoch": 0.0361025641025641,
|
| 623 |
+
"grad_norm": 4.46875,
|
| 624 |
+
"learning_rate": 4.9204688154039346e-06,
|
| 625 |
+
"loss": 1.3949699401855469,
|
| 626 |
+
"step": 88
|
| 627 |
+
},
|
| 628 |
+
{
|
| 629 |
+
"epoch": 0.03651282051282051,
|
| 630 |
+
"grad_norm": 4.4375,
|
| 631 |
+
"learning_rate": 4.918375889493512e-06,
|
| 632 |
+
"loss": 1.201745867729187,
|
| 633 |
+
"step": 89
|
| 634 |
+
},
|
| 635 |
+
{
|
| 636 |
+
"epoch": 0.036923076923076927,
|
| 637 |
+
"grad_norm": 4.40625,
|
| 638 |
+
"learning_rate": 4.91628296358309e-06,
|
| 639 |
+
"loss": 1.3432514667510986,
|
| 640 |
+
"step": 90
|
| 641 |
+
},
|
| 642 |
+
{
|
| 643 |
+
"epoch": 0.037333333333333336,
|
| 644 |
+
"grad_norm": 3.75,
|
| 645 |
+
"learning_rate": 4.914190037672667e-06,
|
| 646 |
+
"loss": 1.1106362342834473,
|
| 647 |
+
"step": 91
|
| 648 |
+
},
|
| 649 |
+
{
|
| 650 |
+
"epoch": 0.037743589743589746,
|
| 651 |
+
"grad_norm": 3.96875,
|
| 652 |
+
"learning_rate": 4.912097111762244e-06,
|
| 653 |
+
"loss": 1.2580790519714355,
|
| 654 |
+
"step": 92
|
| 655 |
+
},
|
| 656 |
+
{
|
| 657 |
+
"epoch": 0.038153846153846156,
|
| 658 |
+
"grad_norm": 4.09375,
|
| 659 |
+
"learning_rate": 4.910004185851821e-06,
|
| 660 |
+
"loss": 1.2685620784759521,
|
| 661 |
+
"step": 93
|
| 662 |
+
},
|
| 663 |
+
{
|
| 664 |
+
"epoch": 0.038564102564102566,
|
| 665 |
+
"grad_norm": 4.28125,
|
| 666 |
+
"learning_rate": 4.907911259941399e-06,
|
| 667 |
+
"loss": 1.2353464365005493,
|
| 668 |
+
"step": 94
|
| 669 |
+
},
|
| 670 |
+
{
|
| 671 |
+
"epoch": 0.038974358974358976,
|
| 672 |
+
"grad_norm": 3.859375,
|
| 673 |
+
"learning_rate": 4.9058183340309755e-06,
|
| 674 |
+
"loss": 1.126069188117981,
|
| 675 |
+
"step": 95
|
| 676 |
+
},
|
| 677 |
+
{
|
| 678 |
+
"epoch": 0.039384615384615386,
|
| 679 |
+
"grad_norm": 3.84375,
|
| 680 |
+
"learning_rate": 4.903725408120553e-06,
|
| 681 |
+
"loss": 1.1509438753128052,
|
| 682 |
+
"step": 96
|
| 683 |
+
},
|
| 684 |
+
{
|
| 685 |
+
"epoch": 0.039794871794871796,
|
| 686 |
+
"grad_norm": 4.09375,
|
| 687 |
+
"learning_rate": 4.90163248221013e-06,
|
| 688 |
+
"loss": 1.1883344650268555,
|
| 689 |
+
"step": 97
|
| 690 |
+
},
|
| 691 |
+
{
|
| 692 |
+
"epoch": 0.040205128205128206,
|
| 693 |
+
"grad_norm": 4.46875,
|
| 694 |
+
"learning_rate": 4.8995395562997075e-06,
|
| 695 |
+
"loss": 1.154602289199829,
|
| 696 |
+
"step": 98
|
| 697 |
+
},
|
| 698 |
+
{
|
| 699 |
+
"epoch": 0.040615384615384616,
|
| 700 |
+
"grad_norm": 4.8125,
|
| 701 |
+
"learning_rate": 4.897446630389285e-06,
|
| 702 |
+
"loss": 1.436761498451233,
|
| 703 |
+
"step": 99
|
| 704 |
+
},
|
| 705 |
+
{
|
| 706 |
+
"epoch": 0.041025641025641026,
|
| 707 |
+
"grad_norm": 4.40625,
|
| 708 |
+
"learning_rate": 4.895353704478862e-06,
|
| 709 |
+
"loss": 1.3350646495819092,
|
| 710 |
+
"step": 100
|
| 711 |
+
},
|
| 712 |
+
{
|
| 713 |
+
"epoch": 0.041435897435897436,
|
| 714 |
+
"grad_norm": 4.3125,
|
| 715 |
+
"learning_rate": 4.893260778568439e-06,
|
| 716 |
+
"loss": 1.2260431051254272,
|
| 717 |
+
"step": 101
|
| 718 |
+
},
|
| 719 |
+
{
|
| 720 |
+
"epoch": 0.041846153846153845,
|
| 721 |
+
"grad_norm": 4.65625,
|
| 722 |
+
"learning_rate": 4.891167852658016e-06,
|
| 723 |
+
"loss": 1.1487889289855957,
|
| 724 |
+
"step": 102
|
| 725 |
+
},
|
| 726 |
+
{
|
| 727 |
+
"epoch": 0.042256410256410255,
|
| 728 |
+
"grad_norm": 4.4375,
|
| 729 |
+
"learning_rate": 4.889074926747593e-06,
|
| 730 |
+
"loss": 1.3355108499526978,
|
| 731 |
+
"step": 103
|
| 732 |
+
},
|
| 733 |
+
{
|
| 734 |
+
"epoch": 0.042666666666666665,
|
| 735 |
+
"grad_norm": 4.46875,
|
| 736 |
+
"learning_rate": 4.886982000837171e-06,
|
| 737 |
+
"loss": 1.157799482345581,
|
| 738 |
+
"step": 104
|
| 739 |
+
},
|
| 740 |
+
{
|
| 741 |
+
"epoch": 0.043076923076923075,
|
| 742 |
+
"grad_norm": 4.5,
|
| 743 |
+
"learning_rate": 4.884889074926748e-06,
|
| 744 |
+
"loss": 1.2815593481063843,
|
| 745 |
+
"step": 105
|
| 746 |
+
},
|
| 747 |
+
{
|
| 748 |
+
"epoch": 0.043487179487179485,
|
| 749 |
+
"grad_norm": 4.46875,
|
| 750 |
+
"learning_rate": 4.882796149016325e-06,
|
| 751 |
+
"loss": 1.3688324689865112,
|
| 752 |
+
"step": 106
|
| 753 |
+
},
|
| 754 |
+
{
|
| 755 |
+
"epoch": 0.043897435897435895,
|
| 756 |
+
"grad_norm": 4.46875,
|
| 757 |
+
"learning_rate": 4.880703223105903e-06,
|
| 758 |
+
"loss": 1.4537888765335083,
|
| 759 |
+
"step": 107
|
| 760 |
+
},
|
| 761 |
+
{
|
| 762 |
+
"epoch": 0.044307692307692305,
|
| 763 |
+
"grad_norm": 4.1875,
|
| 764 |
+
"learning_rate": 4.87861029719548e-06,
|
| 765 |
+
"loss": 1.2465651035308838,
|
| 766 |
+
"step": 108
|
| 767 |
+
},
|
| 768 |
+
{
|
| 769 |
+
"epoch": 0.044717948717948715,
|
| 770 |
+
"grad_norm": 4.15625,
|
| 771 |
+
"learning_rate": 4.8765173712850565e-06,
|
| 772 |
+
"loss": 1.221631646156311,
|
| 773 |
+
"step": 109
|
| 774 |
+
},
|
| 775 |
+
{
|
| 776 |
+
"epoch": 0.04512820512820513,
|
| 777 |
+
"grad_norm": 4.4375,
|
| 778 |
+
"learning_rate": 4.874424445374634e-06,
|
| 779 |
+
"loss": 1.2601536512374878,
|
| 780 |
+
"step": 110
|
| 781 |
+
},
|
| 782 |
+
{
|
| 783 |
+
"epoch": 0.04553846153846154,
|
| 784 |
+
"grad_norm": 4.5625,
|
| 785 |
+
"learning_rate": 4.872331519464211e-06,
|
| 786 |
+
"loss": 1.3444571495056152,
|
| 787 |
+
"step": 111
|
| 788 |
+
},
|
| 789 |
+
{
|
| 790 |
+
"epoch": 0.04594871794871795,
|
| 791 |
+
"grad_norm": 4.40625,
|
| 792 |
+
"learning_rate": 4.8702385935537886e-06,
|
| 793 |
+
"loss": 1.2017529010772705,
|
| 794 |
+
"step": 112
|
| 795 |
+
},
|
| 796 |
+
{
|
| 797 |
+
"epoch": 0.04635897435897436,
|
| 798 |
+
"grad_norm": 4.34375,
|
| 799 |
+
"learning_rate": 4.868145667643366e-06,
|
| 800 |
+
"loss": 1.2831099033355713,
|
| 801 |
+
"step": 113
|
| 802 |
+
},
|
| 803 |
+
{
|
| 804 |
+
"epoch": 0.04676923076923077,
|
| 805 |
+
"grad_norm": 4.625,
|
| 806 |
+
"learning_rate": 4.866052741732943e-06,
|
| 807 |
+
"loss": 1.2186903953552246,
|
| 808 |
+
"step": 114
|
| 809 |
+
},
|
| 810 |
+
{
|
| 811 |
+
"epoch": 0.04717948717948718,
|
| 812 |
+
"grad_norm": 4.15625,
|
| 813 |
+
"learning_rate": 4.863959815822521e-06,
|
| 814 |
+
"loss": 1.2090831995010376,
|
| 815 |
+
"step": 115
|
| 816 |
+
},
|
| 817 |
+
{
|
| 818 |
+
"epoch": 0.04758974358974359,
|
| 819 |
+
"grad_norm": 4.21875,
|
| 820 |
+
"learning_rate": 4.861866889912097e-06,
|
| 821 |
+
"loss": 1.3479111194610596,
|
| 822 |
+
"step": 116
|
| 823 |
+
},
|
| 824 |
+
{
|
| 825 |
+
"epoch": 0.048,
|
| 826 |
+
"grad_norm": 4.21875,
|
| 827 |
+
"learning_rate": 4.859773964001675e-06,
|
| 828 |
+
"loss": 1.1992478370666504,
|
| 829 |
+
"step": 117
|
| 830 |
+
},
|
| 831 |
+
{
|
| 832 |
+
"epoch": 0.04841025641025641,
|
| 833 |
+
"grad_norm": 4.5625,
|
| 834 |
+
"learning_rate": 4.857681038091252e-06,
|
| 835 |
+
"loss": 1.219160795211792,
|
| 836 |
+
"step": 118
|
| 837 |
+
},
|
| 838 |
+
{
|
| 839 |
+
"epoch": 0.04882051282051282,
|
| 840 |
+
"grad_norm": 4.15625,
|
| 841 |
+
"learning_rate": 4.855588112180829e-06,
|
| 842 |
+
"loss": 1.313496708869934,
|
| 843 |
+
"step": 119
|
| 844 |
+
},
|
| 845 |
+
{
|
| 846 |
+
"epoch": 0.04923076923076923,
|
| 847 |
+
"grad_norm": 4.75,
|
| 848 |
+
"learning_rate": 4.853495186270406e-06,
|
| 849 |
+
"loss": 1.3205606937408447,
|
| 850 |
+
"step": 120
|
| 851 |
+
},
|
| 852 |
+
{
|
| 853 |
+
"epoch": 0.04964102564102564,
|
| 854 |
+
"grad_norm": 4.6875,
|
| 855 |
+
"learning_rate": 4.851402260359984e-06,
|
| 856 |
+
"loss": 1.2018580436706543,
|
| 857 |
+
"step": 121
|
| 858 |
+
},
|
| 859 |
+
{
|
| 860 |
+
"epoch": 0.05005128205128205,
|
| 861 |
+
"grad_norm": 4.5625,
|
| 862 |
+
"learning_rate": 4.849309334449561e-06,
|
| 863 |
+
"loss": 1.3555666208267212,
|
| 864 |
+
"step": 122
|
| 865 |
+
},
|
| 866 |
+
{
|
| 867 |
+
"epoch": 0.05046153846153846,
|
| 868 |
+
"grad_norm": 4.375,
|
| 869 |
+
"learning_rate": 4.847216408539138e-06,
|
| 870 |
+
"loss": 1.3178867101669312,
|
| 871 |
+
"step": 123
|
| 872 |
+
},
|
| 873 |
+
{
|
| 874 |
+
"epoch": 0.05087179487179487,
|
| 875 |
+
"grad_norm": 4.21875,
|
| 876 |
+
"learning_rate": 4.845123482628715e-06,
|
| 877 |
+
"loss": 1.3092674016952515,
|
| 878 |
+
"step": 124
|
| 879 |
+
},
|
| 880 |
+
{
|
| 881 |
+
"epoch": 0.05128205128205128,
|
| 882 |
+
"grad_norm": 4.15625,
|
| 883 |
+
"learning_rate": 4.843030556718293e-06,
|
| 884 |
+
"loss": 1.3327960968017578,
|
| 885 |
+
"step": 125
|
| 886 |
+
},
|
| 887 |
+
{
|
| 888 |
+
"epoch": 0.05169230769230769,
|
| 889 |
+
"grad_norm": 4.28125,
|
| 890 |
+
"learning_rate": 4.8409376308078696e-06,
|
| 891 |
+
"loss": 1.2504072189331055,
|
| 892 |
+
"step": 126
|
| 893 |
+
},
|
| 894 |
+
{
|
| 895 |
+
"epoch": 0.0521025641025641,
|
| 896 |
+
"grad_norm": 4.5,
|
| 897 |
+
"learning_rate": 4.838844704897447e-06,
|
| 898 |
+
"loss": 1.1575727462768555,
|
| 899 |
+
"step": 127
|
| 900 |
+
},
|
| 901 |
+
{
|
| 902 |
+
"epoch": 0.05251282051282051,
|
| 903 |
+
"grad_norm": 4.3125,
|
| 904 |
+
"learning_rate": 4.836751778987025e-06,
|
| 905 |
+
"loss": 1.3626970052719116,
|
| 906 |
+
"step": 128
|
| 907 |
+
},
|
| 908 |
+
{
|
| 909 |
+
"epoch": 0.05292307692307692,
|
| 910 |
+
"grad_norm": 4.4375,
|
| 911 |
+
"learning_rate": 4.834658853076602e-06,
|
| 912 |
+
"loss": 1.1424509286880493,
|
| 913 |
+
"step": 129
|
| 914 |
+
},
|
| 915 |
+
{
|
| 916 |
+
"epoch": 0.05333333333333334,
|
| 917 |
+
"grad_norm": 4.53125,
|
| 918 |
+
"learning_rate": 4.832565927166178e-06,
|
| 919 |
+
"loss": 1.3161128759384155,
|
| 920 |
+
"step": 130
|
| 921 |
+
},
|
| 922 |
+
{
|
| 923 |
+
"epoch": 0.05374358974358975,
|
| 924 |
+
"grad_norm": 4.21875,
|
| 925 |
+
"learning_rate": 4.830473001255756e-06,
|
| 926 |
+
"loss": 1.1955534219741821,
|
| 927 |
+
"step": 131
|
| 928 |
+
},
|
| 929 |
+
{
|
| 930 |
+
"epoch": 0.05415384615384616,
|
| 931 |
+
"grad_norm": 4.28125,
|
| 932 |
+
"learning_rate": 4.828380075345333e-06,
|
| 933 |
+
"loss": 1.2027688026428223,
|
| 934 |
+
"step": 132
|
| 935 |
+
},
|
| 936 |
+
{
|
| 937 |
+
"epoch": 0.054564102564102566,
|
| 938 |
+
"grad_norm": 4.1875,
|
| 939 |
+
"learning_rate": 4.8262871494349105e-06,
|
| 940 |
+
"loss": 1.255806565284729,
|
| 941 |
+
"step": 133
|
| 942 |
+
},
|
| 943 |
+
{
|
| 944 |
+
"epoch": 0.054974358974358976,
|
| 945 |
+
"grad_norm": 4.65625,
|
| 946 |
+
"learning_rate": 4.824194223524487e-06,
|
| 947 |
+
"loss": 1.2906789779663086,
|
| 948 |
+
"step": 134
|
| 949 |
+
},
|
| 950 |
+
{
|
| 951 |
+
"epoch": 0.055384615384615386,
|
| 952 |
+
"grad_norm": 4.0625,
|
| 953 |
+
"learning_rate": 4.822101297614065e-06,
|
| 954 |
+
"loss": 1.2715399265289307,
|
| 955 |
+
"step": 135
|
| 956 |
+
},
|
| 957 |
+
{
|
| 958 |
+
"epoch": 0.055794871794871796,
|
| 959 |
+
"grad_norm": 4.34375,
|
| 960 |
+
"learning_rate": 4.8200083717036425e-06,
|
| 961 |
+
"loss": 1.270885705947876,
|
| 962 |
+
"step": 136
|
| 963 |
+
},
|
| 964 |
+
{
|
| 965 |
+
"epoch": 0.056205128205128206,
|
| 966 |
+
"grad_norm": 4.28125,
|
| 967 |
+
"learning_rate": 4.817915445793219e-06,
|
| 968 |
+
"loss": 1.2684006690979004,
|
| 969 |
+
"step": 137
|
| 970 |
+
},
|
| 971 |
+
{
|
| 972 |
+
"epoch": 0.056615384615384616,
|
| 973 |
+
"grad_norm": 4.0625,
|
| 974 |
+
"learning_rate": 4.815822519882797e-06,
|
| 975 |
+
"loss": 1.2454103231430054,
|
| 976 |
+
"step": 138
|
| 977 |
+
},
|
| 978 |
+
{
|
| 979 |
+
"epoch": 0.057025641025641026,
|
| 980 |
+
"grad_norm": 3.75,
|
| 981 |
+
"learning_rate": 4.813729593972374e-06,
|
| 982 |
+
"loss": 1.2314739227294922,
|
| 983 |
+
"step": 139
|
| 984 |
+
},
|
| 985 |
+
{
|
| 986 |
+
"epoch": 0.057435897435897436,
|
| 987 |
+
"grad_norm": 3.65625,
|
| 988 |
+
"learning_rate": 4.8116366680619506e-06,
|
| 989 |
+
"loss": 1.2362333536148071,
|
| 990 |
+
"step": 140
|
| 991 |
+
},
|
| 992 |
+
{
|
| 993 |
+
"epoch": 0.057846153846153846,
|
| 994 |
+
"grad_norm": 3.875,
|
| 995 |
+
"learning_rate": 4.809543742151528e-06,
|
| 996 |
+
"loss": 1.1714919805526733,
|
| 997 |
+
"step": 141
|
| 998 |
+
},
|
| 999 |
+
{
|
| 1000 |
+
"epoch": 0.058256410256410256,
|
| 1001 |
+
"grad_norm": 4.78125,
|
| 1002 |
+
"learning_rate": 4.807450816241106e-06,
|
| 1003 |
+
"loss": 1.340531826019287,
|
| 1004 |
+
"step": 142
|
| 1005 |
+
},
|
| 1006 |
+
{
|
| 1007 |
+
"epoch": 0.058666666666666666,
|
| 1008 |
+
"grad_norm": 4.03125,
|
| 1009 |
+
"learning_rate": 4.805357890330683e-06,
|
| 1010 |
+
"loss": 1.2856035232543945,
|
| 1011 |
+
"step": 143
|
| 1012 |
+
},
|
| 1013 |
+
{
|
| 1014 |
+
"epoch": 0.059076923076923075,
|
| 1015 |
+
"grad_norm": 4.09375,
|
| 1016 |
+
"learning_rate": 4.80326496442026e-06,
|
| 1017 |
+
"loss": 1.183943748474121,
|
| 1018 |
+
"step": 144
|
| 1019 |
+
},
|
| 1020 |
+
{
|
| 1021 |
+
"epoch": 0.059487179487179485,
|
| 1022 |
+
"grad_norm": 4.59375,
|
| 1023 |
+
"learning_rate": 4.801172038509837e-06,
|
| 1024 |
+
"loss": 1.311848521232605,
|
| 1025 |
+
"step": 145
|
| 1026 |
+
},
|
| 1027 |
+
{
|
| 1028 |
+
"epoch": 0.059897435897435895,
|
| 1029 |
+
"grad_norm": 4.28125,
|
| 1030 |
+
"learning_rate": 4.799079112599415e-06,
|
| 1031 |
+
"loss": 1.1645233631134033,
|
| 1032 |
+
"step": 146
|
| 1033 |
+
},
|
| 1034 |
+
{
|
| 1035 |
+
"epoch": 0.060307692307692305,
|
| 1036 |
+
"grad_norm": 3.953125,
|
| 1037 |
+
"learning_rate": 4.7969861866889915e-06,
|
| 1038 |
+
"loss": 1.2566933631896973,
|
| 1039 |
+
"step": 147
|
| 1040 |
+
},
|
| 1041 |
+
{
|
| 1042 |
+
"epoch": 0.060717948717948715,
|
| 1043 |
+
"grad_norm": 4.21875,
|
| 1044 |
+
"learning_rate": 4.794893260778568e-06,
|
| 1045 |
+
"loss": 1.2860949039459229,
|
| 1046 |
+
"step": 148
|
| 1047 |
+
},
|
| 1048 |
+
{
|
| 1049 |
+
"epoch": 0.061128205128205125,
|
| 1050 |
+
"grad_norm": 4.03125,
|
| 1051 |
+
"learning_rate": 4.792800334868146e-06,
|
| 1052 |
+
"loss": 1.2152372598648071,
|
| 1053 |
+
"step": 149
|
| 1054 |
+
},
|
| 1055 |
+
{
|
| 1056 |
+
"epoch": 0.06153846153846154,
|
| 1057 |
+
"grad_norm": 4.40625,
|
| 1058 |
+
"learning_rate": 4.7907074089577235e-06,
|
| 1059 |
+
"loss": 1.2442578077316284,
|
| 1060 |
+
"step": 150
|
| 1061 |
+
},
|
| 1062 |
+
{
|
| 1063 |
+
"epoch": 0.06194871794871795,
|
| 1064 |
+
"grad_norm": 4.625,
|
| 1065 |
+
"learning_rate": 4.7886144830473e-06,
|
| 1066 |
+
"loss": 1.3592571020126343,
|
| 1067 |
+
"step": 151
|
| 1068 |
+
},
|
| 1069 |
+
{
|
| 1070 |
+
"epoch": 0.06235897435897436,
|
| 1071 |
+
"grad_norm": 3.8125,
|
| 1072 |
+
"learning_rate": 4.786521557136878e-06,
|
| 1073 |
+
"loss": 1.2471686601638794,
|
| 1074 |
+
"step": 152
|
| 1075 |
+
},
|
| 1076 |
+
{
|
| 1077 |
+
"epoch": 0.06276923076923077,
|
| 1078 |
+
"grad_norm": 3.9375,
|
| 1079 |
+
"learning_rate": 4.784428631226455e-06,
|
| 1080 |
+
"loss": 1.414598822593689,
|
| 1081 |
+
"step": 153
|
| 1082 |
+
},
|
| 1083 |
+
{
|
| 1084 |
+
"epoch": 0.06317948717948718,
|
| 1085 |
+
"grad_norm": 4.0625,
|
| 1086 |
+
"learning_rate": 4.782335705316032e-06,
|
| 1087 |
+
"loss": 1.2605376243591309,
|
| 1088 |
+
"step": 154
|
| 1089 |
+
},
|
| 1090 |
+
{
|
| 1091 |
+
"epoch": 0.06358974358974359,
|
| 1092 |
+
"grad_norm": 4.5,
|
| 1093 |
+
"learning_rate": 4.780242779405609e-06,
|
| 1094 |
+
"loss": 1.2983124256134033,
|
| 1095 |
+
"step": 155
|
| 1096 |
+
},
|
| 1097 |
+
{
|
| 1098 |
+
"epoch": 0.064,
|
| 1099 |
+
"grad_norm": 4.28125,
|
| 1100 |
+
"learning_rate": 4.778149853495187e-06,
|
| 1101 |
+
"loss": 1.320411205291748,
|
| 1102 |
+
"step": 156
|
| 1103 |
+
},
|
| 1104 |
+
{
|
| 1105 |
+
"epoch": 0.06441025641025641,
|
| 1106 |
+
"grad_norm": 4.3125,
|
| 1107 |
+
"learning_rate": 4.776056927584764e-06,
|
| 1108 |
+
"loss": 1.3486765623092651,
|
| 1109 |
+
"step": 157
|
| 1110 |
+
},
|
| 1111 |
+
{
|
| 1112 |
+
"epoch": 0.06482051282051282,
|
| 1113 |
+
"grad_norm": 3.75,
|
| 1114 |
+
"learning_rate": 4.773964001674341e-06,
|
| 1115 |
+
"loss": 1.2210478782653809,
|
| 1116 |
+
"step": 158
|
| 1117 |
+
},
|
| 1118 |
+
{
|
| 1119 |
+
"epoch": 0.06523076923076923,
|
| 1120 |
+
"grad_norm": 4.09375,
|
| 1121 |
+
"learning_rate": 4.771871075763919e-06,
|
| 1122 |
+
"loss": 1.1709284782409668,
|
| 1123 |
+
"step": 159
|
| 1124 |
+
},
|
| 1125 |
+
{
|
| 1126 |
+
"epoch": 0.06564102564102564,
|
| 1127 |
+
"grad_norm": 4.09375,
|
| 1128 |
+
"learning_rate": 4.769778149853496e-06,
|
| 1129 |
+
"loss": 1.345170497894287,
|
| 1130 |
+
"step": 160
|
| 1131 |
+
},
|
| 1132 |
+
{
|
| 1133 |
+
"epoch": 0.06605128205128205,
|
| 1134 |
+
"grad_norm": 3.828125,
|
| 1135 |
+
"learning_rate": 4.7676852239430725e-06,
|
| 1136 |
+
"loss": 1.1507529020309448,
|
| 1137 |
+
"step": 161
|
| 1138 |
+
},
|
| 1139 |
+
{
|
| 1140 |
+
"epoch": 0.06646153846153846,
|
| 1141 |
+
"grad_norm": 3.921875,
|
| 1142 |
+
"learning_rate": 4.76559229803265e-06,
|
| 1143 |
+
"loss": 1.3053486347198486,
|
| 1144 |
+
"step": 162
|
| 1145 |
+
},
|
| 1146 |
+
{
|
| 1147 |
+
"epoch": 0.06687179487179487,
|
| 1148 |
+
"grad_norm": 4.125,
|
| 1149 |
+
"learning_rate": 4.763499372122227e-06,
|
| 1150 |
+
"loss": 1.163478136062622,
|
| 1151 |
+
"step": 163
|
| 1152 |
+
},
|
| 1153 |
+
{
|
| 1154 |
+
"epoch": 0.06728205128205128,
|
| 1155 |
+
"grad_norm": 4.65625,
|
| 1156 |
+
"learning_rate": 4.7614064462118045e-06,
|
| 1157 |
+
"loss": 1.3403921127319336,
|
| 1158 |
+
"step": 164
|
| 1159 |
+
},
|
| 1160 |
+
{
|
| 1161 |
+
"epoch": 0.06769230769230769,
|
| 1162 |
+
"grad_norm": 4.0625,
|
| 1163 |
+
"learning_rate": 4.759313520301382e-06,
|
| 1164 |
+
"loss": 1.2189948558807373,
|
| 1165 |
+
"step": 165
|
| 1166 |
+
},
|
| 1167 |
+
{
|
| 1168 |
+
"epoch": 0.0681025641025641,
|
| 1169 |
+
"grad_norm": 3.90625,
|
| 1170 |
+
"learning_rate": 4.757220594390959e-06,
|
| 1171 |
+
"loss": 1.2726385593414307,
|
| 1172 |
+
"step": 166
|
| 1173 |
+
},
|
| 1174 |
+
{
|
| 1175 |
+
"epoch": 0.06851282051282051,
|
| 1176 |
+
"grad_norm": 4.125,
|
| 1177 |
+
"learning_rate": 4.755127668480537e-06,
|
| 1178 |
+
"loss": 1.2154686450958252,
|
| 1179 |
+
"step": 167
|
| 1180 |
+
},
|
| 1181 |
+
{
|
| 1182 |
+
"epoch": 0.06892307692307692,
|
| 1183 |
+
"grad_norm": 4.03125,
|
| 1184 |
+
"learning_rate": 4.753034742570113e-06,
|
| 1185 |
+
"loss": 1.1780462265014648,
|
| 1186 |
+
"step": 168
|
| 1187 |
+
},
|
| 1188 |
+
{
|
| 1189 |
+
"epoch": 0.06933333333333333,
|
| 1190 |
+
"grad_norm": 4.375,
|
| 1191 |
+
"learning_rate": 4.75094181665969e-06,
|
| 1192 |
+
"loss": 1.234800934791565,
|
| 1193 |
+
"step": 169
|
| 1194 |
+
},
|
| 1195 |
+
{
|
| 1196 |
+
"epoch": 0.06974358974358974,
|
| 1197 |
+
"grad_norm": 4.3125,
|
| 1198 |
+
"learning_rate": 4.748848890749268e-06,
|
| 1199 |
+
"loss": 1.3003370761871338,
|
| 1200 |
+
"step": 170
|
| 1201 |
+
},
|
| 1202 |
+
{
|
| 1203 |
+
"epoch": 0.07015384615384615,
|
| 1204 |
+
"grad_norm": 4.1875,
|
| 1205 |
+
"learning_rate": 4.746755964838845e-06,
|
| 1206 |
+
"loss": 1.372632384300232,
|
| 1207 |
+
"step": 171
|
| 1208 |
+
},
|
| 1209 |
+
{
|
| 1210 |
+
"epoch": 0.07056410256410256,
|
| 1211 |
+
"grad_norm": 4.0,
|
| 1212 |
+
"learning_rate": 4.744663038928422e-06,
|
| 1213 |
+
"loss": 1.170872449874878,
|
| 1214 |
+
"step": 172
|
| 1215 |
+
},
|
| 1216 |
+
{
|
| 1217 |
+
"epoch": 0.07097435897435897,
|
| 1218 |
+
"grad_norm": 3.859375,
|
| 1219 |
+
"learning_rate": 4.742570113018e-06,
|
| 1220 |
+
"loss": 1.100975751876831,
|
| 1221 |
+
"step": 173
|
| 1222 |
+
},
|
| 1223 |
+
{
|
| 1224 |
+
"epoch": 0.07138461538461538,
|
| 1225 |
+
"grad_norm": 4.125,
|
| 1226 |
+
"learning_rate": 4.740477187107577e-06,
|
| 1227 |
+
"loss": 1.2144557237625122,
|
| 1228 |
+
"step": 174
|
| 1229 |
+
},
|
| 1230 |
+
{
|
| 1231 |
+
"epoch": 0.07179487179487179,
|
| 1232 |
+
"grad_norm": 3.90625,
|
| 1233 |
+
"learning_rate": 4.738384261197154e-06,
|
| 1234 |
+
"loss": 1.2314698696136475,
|
| 1235 |
+
"step": 175
|
| 1236 |
+
},
|
| 1237 |
+
{
|
| 1238 |
+
"epoch": 0.0722051282051282,
|
| 1239 |
+
"grad_norm": 4.3125,
|
| 1240 |
+
"learning_rate": 4.736291335286731e-06,
|
| 1241 |
+
"loss": 1.1676030158996582,
|
| 1242 |
+
"step": 176
|
| 1243 |
+
},
|
| 1244 |
+
{
|
| 1245 |
+
"epoch": 0.07261538461538461,
|
| 1246 |
+
"grad_norm": 4.0625,
|
| 1247 |
+
"learning_rate": 4.734198409376308e-06,
|
| 1248 |
+
"loss": 1.382598638534546,
|
| 1249 |
+
"step": 177
|
| 1250 |
+
},
|
| 1251 |
+
{
|
| 1252 |
+
"epoch": 0.07302564102564102,
|
| 1253 |
+
"grad_norm": 4.53125,
|
| 1254 |
+
"learning_rate": 4.7321054834658856e-06,
|
| 1255 |
+
"loss": 1.3178304433822632,
|
| 1256 |
+
"step": 178
|
| 1257 |
+
},
|
| 1258 |
+
{
|
| 1259 |
+
"epoch": 0.07343589743589743,
|
| 1260 |
+
"grad_norm": 4.3125,
|
| 1261 |
+
"learning_rate": 4.730012557555463e-06,
|
| 1262 |
+
"loss": 1.3160381317138672,
|
| 1263 |
+
"step": 179
|
| 1264 |
+
},
|
| 1265 |
+
{
|
| 1266 |
+
"epoch": 0.07384615384615385,
|
| 1267 |
+
"grad_norm": 3.890625,
|
| 1268 |
+
"learning_rate": 4.727919631645041e-06,
|
| 1269 |
+
"loss": 1.213687539100647,
|
| 1270 |
+
"step": 180
|
| 1271 |
+
},
|
| 1272 |
+
{
|
| 1273 |
+
"epoch": 0.07425641025641026,
|
| 1274 |
+
"grad_norm": 4.21875,
|
| 1275 |
+
"learning_rate": 4.725826705734618e-06,
|
| 1276 |
+
"loss": 1.275590419769287,
|
| 1277 |
+
"step": 181
|
| 1278 |
+
},
|
| 1279 |
+
{
|
| 1280 |
+
"epoch": 0.07466666666666667,
|
| 1281 |
+
"grad_norm": 4.0625,
|
| 1282 |
+
"learning_rate": 4.723733779824194e-06,
|
| 1283 |
+
"loss": 1.032568335533142,
|
| 1284 |
+
"step": 182
|
| 1285 |
+
},
|
| 1286 |
+
{
|
| 1287 |
+
"epoch": 0.07507692307692308,
|
| 1288 |
+
"grad_norm": 4.3125,
|
| 1289 |
+
"learning_rate": 4.721640853913772e-06,
|
| 1290 |
+
"loss": 1.2059623003005981,
|
| 1291 |
+
"step": 183
|
| 1292 |
+
},
|
| 1293 |
+
{
|
| 1294 |
+
"epoch": 0.07548717948717949,
|
| 1295 |
+
"grad_norm": 4.25,
|
| 1296 |
+
"learning_rate": 4.719547928003349e-06,
|
| 1297 |
+
"loss": 1.2014434337615967,
|
| 1298 |
+
"step": 184
|
| 1299 |
+
},
|
| 1300 |
+
{
|
| 1301 |
+
"epoch": 0.0758974358974359,
|
| 1302 |
+
"grad_norm": 4.15625,
|
| 1303 |
+
"learning_rate": 4.7174550020929265e-06,
|
| 1304 |
+
"loss": 1.3291563987731934,
|
| 1305 |
+
"step": 185
|
| 1306 |
+
},
|
| 1307 |
+
{
|
| 1308 |
+
"epoch": 0.07630769230769231,
|
| 1309 |
+
"grad_norm": 4.5625,
|
| 1310 |
+
"learning_rate": 4.715362076182503e-06,
|
| 1311 |
+
"loss": 1.2368932962417603,
|
| 1312 |
+
"step": 186
|
| 1313 |
+
},
|
| 1314 |
+
{
|
| 1315 |
+
"epoch": 0.07671794871794872,
|
| 1316 |
+
"grad_norm": 3.984375,
|
| 1317 |
+
"learning_rate": 4.713269150272081e-06,
|
| 1318 |
+
"loss": 1.2348570823669434,
|
| 1319 |
+
"step": 187
|
| 1320 |
+
},
|
| 1321 |
+
{
|
| 1322 |
+
"epoch": 0.07712820512820513,
|
| 1323 |
+
"grad_norm": 4.3125,
|
| 1324 |
+
"learning_rate": 4.7111762243616585e-06,
|
| 1325 |
+
"loss": 1.2440792322158813,
|
| 1326 |
+
"step": 188
|
| 1327 |
+
},
|
| 1328 |
+
{
|
| 1329 |
+
"epoch": 0.07753846153846154,
|
| 1330 |
+
"grad_norm": 4.03125,
|
| 1331 |
+
"learning_rate": 4.709083298451235e-06,
|
| 1332 |
+
"loss": 1.2497843503952026,
|
| 1333 |
+
"step": 189
|
| 1334 |
+
},
|
| 1335 |
+
{
|
| 1336 |
+
"epoch": 0.07794871794871795,
|
| 1337 |
+
"grad_norm": 4.28125,
|
| 1338 |
+
"learning_rate": 4.706990372540812e-06,
|
| 1339 |
+
"loss": 1.4095282554626465,
|
| 1340 |
+
"step": 190
|
| 1341 |
+
},
|
| 1342 |
+
{
|
| 1343 |
+
"epoch": 0.07835897435897436,
|
| 1344 |
+
"grad_norm": 4.4375,
|
| 1345 |
+
"learning_rate": 4.70489744663039e-06,
|
| 1346 |
+
"loss": 1.2543988227844238,
|
| 1347 |
+
"step": 191
|
| 1348 |
+
},
|
| 1349 |
+
{
|
| 1350 |
+
"epoch": 0.07876923076923077,
|
| 1351 |
+
"grad_norm": 4.5625,
|
| 1352 |
+
"learning_rate": 4.7028045207199666e-06,
|
| 1353 |
+
"loss": 1.2789925336837769,
|
| 1354 |
+
"step": 192
|
| 1355 |
+
},
|
| 1356 |
+
{
|
| 1357 |
+
"epoch": 0.07917948717948718,
|
| 1358 |
+
"grad_norm": 3.875,
|
| 1359 |
+
"learning_rate": 4.700711594809544e-06,
|
| 1360 |
+
"loss": 1.307772159576416,
|
| 1361 |
+
"step": 193
|
| 1362 |
+
},
|
| 1363 |
+
{
|
| 1364 |
+
"epoch": 0.07958974358974359,
|
| 1365 |
+
"grad_norm": 4.21875,
|
| 1366 |
+
"learning_rate": 4.698618668899122e-06,
|
| 1367 |
+
"loss": 1.3721433877944946,
|
| 1368 |
+
"step": 194
|
| 1369 |
+
},
|
| 1370 |
+
{
|
| 1371 |
+
"epoch": 0.08,
|
| 1372 |
+
"grad_norm": 4.25,
|
| 1373 |
+
"learning_rate": 4.696525742988699e-06,
|
| 1374 |
+
"loss": 1.2024502754211426,
|
| 1375 |
+
"step": 195
|
| 1376 |
+
},
|
| 1377 |
+
{
|
| 1378 |
+
"epoch": 0.08041025641025641,
|
| 1379 |
+
"grad_norm": 4.1875,
|
| 1380 |
+
"learning_rate": 4.694432817078276e-06,
|
| 1381 |
+
"loss": 1.2736141681671143,
|
| 1382 |
+
"step": 196
|
| 1383 |
+
},
|
| 1384 |
+
{
|
| 1385 |
+
"epoch": 0.08082051282051282,
|
| 1386 |
+
"grad_norm": 3.875,
|
| 1387 |
+
"learning_rate": 4.692339891167853e-06,
|
| 1388 |
+
"loss": 1.186525583267212,
|
| 1389 |
+
"step": 197
|
| 1390 |
+
},
|
| 1391 |
+
{
|
| 1392 |
+
"epoch": 0.08123076923076923,
|
| 1393 |
+
"grad_norm": 4.21875,
|
| 1394 |
+
"learning_rate": 4.69024696525743e-06,
|
| 1395 |
+
"loss": 1.2935118675231934,
|
| 1396 |
+
"step": 198
|
| 1397 |
+
},
|
| 1398 |
+
{
|
| 1399 |
+
"epoch": 0.08164102564102564,
|
| 1400 |
+
"grad_norm": 4.3125,
|
| 1401 |
+
"learning_rate": 4.6881540393470075e-06,
|
| 1402 |
+
"loss": 1.310086965560913,
|
| 1403 |
+
"step": 199
|
| 1404 |
+
},
|
| 1405 |
+
{
|
| 1406 |
+
"epoch": 0.08205128205128205,
|
| 1407 |
+
"grad_norm": 5.0625,
|
| 1408 |
+
"learning_rate": 4.686061113436584e-06,
|
| 1409 |
+
"loss": 1.28630793094635,
|
| 1410 |
+
"step": 200
|
| 1411 |
+
},
|
| 1412 |
+
{
|
| 1413 |
+
"epoch": 0.08246153846153846,
|
| 1414 |
+
"grad_norm": 4.59375,
|
| 1415 |
+
"learning_rate": 4.683968187526162e-06,
|
| 1416 |
+
"loss": 1.1203373670578003,
|
| 1417 |
+
"step": 201
|
| 1418 |
+
},
|
| 1419 |
+
{
|
| 1420 |
+
"epoch": 0.08287179487179487,
|
| 1421 |
+
"grad_norm": 4.25,
|
| 1422 |
+
"learning_rate": 4.6818752616157395e-06,
|
| 1423 |
+
"loss": 1.2714130878448486,
|
| 1424 |
+
"step": 202
|
| 1425 |
+
},
|
| 1426 |
+
{
|
| 1427 |
+
"epoch": 0.08328205128205128,
|
| 1428 |
+
"grad_norm": 4.0,
|
| 1429 |
+
"learning_rate": 4.679782335705316e-06,
|
| 1430 |
+
"loss": 1.228044033050537,
|
| 1431 |
+
"step": 203
|
| 1432 |
+
},
|
| 1433 |
+
{
|
| 1434 |
+
"epoch": 0.08369230769230769,
|
| 1435 |
+
"grad_norm": 3.640625,
|
| 1436 |
+
"learning_rate": 4.677689409794894e-06,
|
| 1437 |
+
"loss": 1.2472220659255981,
|
| 1438 |
+
"step": 204
|
| 1439 |
+
},
|
| 1440 |
+
{
|
| 1441 |
+
"epoch": 0.0841025641025641,
|
| 1442 |
+
"grad_norm": 5.21875,
|
| 1443 |
+
"learning_rate": 4.675596483884471e-06,
|
| 1444 |
+
"loss": 1.182159185409546,
|
| 1445 |
+
"step": 205
|
| 1446 |
+
},
|
| 1447 |
+
{
|
| 1448 |
+
"epoch": 0.08451282051282051,
|
| 1449 |
+
"grad_norm": 3.859375,
|
| 1450 |
+
"learning_rate": 4.6735035579740476e-06,
|
| 1451 |
+
"loss": 1.2498165369033813,
|
| 1452 |
+
"step": 206
|
| 1453 |
+
},
|
| 1454 |
+
{
|
| 1455 |
+
"epoch": 0.08492307692307692,
|
| 1456 |
+
"grad_norm": 3.640625,
|
| 1457 |
+
"learning_rate": 4.671410632063625e-06,
|
| 1458 |
+
"loss": 1.190643548965454,
|
| 1459 |
+
"step": 207
|
| 1460 |
+
},
|
| 1461 |
+
{
|
| 1462 |
+
"epoch": 0.08533333333333333,
|
| 1463 |
+
"grad_norm": 4.4375,
|
| 1464 |
+
"learning_rate": 4.669317706153203e-06,
|
| 1465 |
+
"loss": 1.233026385307312,
|
| 1466 |
+
"step": 208
|
| 1467 |
+
},
|
| 1468 |
+
{
|
| 1469 |
+
"epoch": 0.08574358974358974,
|
| 1470 |
+
"grad_norm": 4.28125,
|
| 1471 |
+
"learning_rate": 4.66722478024278e-06,
|
| 1472 |
+
"loss": 1.2331702709197998,
|
| 1473 |
+
"step": 209
|
| 1474 |
+
},
|
| 1475 |
+
{
|
| 1476 |
+
"epoch": 0.08615384615384615,
|
| 1477 |
+
"grad_norm": 4.15625,
|
| 1478 |
+
"learning_rate": 4.665131854332357e-06,
|
| 1479 |
+
"loss": 1.151205062866211,
|
| 1480 |
+
"step": 210
|
| 1481 |
+
},
|
| 1482 |
+
{
|
| 1483 |
+
"epoch": 0.08656410256410256,
|
| 1484 |
+
"grad_norm": 5.1875,
|
| 1485 |
+
"learning_rate": 4.663038928421934e-06,
|
| 1486 |
+
"loss": 1.2539771795272827,
|
| 1487 |
+
"step": 211
|
| 1488 |
+
},
|
| 1489 |
+
{
|
| 1490 |
+
"epoch": 0.08697435897435897,
|
| 1491 |
+
"grad_norm": 4.53125,
|
| 1492 |
+
"learning_rate": 4.660946002511512e-06,
|
| 1493 |
+
"loss": 1.5276305675506592,
|
| 1494 |
+
"step": 212
|
| 1495 |
+
},
|
| 1496 |
+
{
|
| 1497 |
+
"epoch": 0.08738461538461538,
|
| 1498 |
+
"grad_norm": 4.40625,
|
| 1499 |
+
"learning_rate": 4.6588530766010885e-06,
|
| 1500 |
+
"loss": 1.1912729740142822,
|
| 1501 |
+
"step": 213
|
| 1502 |
+
},
|
| 1503 |
+
{
|
| 1504 |
+
"epoch": 0.08779487179487179,
|
| 1505 |
+
"grad_norm": 4.40625,
|
| 1506 |
+
"learning_rate": 4.656760150690665e-06,
|
| 1507 |
+
"loss": 1.0854434967041016,
|
| 1508 |
+
"step": 214
|
| 1509 |
+
},
|
| 1510 |
+
{
|
| 1511 |
+
"epoch": 0.0882051282051282,
|
| 1512 |
+
"grad_norm": 4.5,
|
| 1513 |
+
"learning_rate": 4.654667224780243e-06,
|
| 1514 |
+
"loss": 1.2251269817352295,
|
| 1515 |
+
"step": 215
|
| 1516 |
+
},
|
| 1517 |
+
{
|
| 1518 |
+
"epoch": 0.08861538461538461,
|
| 1519 |
+
"grad_norm": 4.125,
|
| 1520 |
+
"learning_rate": 4.6525742988698205e-06,
|
| 1521 |
+
"loss": 1.2022786140441895,
|
| 1522 |
+
"step": 216
|
| 1523 |
+
},
|
| 1524 |
+
{
|
| 1525 |
+
"epoch": 0.08902564102564102,
|
| 1526 |
+
"grad_norm": 5.34375,
|
| 1527 |
+
"learning_rate": 4.650481372959398e-06,
|
| 1528 |
+
"loss": 1.2680422067642212,
|
| 1529 |
+
"step": 217
|
| 1530 |
+
},
|
| 1531 |
+
{
|
| 1532 |
+
"epoch": 0.08943589743589743,
|
| 1533 |
+
"grad_norm": 4.59375,
|
| 1534 |
+
"learning_rate": 4.648388447048975e-06,
|
| 1535 |
+
"loss": 1.2347216606140137,
|
| 1536 |
+
"step": 218
|
| 1537 |
+
},
|
| 1538 |
+
{
|
| 1539 |
+
"epoch": 0.08984615384615384,
|
| 1540 |
+
"grad_norm": 4.375,
|
| 1541 |
+
"learning_rate": 4.646295521138552e-06,
|
| 1542 |
+
"loss": 1.3959842920303345,
|
| 1543 |
+
"step": 219
|
| 1544 |
+
},
|
| 1545 |
+
{
|
| 1546 |
+
"epoch": 0.09025641025641026,
|
| 1547 |
+
"grad_norm": 4.0,
|
| 1548 |
+
"learning_rate": 4.644202595228129e-06,
|
| 1549 |
+
"loss": 1.2361531257629395,
|
| 1550 |
+
"step": 220
|
| 1551 |
+
},
|
| 1552 |
+
{
|
| 1553 |
+
"epoch": 0.09066666666666667,
|
| 1554 |
+
"grad_norm": 4.375,
|
| 1555 |
+
"learning_rate": 4.642109669317706e-06,
|
| 1556 |
+
"loss": 1.2429603338241577,
|
| 1557 |
+
"step": 221
|
| 1558 |
+
},
|
| 1559 |
+
{
|
| 1560 |
+
"epoch": 0.09107692307692308,
|
| 1561 |
+
"grad_norm": 4.25,
|
| 1562 |
+
"learning_rate": 4.640016743407284e-06,
|
| 1563 |
+
"loss": 1.401918649673462,
|
| 1564 |
+
"step": 222
|
| 1565 |
+
},
|
| 1566 |
+
{
|
| 1567 |
+
"epoch": 0.09148717948717949,
|
| 1568 |
+
"grad_norm": 3.796875,
|
| 1569 |
+
"learning_rate": 4.637923817496861e-06,
|
| 1570 |
+
"loss": 1.123022198677063,
|
| 1571 |
+
"step": 223
|
| 1572 |
+
},
|
| 1573 |
+
{
|
| 1574 |
+
"epoch": 0.0918974358974359,
|
| 1575 |
+
"grad_norm": 4.53125,
|
| 1576 |
+
"learning_rate": 4.635830891586438e-06,
|
| 1577 |
+
"loss": 1.4398316144943237,
|
| 1578 |
+
"step": 224
|
| 1579 |
+
},
|
| 1580 |
+
{
|
| 1581 |
+
"epoch": 0.09230769230769231,
|
| 1582 |
+
"grad_norm": 4.34375,
|
| 1583 |
+
"learning_rate": 4.633737965676016e-06,
|
| 1584 |
+
"loss": 1.2587261199951172,
|
| 1585 |
+
"step": 225
|
| 1586 |
+
},
|
| 1587 |
+
{
|
| 1588 |
+
"epoch": 0.09271794871794872,
|
| 1589 |
+
"grad_norm": 3.453125,
|
| 1590 |
+
"learning_rate": 4.631645039765593e-06,
|
| 1591 |
+
"loss": 1.1854766607284546,
|
| 1592 |
+
"step": 226
|
| 1593 |
+
},
|
| 1594 |
+
{
|
| 1595 |
+
"epoch": 0.09312820512820513,
|
| 1596 |
+
"grad_norm": 4.25,
|
| 1597 |
+
"learning_rate": 4.6295521138551695e-06,
|
| 1598 |
+
"loss": 1.2904247045516968,
|
| 1599 |
+
"step": 227
|
| 1600 |
+
},
|
| 1601 |
+
{
|
| 1602 |
+
"epoch": 0.09353846153846154,
|
| 1603 |
+
"grad_norm": 4.4375,
|
| 1604 |
+
"learning_rate": 4.627459187944747e-06,
|
| 1605 |
+
"loss": 1.1688357591629028,
|
| 1606 |
+
"step": 228
|
| 1607 |
+
},
|
| 1608 |
+
{
|
| 1609 |
+
"epoch": 0.09394871794871795,
|
| 1610 |
+
"grad_norm": 3.859375,
|
| 1611 |
+
"learning_rate": 4.625366262034324e-06,
|
| 1612 |
+
"loss": 1.247882604598999,
|
| 1613 |
+
"step": 229
|
| 1614 |
+
},
|
| 1615 |
+
{
|
| 1616 |
+
"epoch": 0.09435897435897436,
|
| 1617 |
+
"grad_norm": 4.15625,
|
| 1618 |
+
"learning_rate": 4.6232733361239015e-06,
|
| 1619 |
+
"loss": 1.2831999063491821,
|
| 1620 |
+
"step": 230
|
| 1621 |
+
},
|
| 1622 |
+
{
|
| 1623 |
+
"epoch": 0.09476923076923077,
|
| 1624 |
+
"grad_norm": 4.09375,
|
| 1625 |
+
"learning_rate": 4.621180410213479e-06,
|
| 1626 |
+
"loss": 1.2747775316238403,
|
| 1627 |
+
"step": 231
|
| 1628 |
+
},
|
| 1629 |
+
{
|
| 1630 |
+
"epoch": 0.09517948717948718,
|
| 1631 |
+
"grad_norm": 4.21875,
|
| 1632 |
+
"learning_rate": 4.619087484303056e-06,
|
| 1633 |
+
"loss": 1.293259859085083,
|
| 1634 |
+
"step": 232
|
| 1635 |
+
},
|
| 1636 |
+
{
|
| 1637 |
+
"epoch": 0.09558974358974359,
|
| 1638 |
+
"grad_norm": 4.375,
|
| 1639 |
+
"learning_rate": 4.616994558392634e-06,
|
| 1640 |
+
"loss": 1.2811206579208374,
|
| 1641 |
+
"step": 233
|
| 1642 |
+
},
|
| 1643 |
+
{
|
| 1644 |
+
"epoch": 0.096,
|
| 1645 |
+
"grad_norm": 4.25,
|
| 1646 |
+
"learning_rate": 4.61490163248221e-06,
|
| 1647 |
+
"loss": 1.1852827072143555,
|
| 1648 |
+
"step": 234
|
| 1649 |
+
},
|
| 1650 |
+
{
|
| 1651 |
+
"epoch": 0.09641025641025641,
|
| 1652 |
+
"grad_norm": 3.96875,
|
| 1653 |
+
"learning_rate": 4.612808706571787e-06,
|
| 1654 |
+
"loss": 1.3051080703735352,
|
| 1655 |
+
"step": 235
|
| 1656 |
+
},
|
| 1657 |
+
{
|
| 1658 |
+
"epoch": 0.09682051282051282,
|
| 1659 |
+
"grad_norm": 4.3125,
|
| 1660 |
+
"learning_rate": 4.610715780661365e-06,
|
| 1661 |
+
"loss": 1.2771573066711426,
|
| 1662 |
+
"step": 236
|
| 1663 |
+
},
|
| 1664 |
+
{
|
| 1665 |
+
"epoch": 0.09723076923076923,
|
| 1666 |
+
"grad_norm": 4.5,
|
| 1667 |
+
"learning_rate": 4.6086228547509425e-06,
|
| 1668 |
+
"loss": 1.4060215950012207,
|
| 1669 |
+
"step": 237
|
| 1670 |
+
},
|
| 1671 |
+
{
|
| 1672 |
+
"epoch": 0.09764102564102564,
|
| 1673 |
+
"grad_norm": 4.1875,
|
| 1674 |
+
"learning_rate": 4.606529928840519e-06,
|
| 1675 |
+
"loss": 1.3866937160491943,
|
| 1676 |
+
"step": 238
|
| 1677 |
+
},
|
| 1678 |
+
{
|
| 1679 |
+
"epoch": 0.09805128205128205,
|
| 1680 |
+
"grad_norm": 3.921875,
|
| 1681 |
+
"learning_rate": 4.604437002930097e-06,
|
| 1682 |
+
"loss": 1.1325360536575317,
|
| 1683 |
+
"step": 239
|
| 1684 |
+
},
|
| 1685 |
+
{
|
| 1686 |
+
"epoch": 0.09846153846153846,
|
| 1687 |
+
"grad_norm": 4.09375,
|
| 1688 |
+
"learning_rate": 4.602344077019674e-06,
|
| 1689 |
+
"loss": 1.1768242120742798,
|
| 1690 |
+
"step": 240
|
| 1691 |
+
},
|
| 1692 |
+
{
|
| 1693 |
+
"epoch": 0.09887179487179487,
|
| 1694 |
+
"grad_norm": 3.9375,
|
| 1695 |
+
"learning_rate": 4.600251151109251e-06,
|
| 1696 |
+
"loss": 1.2350736856460571,
|
| 1697 |
+
"step": 241
|
| 1698 |
+
},
|
| 1699 |
+
{
|
| 1700 |
+
"epoch": 0.09928205128205128,
|
| 1701 |
+
"grad_norm": 4.53125,
|
| 1702 |
+
"learning_rate": 4.598158225198828e-06,
|
| 1703 |
+
"loss": 1.2066761255264282,
|
| 1704 |
+
"step": 242
|
| 1705 |
+
},
|
| 1706 |
+
{
|
| 1707 |
+
"epoch": 0.09969230769230769,
|
| 1708 |
+
"grad_norm": 4.21875,
|
| 1709 |
+
"learning_rate": 4.596065299288406e-06,
|
| 1710 |
+
"loss": 1.3870455026626587,
|
| 1711 |
+
"step": 243
|
| 1712 |
+
},
|
| 1713 |
+
{
|
| 1714 |
+
"epoch": 0.1001025641025641,
|
| 1715 |
+
"grad_norm": 4.4375,
|
| 1716 |
+
"learning_rate": 4.5939723733779826e-06,
|
| 1717 |
+
"loss": 1.383860468864441,
|
| 1718 |
+
"step": 244
|
| 1719 |
+
},
|
| 1720 |
+
{
|
| 1721 |
+
"epoch": 0.10051282051282051,
|
| 1722 |
+
"grad_norm": 4.125,
|
| 1723 |
+
"learning_rate": 4.59187944746756e-06,
|
| 1724 |
+
"loss": 1.2680752277374268,
|
| 1725 |
+
"step": 245
|
| 1726 |
+
},
|
| 1727 |
+
{
|
| 1728 |
+
"epoch": 0.10092307692307692,
|
| 1729 |
+
"grad_norm": 4.0625,
|
| 1730 |
+
"learning_rate": 4.589786521557138e-06,
|
| 1731 |
+
"loss": 1.3729698657989502,
|
| 1732 |
+
"step": 246
|
| 1733 |
+
},
|
| 1734 |
+
{
|
| 1735 |
+
"epoch": 0.10133333333333333,
|
| 1736 |
+
"grad_norm": 4.09375,
|
| 1737 |
+
"learning_rate": 4.587693595646715e-06,
|
| 1738 |
+
"loss": 1.258432388305664,
|
| 1739 |
+
"step": 247
|
| 1740 |
+
},
|
| 1741 |
+
{
|
| 1742 |
+
"epoch": 0.10174358974358974,
|
| 1743 |
+
"grad_norm": 4.0625,
|
| 1744 |
+
"learning_rate": 4.585600669736291e-06,
|
| 1745 |
+
"loss": 1.37735116481781,
|
| 1746 |
+
"step": 248
|
| 1747 |
+
},
|
| 1748 |
+
{
|
| 1749 |
+
"epoch": 0.10215384615384615,
|
| 1750 |
+
"grad_norm": 4.6875,
|
| 1751 |
+
"learning_rate": 4.583507743825869e-06,
|
| 1752 |
+
"loss": 1.2726860046386719,
|
| 1753 |
+
"step": 249
|
| 1754 |
+
},
|
| 1755 |
+
{
|
| 1756 |
+
"epoch": 0.10256410256410256,
|
| 1757 |
+
"grad_norm": 3.65625,
|
| 1758 |
+
"learning_rate": 4.581414817915446e-06,
|
| 1759 |
+
"loss": 1.2707538604736328,
|
| 1760 |
+
"step": 250
|
| 1761 |
+
},
|
| 1762 |
+
{
|
| 1763 |
+
"epoch": 0.10297435897435897,
|
| 1764 |
+
"grad_norm": 4.1875,
|
| 1765 |
+
"learning_rate": 4.5793218920050235e-06,
|
| 1766 |
+
"loss": 1.2484283447265625,
|
| 1767 |
+
"step": 251
|
| 1768 |
+
},
|
| 1769 |
+
{
|
| 1770 |
+
"epoch": 0.10338461538461538,
|
| 1771 |
+
"grad_norm": 3.96875,
|
| 1772 |
+
"learning_rate": 4.5772289660946e-06,
|
| 1773 |
+
"loss": 1.1953047513961792,
|
| 1774 |
+
"step": 252
|
| 1775 |
+
},
|
| 1776 |
+
{
|
| 1777 |
+
"epoch": 0.10379487179487179,
|
| 1778 |
+
"grad_norm": 4.46875,
|
| 1779 |
+
"learning_rate": 4.575136040184178e-06,
|
| 1780 |
+
"loss": 1.1912050247192383,
|
| 1781 |
+
"step": 253
|
| 1782 |
+
},
|
| 1783 |
+
{
|
| 1784 |
+
"epoch": 0.1042051282051282,
|
| 1785 |
+
"grad_norm": 3.84375,
|
| 1786 |
+
"learning_rate": 4.5730431142737555e-06,
|
| 1787 |
+
"loss": 1.1780365705490112,
|
| 1788 |
+
"step": 254
|
| 1789 |
+
},
|
| 1790 |
+
{
|
| 1791 |
+
"epoch": 0.10461538461538461,
|
| 1792 |
+
"grad_norm": 4.625,
|
| 1793 |
+
"learning_rate": 4.570950188363332e-06,
|
| 1794 |
+
"loss": 1.330933690071106,
|
| 1795 |
+
"step": 255
|
| 1796 |
+
},
|
| 1797 |
+
{
|
| 1798 |
+
"epoch": 0.10502564102564102,
|
| 1799 |
+
"grad_norm": 4.375,
|
| 1800 |
+
"learning_rate": 4.568857262452909e-06,
|
| 1801 |
+
"loss": 1.2700480222702026,
|
| 1802 |
+
"step": 256
|
| 1803 |
+
},
|
| 1804 |
+
{
|
| 1805 |
+
"epoch": 0.10543589743589743,
|
| 1806 |
+
"grad_norm": 3.828125,
|
| 1807 |
+
"learning_rate": 4.566764336542487e-06,
|
| 1808 |
+
"loss": 1.2775812149047852,
|
| 1809 |
+
"step": 257
|
| 1810 |
+
},
|
| 1811 |
+
{
|
| 1812 |
+
"epoch": 0.10584615384615384,
|
| 1813 |
+
"grad_norm": 3.78125,
|
| 1814 |
+
"learning_rate": 4.5646714106320636e-06,
|
| 1815 |
+
"loss": 1.348310947418213,
|
| 1816 |
+
"step": 258
|
| 1817 |
+
},
|
| 1818 |
+
{
|
| 1819 |
+
"epoch": 0.10625641025641025,
|
| 1820 |
+
"grad_norm": 4.5,
|
| 1821 |
+
"learning_rate": 4.562578484721641e-06,
|
| 1822 |
+
"loss": 1.4222533702850342,
|
| 1823 |
+
"step": 259
|
| 1824 |
+
},
|
| 1825 |
+
{
|
| 1826 |
+
"epoch": 0.10666666666666667,
|
| 1827 |
+
"grad_norm": 4.21875,
|
| 1828 |
+
"learning_rate": 4.560485558811219e-06,
|
| 1829 |
+
"loss": 1.369185447692871,
|
| 1830 |
+
"step": 260
|
| 1831 |
+
},
|
| 1832 |
+
{
|
| 1833 |
+
"epoch": 0.10707692307692308,
|
| 1834 |
+
"grad_norm": 4.03125,
|
| 1835 |
+
"learning_rate": 4.558392632900796e-06,
|
| 1836 |
+
"loss": 1.157382845878601,
|
| 1837 |
+
"step": 261
|
| 1838 |
+
},
|
| 1839 |
+
{
|
| 1840 |
+
"epoch": 0.1074871794871795,
|
| 1841 |
+
"grad_norm": 4.375,
|
| 1842 |
+
"learning_rate": 4.556299706990373e-06,
|
| 1843 |
+
"loss": 1.1591321229934692,
|
| 1844 |
+
"step": 262
|
| 1845 |
+
},
|
| 1846 |
+
{
|
| 1847 |
+
"epoch": 0.1078974358974359,
|
| 1848 |
+
"grad_norm": 4.0625,
|
| 1849 |
+
"learning_rate": 4.55420678107995e-06,
|
| 1850 |
+
"loss": 1.2198688983917236,
|
| 1851 |
+
"step": 263
|
| 1852 |
+
},
|
| 1853 |
+
{
|
| 1854 |
+
"epoch": 0.10830769230769231,
|
| 1855 |
+
"grad_norm": 4.0,
|
| 1856 |
+
"learning_rate": 4.552113855169528e-06,
|
| 1857 |
+
"loss": 1.1403449773788452,
|
| 1858 |
+
"step": 264
|
| 1859 |
+
},
|
| 1860 |
+
{
|
| 1861 |
+
"epoch": 0.10871794871794872,
|
| 1862 |
+
"grad_norm": 4.125,
|
| 1863 |
+
"learning_rate": 4.5500209292591045e-06,
|
| 1864 |
+
"loss": 1.2642017602920532,
|
| 1865 |
+
"step": 265
|
| 1866 |
+
},
|
| 1867 |
+
{
|
| 1868 |
+
"epoch": 0.10912820512820513,
|
| 1869 |
+
"grad_norm": 4.1875,
|
| 1870 |
+
"learning_rate": 4.547928003348681e-06,
|
| 1871 |
+
"loss": 1.4000636339187622,
|
| 1872 |
+
"step": 266
|
| 1873 |
+
},
|
| 1874 |
+
{
|
| 1875 |
+
"epoch": 0.10953846153846154,
|
| 1876 |
+
"grad_norm": 4.0,
|
| 1877 |
+
"learning_rate": 4.545835077438259e-06,
|
| 1878 |
+
"loss": 1.279004693031311,
|
| 1879 |
+
"step": 267
|
| 1880 |
+
},
|
| 1881 |
+
{
|
| 1882 |
+
"epoch": 0.10994871794871795,
|
| 1883 |
+
"grad_norm": 4.46875,
|
| 1884 |
+
"learning_rate": 4.5437421515278365e-06,
|
| 1885 |
+
"loss": 1.2803137302398682,
|
| 1886 |
+
"step": 268
|
| 1887 |
+
},
|
| 1888 |
+
{
|
| 1889 |
+
"epoch": 0.11035897435897436,
|
| 1890 |
+
"grad_norm": 4.0625,
|
| 1891 |
+
"learning_rate": 4.541649225617413e-06,
|
| 1892 |
+
"loss": 1.2438591718673706,
|
| 1893 |
+
"step": 269
|
| 1894 |
+
},
|
| 1895 |
+
{
|
| 1896 |
+
"epoch": 0.11076923076923077,
|
| 1897 |
+
"grad_norm": 4.0625,
|
| 1898 |
+
"learning_rate": 4.539556299706991e-06,
|
| 1899 |
+
"loss": 1.3200500011444092,
|
| 1900 |
+
"step": 270
|
| 1901 |
+
},
|
| 1902 |
+
{
|
| 1903 |
+
"epoch": 0.11117948717948718,
|
| 1904 |
+
"grad_norm": 4.15625,
|
| 1905 |
+
"learning_rate": 4.537463373796568e-06,
|
| 1906 |
+
"loss": 1.1718605756759644,
|
| 1907 |
+
"step": 271
|
| 1908 |
+
},
|
| 1909 |
+
{
|
| 1910 |
+
"epoch": 0.11158974358974359,
|
| 1911 |
+
"grad_norm": 3.953125,
|
| 1912 |
+
"learning_rate": 4.535370447886145e-06,
|
| 1913 |
+
"loss": 1.036754846572876,
|
| 1914 |
+
"step": 272
|
| 1915 |
+
},
|
| 1916 |
+
{
|
| 1917 |
+
"epoch": 0.112,
|
| 1918 |
+
"grad_norm": 3.875,
|
| 1919 |
+
"learning_rate": 4.533277521975722e-06,
|
| 1920 |
+
"loss": 1.1870700120925903,
|
| 1921 |
+
"step": 273
|
| 1922 |
+
},
|
| 1923 |
+
{
|
| 1924 |
+
"epoch": 0.11241025641025641,
|
| 1925 |
+
"grad_norm": 4.34375,
|
| 1926 |
+
"learning_rate": 4.5311845960653e-06,
|
| 1927 |
+
"loss": 1.3146507740020752,
|
| 1928 |
+
"step": 274
|
| 1929 |
+
},
|
| 1930 |
+
{
|
| 1931 |
+
"epoch": 0.11282051282051282,
|
| 1932 |
+
"grad_norm": 3.796875,
|
| 1933 |
+
"learning_rate": 4.5290916701548775e-06,
|
| 1934 |
+
"loss": 1.1766315698623657,
|
| 1935 |
+
"step": 275
|
| 1936 |
+
},
|
| 1937 |
+
{
|
| 1938 |
+
"epoch": 0.11323076923076923,
|
| 1939 |
+
"grad_norm": 4.09375,
|
| 1940 |
+
"learning_rate": 4.526998744244454e-06,
|
| 1941 |
+
"loss": 1.175342082977295,
|
| 1942 |
+
"step": 276
|
| 1943 |
+
},
|
| 1944 |
+
{
|
| 1945 |
+
"epoch": 0.11364102564102564,
|
| 1946 |
+
"grad_norm": 3.828125,
|
| 1947 |
+
"learning_rate": 4.524905818334031e-06,
|
| 1948 |
+
"loss": 1.3296524286270142,
|
| 1949 |
+
"step": 277
|
| 1950 |
+
},
|
| 1951 |
+
{
|
| 1952 |
+
"epoch": 0.11405128205128205,
|
| 1953 |
+
"grad_norm": 4.375,
|
| 1954 |
+
"learning_rate": 4.522812892423609e-06,
|
| 1955 |
+
"loss": 1.124447226524353,
|
| 1956 |
+
"step": 278
|
| 1957 |
+
},
|
| 1958 |
+
{
|
| 1959 |
+
"epoch": 0.11446153846153846,
|
| 1960 |
+
"grad_norm": 4.59375,
|
| 1961 |
+
"learning_rate": 4.5207199665131855e-06,
|
| 1962 |
+
"loss": 1.2079598903656006,
|
| 1963 |
+
"step": 279
|
| 1964 |
+
},
|
| 1965 |
+
{
|
| 1966 |
+
"epoch": 0.11487179487179487,
|
| 1967 |
+
"grad_norm": 4.1875,
|
| 1968 |
+
"learning_rate": 4.518627040602763e-06,
|
| 1969 |
+
"loss": 1.309114933013916,
|
| 1970 |
+
"step": 280
|
| 1971 |
+
},
|
| 1972 |
+
{
|
| 1973 |
+
"epoch": 0.11528205128205128,
|
| 1974 |
+
"grad_norm": 3.984375,
|
| 1975 |
+
"learning_rate": 4.51653411469234e-06,
|
| 1976 |
+
"loss": 1.150520920753479,
|
| 1977 |
+
"step": 281
|
| 1978 |
+
},
|
| 1979 |
+
{
|
| 1980 |
+
"epoch": 0.11569230769230769,
|
| 1981 |
+
"grad_norm": 3.828125,
|
| 1982 |
+
"learning_rate": 4.5144411887819175e-06,
|
| 1983 |
+
"loss": 1.1144882440567017,
|
| 1984 |
+
"step": 282
|
| 1985 |
+
},
|
| 1986 |
+
{
|
| 1987 |
+
"epoch": 0.1161025641025641,
|
| 1988 |
+
"grad_norm": 4.09375,
|
| 1989 |
+
"learning_rate": 4.512348262871495e-06,
|
| 1990 |
+
"loss": 1.2613205909729004,
|
| 1991 |
+
"step": 283
|
| 1992 |
+
},
|
| 1993 |
+
{
|
| 1994 |
+
"epoch": 0.11651282051282051,
|
| 1995 |
+
"grad_norm": 4.40625,
|
| 1996 |
+
"learning_rate": 4.510255336961072e-06,
|
| 1997 |
+
"loss": 1.2692985534667969,
|
| 1998 |
+
"step": 284
|
| 1999 |
+
},
|
| 2000 |
+
{
|
| 2001 |
+
"epoch": 0.11692307692307692,
|
| 2002 |
+
"grad_norm": 4.0,
|
| 2003 |
+
"learning_rate": 4.50816241105065e-06,
|
| 2004 |
+
"loss": 1.2029545307159424,
|
| 2005 |
+
"step": 285
|
| 2006 |
+
},
|
| 2007 |
+
{
|
| 2008 |
+
"epoch": 0.11733333333333333,
|
| 2009 |
+
"grad_norm": 4.34375,
|
| 2010 |
+
"learning_rate": 4.506069485140226e-06,
|
| 2011 |
+
"loss": 1.180769681930542,
|
| 2012 |
+
"step": 286
|
| 2013 |
+
},
|
| 2014 |
+
{
|
| 2015 |
+
"epoch": 0.11774358974358974,
|
| 2016 |
+
"grad_norm": 3.859375,
|
| 2017 |
+
"learning_rate": 4.503976559229803e-06,
|
| 2018 |
+
"loss": 1.1751469373703003,
|
| 2019 |
+
"step": 287
|
| 2020 |
+
},
|
| 2021 |
+
{
|
| 2022 |
+
"epoch": 0.11815384615384615,
|
| 2023 |
+
"grad_norm": 4.375,
|
| 2024 |
+
"learning_rate": 4.501883633319381e-06,
|
| 2025 |
+
"loss": 1.1570502519607544,
|
| 2026 |
+
"step": 288
|
| 2027 |
+
},
|
| 2028 |
+
{
|
| 2029 |
+
"epoch": 0.11856410256410256,
|
| 2030 |
+
"grad_norm": 4.3125,
|
| 2031 |
+
"learning_rate": 4.4997907074089585e-06,
|
| 2032 |
+
"loss": 1.296753168106079,
|
| 2033 |
+
"step": 289
|
| 2034 |
+
},
|
| 2035 |
+
{
|
| 2036 |
+
"epoch": 0.11897435897435897,
|
| 2037 |
+
"grad_norm": 4.125,
|
| 2038 |
+
"learning_rate": 4.497697781498535e-06,
|
| 2039 |
+
"loss": 1.2842097282409668,
|
| 2040 |
+
"step": 290
|
| 2041 |
+
},
|
| 2042 |
+
{
|
| 2043 |
+
"epoch": 0.11938461538461538,
|
| 2044 |
+
"grad_norm": 4.21875,
|
| 2045 |
+
"learning_rate": 4.495604855588113e-06,
|
| 2046 |
+
"loss": 1.2111274003982544,
|
| 2047 |
+
"step": 291
|
| 2048 |
+
},
|
| 2049 |
+
{
|
| 2050 |
+
"epoch": 0.11979487179487179,
|
| 2051 |
+
"grad_norm": 4.03125,
|
| 2052 |
+
"learning_rate": 4.49351192967769e-06,
|
| 2053 |
+
"loss": 1.1865339279174805,
|
| 2054 |
+
"step": 292
|
| 2055 |
+
},
|
| 2056 |
+
{
|
| 2057 |
+
"epoch": 0.1202051282051282,
|
| 2058 |
+
"grad_norm": 4.28125,
|
| 2059 |
+
"learning_rate": 4.491419003767267e-06,
|
| 2060 |
+
"loss": 1.2637746334075928,
|
| 2061 |
+
"step": 293
|
| 2062 |
+
},
|
| 2063 |
+
{
|
| 2064 |
+
"epoch": 0.12061538461538461,
|
| 2065 |
+
"grad_norm": 4.40625,
|
| 2066 |
+
"learning_rate": 4.489326077856844e-06,
|
| 2067 |
+
"loss": 1.3115350008010864,
|
| 2068 |
+
"step": 294
|
| 2069 |
+
},
|
| 2070 |
+
{
|
| 2071 |
+
"epoch": 0.12102564102564102,
|
| 2072 |
+
"grad_norm": 4.3125,
|
| 2073 |
+
"learning_rate": 4.487233151946421e-06,
|
| 2074 |
+
"loss": 1.2570511102676392,
|
| 2075 |
+
"step": 295
|
| 2076 |
+
},
|
| 2077 |
+
{
|
| 2078 |
+
"epoch": 0.12143589743589743,
|
| 2079 |
+
"grad_norm": 4.21875,
|
| 2080 |
+
"learning_rate": 4.4851402260359985e-06,
|
| 2081 |
+
"loss": 1.2771306037902832,
|
| 2082 |
+
"step": 296
|
| 2083 |
+
},
|
| 2084 |
+
{
|
| 2085 |
+
"epoch": 0.12184615384615384,
|
| 2086 |
+
"grad_norm": 4.0625,
|
| 2087 |
+
"learning_rate": 4.483047300125576e-06,
|
| 2088 |
+
"loss": 1.2327501773834229,
|
| 2089 |
+
"step": 297
|
| 2090 |
+
},
|
| 2091 |
+
{
|
| 2092 |
+
"epoch": 0.12225641025641025,
|
| 2093 |
+
"grad_norm": 4.34375,
|
| 2094 |
+
"learning_rate": 4.480954374215153e-06,
|
| 2095 |
+
"loss": 1.22713041305542,
|
| 2096 |
+
"step": 298
|
| 2097 |
+
},
|
| 2098 |
+
{
|
| 2099 |
+
"epoch": 0.12266666666666666,
|
| 2100 |
+
"grad_norm": 3.953125,
|
| 2101 |
+
"learning_rate": 4.478861448304731e-06,
|
| 2102 |
+
"loss": 1.1757054328918457,
|
| 2103 |
+
"step": 299
|
| 2104 |
+
},
|
| 2105 |
+
{
|
| 2106 |
+
"epoch": 0.12307692307692308,
|
| 2107 |
+
"grad_norm": 4.3125,
|
| 2108 |
+
"learning_rate": 4.476768522394307e-06,
|
| 2109 |
+
"loss": 1.2990862131118774,
|
| 2110 |
+
"step": 300
|
| 2111 |
+
},
|
| 2112 |
+
{
|
| 2113 |
+
"epoch": 0.1234871794871795,
|
| 2114 |
+
"grad_norm": 3.71875,
|
| 2115 |
+
"learning_rate": 4.474675596483885e-06,
|
| 2116 |
+
"loss": 1.283048152923584,
|
| 2117 |
+
"step": 301
|
| 2118 |
+
},
|
| 2119 |
+
{
|
| 2120 |
+
"epoch": 0.1238974358974359,
|
| 2121 |
+
"grad_norm": 4.6875,
|
| 2122 |
+
"learning_rate": 4.472582670573462e-06,
|
| 2123 |
+
"loss": 1.3461682796478271,
|
| 2124 |
+
"step": 302
|
| 2125 |
+
},
|
| 2126 |
+
{
|
| 2127 |
+
"epoch": 0.12430769230769231,
|
| 2128 |
+
"grad_norm": 4.0625,
|
| 2129 |
+
"learning_rate": 4.4704897446630395e-06,
|
| 2130 |
+
"loss": 1.217667818069458,
|
| 2131 |
+
"step": 303
|
| 2132 |
+
},
|
| 2133 |
+
{
|
| 2134 |
+
"epoch": 0.12471794871794872,
|
| 2135 |
+
"grad_norm": 4.34375,
|
| 2136 |
+
"learning_rate": 4.468396818752616e-06,
|
| 2137 |
+
"loss": 1.2673282623291016,
|
| 2138 |
+
"step": 304
|
| 2139 |
+
},
|
| 2140 |
+
{
|
| 2141 |
+
"epoch": 0.12512820512820513,
|
| 2142 |
+
"grad_norm": 3.796875,
|
| 2143 |
+
"learning_rate": 4.466303892842194e-06,
|
| 2144 |
+
"loss": 1.2111177444458008,
|
| 2145 |
+
"step": 305
|
| 2146 |
+
},
|
| 2147 |
+
{
|
| 2148 |
+
"epoch": 0.12553846153846154,
|
| 2149 |
+
"grad_norm": 4.15625,
|
| 2150 |
+
"learning_rate": 4.4642109669317715e-06,
|
| 2151 |
+
"loss": 1.2601410150527954,
|
| 2152 |
+
"step": 306
|
| 2153 |
+
},
|
| 2154 |
+
{
|
| 2155 |
+
"epoch": 0.12594871794871795,
|
| 2156 |
+
"grad_norm": 4.15625,
|
| 2157 |
+
"learning_rate": 4.462118041021348e-06,
|
| 2158 |
+
"loss": 1.2759875059127808,
|
| 2159 |
+
"step": 307
|
| 2160 |
+
},
|
| 2161 |
+
{
|
| 2162 |
+
"epoch": 0.12635897435897436,
|
| 2163 |
+
"grad_norm": 4.09375,
|
| 2164 |
+
"learning_rate": 4.460025115110925e-06,
|
| 2165 |
+
"loss": 1.2334433794021606,
|
| 2166 |
+
"step": 308
|
| 2167 |
+
},
|
| 2168 |
+
{
|
| 2169 |
+
"epoch": 0.12676923076923077,
|
| 2170 |
+
"grad_norm": 4.28125,
|
| 2171 |
+
"learning_rate": 4.457932189200503e-06,
|
| 2172 |
+
"loss": 1.2810615301132202,
|
| 2173 |
+
"step": 309
|
| 2174 |
+
},
|
| 2175 |
+
{
|
| 2176 |
+
"epoch": 0.12717948717948718,
|
| 2177 |
+
"grad_norm": 4.0625,
|
| 2178 |
+
"learning_rate": 4.4558392632900796e-06,
|
| 2179 |
+
"loss": 1.2481120824813843,
|
| 2180 |
+
"step": 310
|
| 2181 |
+
},
|
| 2182 |
+
{
|
| 2183 |
+
"epoch": 0.1275897435897436,
|
| 2184 |
+
"grad_norm": 4.375,
|
| 2185 |
+
"learning_rate": 4.453746337379657e-06,
|
| 2186 |
+
"loss": 1.2821013927459717,
|
| 2187 |
+
"step": 311
|
| 2188 |
+
},
|
| 2189 |
+
{
|
| 2190 |
+
"epoch": 0.128,
|
| 2191 |
+
"grad_norm": 4.21875,
|
| 2192 |
+
"learning_rate": 4.451653411469235e-06,
|
| 2193 |
+
"loss": 1.0487443208694458,
|
| 2194 |
+
"step": 312
|
| 2195 |
+
},
|
| 2196 |
+
{
|
| 2197 |
+
"epoch": 0.1284102564102564,
|
| 2198 |
+
"grad_norm": 3.78125,
|
| 2199 |
+
"learning_rate": 4.449560485558812e-06,
|
| 2200 |
+
"loss": 1.1363240480422974,
|
| 2201 |
+
"step": 313
|
| 2202 |
+
},
|
| 2203 |
+
{
|
| 2204 |
+
"epoch": 0.12882051282051282,
|
| 2205 |
+
"grad_norm": 3.828125,
|
| 2206 |
+
"learning_rate": 4.447467559648389e-06,
|
| 2207 |
+
"loss": 1.112923502922058,
|
| 2208 |
+
"step": 314
|
| 2209 |
+
},
|
| 2210 |
+
{
|
| 2211 |
+
"epoch": 0.12923076923076923,
|
| 2212 |
+
"grad_norm": 3.765625,
|
| 2213 |
+
"learning_rate": 4.445374633737966e-06,
|
| 2214 |
+
"loss": 1.161982536315918,
|
| 2215 |
+
"step": 315
|
| 2216 |
+
},
|
| 2217 |
+
{
|
| 2218 |
+
"epoch": 0.12964102564102564,
|
| 2219 |
+
"grad_norm": 4.03125,
|
| 2220 |
+
"learning_rate": 4.443281707827543e-06,
|
| 2221 |
+
"loss": 1.288543939590454,
|
| 2222 |
+
"step": 316
|
| 2223 |
+
},
|
| 2224 |
+
{
|
| 2225 |
+
"epoch": 0.13005128205128205,
|
| 2226 |
+
"grad_norm": 4.09375,
|
| 2227 |
+
"learning_rate": 4.4411887819171205e-06,
|
| 2228 |
+
"loss": 1.2287678718566895,
|
| 2229 |
+
"step": 317
|
| 2230 |
+
},
|
| 2231 |
+
{
|
| 2232 |
+
"epoch": 0.13046153846153846,
|
| 2233 |
+
"grad_norm": 3.96875,
|
| 2234 |
+
"learning_rate": 4.439095856006697e-06,
|
| 2235 |
+
"loss": 1.1338226795196533,
|
| 2236 |
+
"step": 318
|
| 2237 |
+
},
|
| 2238 |
+
{
|
| 2239 |
+
"epoch": 0.13087179487179487,
|
| 2240 |
+
"grad_norm": 4.03125,
|
| 2241 |
+
"learning_rate": 4.437002930096275e-06,
|
| 2242 |
+
"loss": 1.1916424036026,
|
| 2243 |
+
"step": 319
|
| 2244 |
+
},
|
| 2245 |
+
{
|
| 2246 |
+
"epoch": 0.13128205128205128,
|
| 2247 |
+
"grad_norm": 3.921875,
|
| 2248 |
+
"learning_rate": 4.4349100041858525e-06,
|
| 2249 |
+
"loss": 1.1514812707901,
|
| 2250 |
+
"step": 320
|
| 2251 |
+
},
|
| 2252 |
+
{
|
| 2253 |
+
"epoch": 0.1316923076923077,
|
| 2254 |
+
"grad_norm": 4.03125,
|
| 2255 |
+
"learning_rate": 4.432817078275429e-06,
|
| 2256 |
+
"loss": 1.203607201576233,
|
| 2257 |
+
"step": 321
|
| 2258 |
+
},
|
| 2259 |
+
{
|
| 2260 |
+
"epoch": 0.1321025641025641,
|
| 2261 |
+
"grad_norm": 4.0,
|
| 2262 |
+
"learning_rate": 4.430724152365007e-06,
|
| 2263 |
+
"loss": 1.2071082592010498,
|
| 2264 |
+
"step": 322
|
| 2265 |
+
},
|
| 2266 |
+
{
|
| 2267 |
+
"epoch": 0.1325128205128205,
|
| 2268 |
+
"grad_norm": 4.625,
|
| 2269 |
+
"learning_rate": 4.428631226454584e-06,
|
| 2270 |
+
"loss": 1.3184986114501953,
|
| 2271 |
+
"step": 323
|
| 2272 |
+
},
|
| 2273 |
+
{
|
| 2274 |
+
"epoch": 0.13292307692307692,
|
| 2275 |
+
"grad_norm": 3.8125,
|
| 2276 |
+
"learning_rate": 4.4265383005441606e-06,
|
| 2277 |
+
"loss": 1.2131890058517456,
|
| 2278 |
+
"step": 324
|
| 2279 |
+
},
|
| 2280 |
+
{
|
| 2281 |
+
"epoch": 0.13333333333333333,
|
| 2282 |
+
"grad_norm": 4.28125,
|
| 2283 |
+
"learning_rate": 4.424445374633738e-06,
|
| 2284 |
+
"loss": 1.3523532152175903,
|
| 2285 |
+
"step": 325
|
| 2286 |
+
},
|
| 2287 |
+
{
|
| 2288 |
+
"epoch": 0.13374358974358974,
|
| 2289 |
+
"grad_norm": 4.28125,
|
| 2290 |
+
"learning_rate": 4.422352448723316e-06,
|
| 2291 |
+
"loss": 1.0918247699737549,
|
| 2292 |
+
"step": 326
|
| 2293 |
+
},
|
| 2294 |
+
{
|
| 2295 |
+
"epoch": 0.13415384615384615,
|
| 2296 |
+
"grad_norm": 4.15625,
|
| 2297 |
+
"learning_rate": 4.4202595228128935e-06,
|
| 2298 |
+
"loss": 1.3244166374206543,
|
| 2299 |
+
"step": 327
|
| 2300 |
+
},
|
| 2301 |
+
{
|
| 2302 |
+
"epoch": 0.13456410256410256,
|
| 2303 |
+
"grad_norm": 3.796875,
|
| 2304 |
+
"learning_rate": 4.41816659690247e-06,
|
| 2305 |
+
"loss": 1.2182252407073975,
|
| 2306 |
+
"step": 328
|
| 2307 |
+
},
|
| 2308 |
+
{
|
| 2309 |
+
"epoch": 0.13497435897435897,
|
| 2310 |
+
"grad_norm": 4.1875,
|
| 2311 |
+
"learning_rate": 4.416073670992047e-06,
|
| 2312 |
+
"loss": 1.2532880306243896,
|
| 2313 |
+
"step": 329
|
| 2314 |
+
},
|
| 2315 |
+
{
|
| 2316 |
+
"epoch": 0.13538461538461538,
|
| 2317 |
+
"grad_norm": 4.28125,
|
| 2318 |
+
"learning_rate": 4.413980745081625e-06,
|
| 2319 |
+
"loss": 1.2696692943572998,
|
| 2320 |
+
"step": 330
|
| 2321 |
+
},
|
| 2322 |
+
{
|
| 2323 |
+
"epoch": 0.1357948717948718,
|
| 2324 |
+
"grad_norm": 4.34375,
|
| 2325 |
+
"learning_rate": 4.4118878191712015e-06,
|
| 2326 |
+
"loss": 1.3265308141708374,
|
| 2327 |
+
"step": 331
|
| 2328 |
+
},
|
| 2329 |
+
{
|
| 2330 |
+
"epoch": 0.1362051282051282,
|
| 2331 |
+
"grad_norm": 3.875,
|
| 2332 |
+
"learning_rate": 4.409794893260779e-06,
|
| 2333 |
+
"loss": 1.0906741619110107,
|
| 2334 |
+
"step": 332
|
| 2335 |
+
},
|
| 2336 |
+
{
|
| 2337 |
+
"epoch": 0.1366153846153846,
|
| 2338 |
+
"grad_norm": 4.0625,
|
| 2339 |
+
"learning_rate": 4.407701967350356e-06,
|
| 2340 |
+
"loss": 1.2565168142318726,
|
| 2341 |
+
"step": 333
|
| 2342 |
+
},
|
| 2343 |
+
{
|
| 2344 |
+
"epoch": 0.13702564102564102,
|
| 2345 |
+
"grad_norm": 4.03125,
|
| 2346 |
+
"learning_rate": 4.4056090414399335e-06,
|
| 2347 |
+
"loss": 1.2574559450149536,
|
| 2348 |
+
"step": 334
|
| 2349 |
+
},
|
| 2350 |
+
{
|
| 2351 |
+
"epoch": 0.13743589743589743,
|
| 2352 |
+
"grad_norm": 3.859375,
|
| 2353 |
+
"learning_rate": 4.403516115529511e-06,
|
| 2354 |
+
"loss": 1.2529447078704834,
|
| 2355 |
+
"step": 335
|
| 2356 |
+
},
|
| 2357 |
+
{
|
| 2358 |
+
"epoch": 0.13784615384615384,
|
| 2359 |
+
"grad_norm": 4.15625,
|
| 2360 |
+
"learning_rate": 4.401423189619088e-06,
|
| 2361 |
+
"loss": 1.2135076522827148,
|
| 2362 |
+
"step": 336
|
| 2363 |
+
},
|
| 2364 |
+
{
|
| 2365 |
+
"epoch": 0.13825641025641025,
|
| 2366 |
+
"grad_norm": 3.75,
|
| 2367 |
+
"learning_rate": 4.399330263708665e-06,
|
| 2368 |
+
"loss": 1.2332839965820312,
|
| 2369 |
+
"step": 337
|
| 2370 |
+
},
|
| 2371 |
+
{
|
| 2372 |
+
"epoch": 0.13866666666666666,
|
| 2373 |
+
"grad_norm": 4.5625,
|
| 2374 |
+
"learning_rate": 4.397237337798242e-06,
|
| 2375 |
+
"loss": 1.3413827419281006,
|
| 2376 |
+
"step": 338
|
| 2377 |
+
},
|
| 2378 |
+
{
|
| 2379 |
+
"epoch": 0.13907692307692307,
|
| 2380 |
+
"grad_norm": 3.828125,
|
| 2381 |
+
"learning_rate": 4.395144411887819e-06,
|
| 2382 |
+
"loss": 1.1728384494781494,
|
| 2383 |
+
"step": 339
|
| 2384 |
+
},
|
| 2385 |
+
{
|
| 2386 |
+
"epoch": 0.13948717948717948,
|
| 2387 |
+
"grad_norm": 4.25,
|
| 2388 |
+
"learning_rate": 4.393051485977397e-06,
|
| 2389 |
+
"loss": 1.2645047903060913,
|
| 2390 |
+
"step": 340
|
| 2391 |
+
},
|
| 2392 |
+
{
|
| 2393 |
+
"epoch": 0.1398974358974359,
|
| 2394 |
+
"grad_norm": 4.1875,
|
| 2395 |
+
"learning_rate": 4.3909585600669745e-06,
|
| 2396 |
+
"loss": 1.0893607139587402,
|
| 2397 |
+
"step": 341
|
| 2398 |
+
},
|
| 2399 |
+
{
|
| 2400 |
+
"epoch": 0.1403076923076923,
|
| 2401 |
+
"grad_norm": 4.65625,
|
| 2402 |
+
"learning_rate": 4.388865634156551e-06,
|
| 2403 |
+
"loss": 1.2841830253601074,
|
| 2404 |
+
"step": 342
|
| 2405 |
+
},
|
| 2406 |
+
{
|
| 2407 |
+
"epoch": 0.1407179487179487,
|
| 2408 |
+
"grad_norm": 3.96875,
|
| 2409 |
+
"learning_rate": 4.386772708246129e-06,
|
| 2410 |
+
"loss": 1.2927216291427612,
|
| 2411 |
+
"step": 343
|
| 2412 |
+
},
|
| 2413 |
+
{
|
| 2414 |
+
"epoch": 0.14112820512820512,
|
| 2415 |
+
"grad_norm": 4.21875,
|
| 2416 |
+
"learning_rate": 4.384679782335706e-06,
|
| 2417 |
+
"loss": 1.1706403493881226,
|
| 2418 |
+
"step": 344
|
| 2419 |
+
},
|
| 2420 |
+
{
|
| 2421 |
+
"epoch": 0.14153846153846153,
|
| 2422 |
+
"grad_norm": 4.0625,
|
| 2423 |
+
"learning_rate": 4.3825868564252825e-06,
|
| 2424 |
+
"loss": 1.3131035566329956,
|
| 2425 |
+
"step": 345
|
| 2426 |
+
},
|
| 2427 |
+
{
|
| 2428 |
+
"epoch": 0.14194871794871794,
|
| 2429 |
+
"grad_norm": 3.625,
|
| 2430 |
+
"learning_rate": 4.38049393051486e-06,
|
| 2431 |
+
"loss": 1.0559601783752441,
|
| 2432 |
+
"step": 346
|
| 2433 |
+
},
|
| 2434 |
+
{
|
| 2435 |
+
"epoch": 0.14235897435897435,
|
| 2436 |
+
"grad_norm": 4.3125,
|
| 2437 |
+
"learning_rate": 4.378401004604437e-06,
|
| 2438 |
+
"loss": 1.299569010734558,
|
| 2439 |
+
"step": 347
|
| 2440 |
+
},
|
| 2441 |
+
{
|
| 2442 |
+
"epoch": 0.14276923076923076,
|
| 2443 |
+
"grad_norm": 4.375,
|
| 2444 |
+
"learning_rate": 4.3763080786940145e-06,
|
| 2445 |
+
"loss": 1.2612409591674805,
|
| 2446 |
+
"step": 348
|
| 2447 |
+
},
|
| 2448 |
+
{
|
| 2449 |
+
"epoch": 0.14317948717948717,
|
| 2450 |
+
"grad_norm": 4.125,
|
| 2451 |
+
"learning_rate": 4.374215152783592e-06,
|
| 2452 |
+
"loss": 1.240541696548462,
|
| 2453 |
+
"step": 349
|
| 2454 |
+
},
|
| 2455 |
+
{
|
| 2456 |
+
"epoch": 0.14358974358974358,
|
| 2457 |
+
"grad_norm": 4.0,
|
| 2458 |
+
"learning_rate": 4.372122226873169e-06,
|
| 2459 |
+
"loss": 1.298937201499939,
|
| 2460 |
+
"step": 350
|
| 2461 |
+
},
|
| 2462 |
+
{
|
| 2463 |
+
"epoch": 0.144,
|
| 2464 |
+
"grad_norm": 4.5,
|
| 2465 |
+
"learning_rate": 4.370029300962747e-06,
|
| 2466 |
+
"loss": 1.153456211090088,
|
| 2467 |
+
"step": 351
|
| 2468 |
+
},
|
| 2469 |
+
{
|
| 2470 |
+
"epoch": 0.1444102564102564,
|
| 2471 |
+
"grad_norm": 4.03125,
|
| 2472 |
+
"learning_rate": 4.367936375052323e-06,
|
| 2473 |
+
"loss": 1.1748931407928467,
|
| 2474 |
+
"step": 352
|
| 2475 |
+
},
|
| 2476 |
+
{
|
| 2477 |
+
"epoch": 0.1448205128205128,
|
| 2478 |
+
"grad_norm": 3.953125,
|
| 2479 |
+
"learning_rate": 4.3658434491419e-06,
|
| 2480 |
+
"loss": 1.264051914215088,
|
| 2481 |
+
"step": 353
|
| 2482 |
+
},
|
| 2483 |
+
{
|
| 2484 |
+
"epoch": 0.14523076923076922,
|
| 2485 |
+
"grad_norm": 3.546875,
|
| 2486 |
+
"learning_rate": 4.363750523231478e-06,
|
| 2487 |
+
"loss": 0.9747573733329773,
|
| 2488 |
+
"step": 354
|
| 2489 |
+
},
|
| 2490 |
+
{
|
| 2491 |
+
"epoch": 0.14564102564102563,
|
| 2492 |
+
"grad_norm": 4.21875,
|
| 2493 |
+
"learning_rate": 4.3616575973210555e-06,
|
| 2494 |
+
"loss": 1.2768664360046387,
|
| 2495 |
+
"step": 355
|
| 2496 |
+
},
|
| 2497 |
+
{
|
| 2498 |
+
"epoch": 0.14605128205128204,
|
| 2499 |
+
"grad_norm": 3.703125,
|
| 2500 |
+
"learning_rate": 4.359564671410632e-06,
|
| 2501 |
+
"loss": 1.0438206195831299,
|
| 2502 |
+
"step": 356
|
| 2503 |
+
},
|
| 2504 |
+
{
|
| 2505 |
+
"epoch": 0.14646153846153845,
|
| 2506 |
+
"grad_norm": 3.859375,
|
| 2507 |
+
"learning_rate": 4.35747174550021e-06,
|
| 2508 |
+
"loss": 1.1907073259353638,
|
| 2509 |
+
"step": 357
|
| 2510 |
+
},
|
| 2511 |
+
{
|
| 2512 |
+
"epoch": 0.14687179487179486,
|
| 2513 |
+
"grad_norm": 4.53125,
|
| 2514 |
+
"learning_rate": 4.355378819589787e-06,
|
| 2515 |
+
"loss": 1.0708136558532715,
|
| 2516 |
+
"step": 358
|
| 2517 |
+
},
|
| 2518 |
+
{
|
| 2519 |
+
"epoch": 0.14728205128205127,
|
| 2520 |
+
"grad_norm": 4.21875,
|
| 2521 |
+
"learning_rate": 4.353285893679364e-06,
|
| 2522 |
+
"loss": 1.3200894594192505,
|
| 2523 |
+
"step": 359
|
| 2524 |
+
},
|
| 2525 |
+
{
|
| 2526 |
+
"epoch": 0.1476923076923077,
|
| 2527 |
+
"grad_norm": 4.25,
|
| 2528 |
+
"learning_rate": 4.351192967768941e-06,
|
| 2529 |
+
"loss": 1.346101999282837,
|
| 2530 |
+
"step": 360
|
| 2531 |
+
},
|
| 2532 |
+
{
|
| 2533 |
+
"epoch": 0.14810256410256412,
|
| 2534 |
+
"grad_norm": 3.828125,
|
| 2535 |
+
"learning_rate": 4.349100041858518e-06,
|
| 2536 |
+
"loss": 1.2198647260665894,
|
| 2537 |
+
"step": 361
|
| 2538 |
+
},
|
| 2539 |
+
{
|
| 2540 |
+
"epoch": 0.14851282051282053,
|
| 2541 |
+
"grad_norm": 3.84375,
|
| 2542 |
+
"learning_rate": 4.3470071159480955e-06,
|
| 2543 |
+
"loss": 1.0523555278778076,
|
| 2544 |
+
"step": 362
|
| 2545 |
+
},
|
| 2546 |
+
{
|
| 2547 |
+
"epoch": 0.14892307692307694,
|
| 2548 |
+
"grad_norm": 4.6875,
|
| 2549 |
+
"learning_rate": 4.344914190037673e-06,
|
| 2550 |
+
"loss": 1.2775603532791138,
|
| 2551 |
+
"step": 363
|
| 2552 |
+
},
|
| 2553 |
+
{
|
| 2554 |
+
"epoch": 0.14933333333333335,
|
| 2555 |
+
"grad_norm": 4.03125,
|
| 2556 |
+
"learning_rate": 4.342821264127251e-06,
|
| 2557 |
+
"loss": 1.0812149047851562,
|
| 2558 |
+
"step": 364
|
| 2559 |
+
},
|
| 2560 |
+
{
|
| 2561 |
+
"epoch": 0.14974358974358976,
|
| 2562 |
+
"grad_norm": 3.984375,
|
| 2563 |
+
"learning_rate": 4.340728338216828e-06,
|
| 2564 |
+
"loss": 1.12788724899292,
|
| 2565 |
+
"step": 365
|
| 2566 |
+
},
|
| 2567 |
+
{
|
| 2568 |
+
"epoch": 0.15015384615384617,
|
| 2569 |
+
"grad_norm": 4.3125,
|
| 2570 |
+
"learning_rate": 4.338635412306404e-06,
|
| 2571 |
+
"loss": 1.2764321565628052,
|
| 2572 |
+
"step": 366
|
| 2573 |
+
},
|
| 2574 |
+
{
|
| 2575 |
+
"epoch": 0.15056410256410258,
|
| 2576 |
+
"grad_norm": 4.03125,
|
| 2577 |
+
"learning_rate": 4.336542486395982e-06,
|
| 2578 |
+
"loss": 1.0494221448898315,
|
| 2579 |
+
"step": 367
|
| 2580 |
+
},
|
| 2581 |
+
{
|
| 2582 |
+
"epoch": 0.15097435897435899,
|
| 2583 |
+
"grad_norm": 4.4375,
|
| 2584 |
+
"learning_rate": 4.334449560485559e-06,
|
| 2585 |
+
"loss": 1.1716052293777466,
|
| 2586 |
+
"step": 368
|
| 2587 |
+
},
|
| 2588 |
+
{
|
| 2589 |
+
"epoch": 0.1513846153846154,
|
| 2590 |
+
"grad_norm": 4.25,
|
| 2591 |
+
"learning_rate": 4.3323566345751365e-06,
|
| 2592 |
+
"loss": 1.3589006662368774,
|
| 2593 |
+
"step": 369
|
| 2594 |
+
},
|
| 2595 |
+
{
|
| 2596 |
+
"epoch": 0.1517948717948718,
|
| 2597 |
+
"grad_norm": 4.125,
|
| 2598 |
+
"learning_rate": 4.330263708664713e-06,
|
| 2599 |
+
"loss": 1.306352138519287,
|
| 2600 |
+
"step": 370
|
| 2601 |
+
},
|
| 2602 |
+
{
|
| 2603 |
+
"epoch": 0.15220512820512822,
|
| 2604 |
+
"grad_norm": 4.375,
|
| 2605 |
+
"learning_rate": 4.328170782754291e-06,
|
| 2606 |
+
"loss": 1.2067084312438965,
|
| 2607 |
+
"step": 371
|
| 2608 |
+
},
|
| 2609 |
+
{
|
| 2610 |
+
"epoch": 0.15261538461538462,
|
| 2611 |
+
"grad_norm": 4.46875,
|
| 2612 |
+
"learning_rate": 4.3260778568438685e-06,
|
| 2613 |
+
"loss": 1.0758178234100342,
|
| 2614 |
+
"step": 372
|
| 2615 |
+
},
|
| 2616 |
+
{
|
| 2617 |
+
"epoch": 0.15302564102564103,
|
| 2618 |
+
"grad_norm": 3.90625,
|
| 2619 |
+
"learning_rate": 4.323984930933445e-06,
|
| 2620 |
+
"loss": 1.152404546737671,
|
| 2621 |
+
"step": 373
|
| 2622 |
+
},
|
| 2623 |
+
{
|
| 2624 |
+
"epoch": 0.15343589743589744,
|
| 2625 |
+
"grad_norm": 3.921875,
|
| 2626 |
+
"learning_rate": 4.321892005023022e-06,
|
| 2627 |
+
"loss": 1.3583828210830688,
|
| 2628 |
+
"step": 374
|
| 2629 |
+
},
|
| 2630 |
+
{
|
| 2631 |
+
"epoch": 0.15384615384615385,
|
| 2632 |
+
"grad_norm": 4.15625,
|
| 2633 |
+
"learning_rate": 4.3197990791126e-06,
|
| 2634 |
+
"loss": 1.3320859670639038,
|
| 2635 |
+
"step": 375
|
| 2636 |
+
},
|
| 2637 |
+
{
|
| 2638 |
+
"epoch": 0.15425641025641026,
|
| 2639 |
+
"grad_norm": 4.125,
|
| 2640 |
+
"learning_rate": 4.3177061532021766e-06,
|
| 2641 |
+
"loss": 1.216289758682251,
|
| 2642 |
+
"step": 376
|
| 2643 |
+
},
|
| 2644 |
+
{
|
| 2645 |
+
"epoch": 0.15466666666666667,
|
| 2646 |
+
"grad_norm": 4.0,
|
| 2647 |
+
"learning_rate": 4.315613227291754e-06,
|
| 2648 |
+
"loss": 1.2127735614776611,
|
| 2649 |
+
"step": 377
|
| 2650 |
+
},
|
| 2651 |
+
{
|
| 2652 |
+
"epoch": 0.15507692307692308,
|
| 2653 |
+
"grad_norm": 3.78125,
|
| 2654 |
+
"learning_rate": 4.313520301381332e-06,
|
| 2655 |
+
"loss": 1.2302391529083252,
|
| 2656 |
+
"step": 378
|
| 2657 |
+
},
|
| 2658 |
+
{
|
| 2659 |
+
"epoch": 0.1554871794871795,
|
| 2660 |
+
"grad_norm": 4.03125,
|
| 2661 |
+
"learning_rate": 4.311427375470909e-06,
|
| 2662 |
+
"loss": 1.150987982749939,
|
| 2663 |
+
"step": 379
|
| 2664 |
+
},
|
| 2665 |
+
{
|
| 2666 |
+
"epoch": 0.1558974358974359,
|
| 2667 |
+
"grad_norm": 3.890625,
|
| 2668 |
+
"learning_rate": 4.309334449560486e-06,
|
| 2669 |
+
"loss": 1.2136256694793701,
|
| 2670 |
+
"step": 380
|
| 2671 |
+
},
|
| 2672 |
+
{
|
| 2673 |
+
"epoch": 0.15630769230769231,
|
| 2674 |
+
"grad_norm": 3.5,
|
| 2675 |
+
"learning_rate": 4.307241523650063e-06,
|
| 2676 |
+
"loss": 1.117332100868225,
|
| 2677 |
+
"step": 381
|
| 2678 |
+
},
|
| 2679 |
+
{
|
| 2680 |
+
"epoch": 0.15671794871794872,
|
| 2681 |
+
"grad_norm": 4.125,
|
| 2682 |
+
"learning_rate": 4.30514859773964e-06,
|
| 2683 |
+
"loss": 1.1436775922775269,
|
| 2684 |
+
"step": 382
|
| 2685 |
+
},
|
| 2686 |
+
{
|
| 2687 |
+
"epoch": 0.15712820512820513,
|
| 2688 |
+
"grad_norm": 3.921875,
|
| 2689 |
+
"learning_rate": 4.3030556718292175e-06,
|
| 2690 |
+
"loss": 1.3296345472335815,
|
| 2691 |
+
"step": 383
|
| 2692 |
+
},
|
| 2693 |
+
{
|
| 2694 |
+
"epoch": 0.15753846153846154,
|
| 2695 |
+
"grad_norm": 3.8125,
|
| 2696 |
+
"learning_rate": 4.300962745918795e-06,
|
| 2697 |
+
"loss": 1.145194411277771,
|
| 2698 |
+
"step": 384
|
| 2699 |
+
},
|
| 2700 |
+
{
|
| 2701 |
+
"epoch": 0.15794871794871795,
|
| 2702 |
+
"grad_norm": 3.859375,
|
| 2703 |
+
"learning_rate": 4.298869820008372e-06,
|
| 2704 |
+
"loss": 1.1882728338241577,
|
| 2705 |
+
"step": 385
|
| 2706 |
+
},
|
| 2707 |
+
{
|
| 2708 |
+
"epoch": 0.15835897435897436,
|
| 2709 |
+
"grad_norm": 3.90625,
|
| 2710 |
+
"learning_rate": 4.2967768940979495e-06,
|
| 2711 |
+
"loss": 1.3066014051437378,
|
| 2712 |
+
"step": 386
|
| 2713 |
+
},
|
| 2714 |
+
{
|
| 2715 |
+
"epoch": 0.15876923076923077,
|
| 2716 |
+
"grad_norm": 4.0625,
|
| 2717 |
+
"learning_rate": 4.294683968187526e-06,
|
| 2718 |
+
"loss": 1.3157684803009033,
|
| 2719 |
+
"step": 387
|
| 2720 |
+
},
|
| 2721 |
+
{
|
| 2722 |
+
"epoch": 0.15917948717948718,
|
| 2723 |
+
"grad_norm": 3.90625,
|
| 2724 |
+
"learning_rate": 4.292591042277104e-06,
|
| 2725 |
+
"loss": 1.2764596939086914,
|
| 2726 |
+
"step": 388
|
| 2727 |
+
},
|
| 2728 |
+
{
|
| 2729 |
+
"epoch": 0.1595897435897436,
|
| 2730 |
+
"grad_norm": 3.84375,
|
| 2731 |
+
"learning_rate": 4.290498116366681e-06,
|
| 2732 |
+
"loss": 1.2815313339233398,
|
| 2733 |
+
"step": 389
|
| 2734 |
+
},
|
| 2735 |
+
{
|
| 2736 |
+
"epoch": 0.16,
|
| 2737 |
+
"grad_norm": 4.3125,
|
| 2738 |
+
"learning_rate": 4.288405190456258e-06,
|
| 2739 |
+
"loss": 1.224855899810791,
|
| 2740 |
+
"step": 390
|
| 2741 |
+
},
|
| 2742 |
+
{
|
| 2743 |
+
"epoch": 0.1604102564102564,
|
| 2744 |
+
"grad_norm": 4.21875,
|
| 2745 |
+
"learning_rate": 4.286312264545835e-06,
|
| 2746 |
+
"loss": 1.2085756063461304,
|
| 2747 |
+
"step": 391
|
| 2748 |
+
},
|
| 2749 |
+
{
|
| 2750 |
+
"epoch": 0.16082051282051282,
|
| 2751 |
+
"grad_norm": 3.875,
|
| 2752 |
+
"learning_rate": 4.284219338635413e-06,
|
| 2753 |
+
"loss": 1.1893349885940552,
|
| 2754 |
+
"step": 392
|
| 2755 |
+
},
|
| 2756 |
+
{
|
| 2757 |
+
"epoch": 0.16123076923076923,
|
| 2758 |
+
"grad_norm": 3.984375,
|
| 2759 |
+
"learning_rate": 4.2821264127249905e-06,
|
| 2760 |
+
"loss": 1.3876506090164185,
|
| 2761 |
+
"step": 393
|
| 2762 |
+
},
|
| 2763 |
+
{
|
| 2764 |
+
"epoch": 0.16164102564102564,
|
| 2765 |
+
"grad_norm": 4.1875,
|
| 2766 |
+
"learning_rate": 4.280033486814567e-06,
|
| 2767 |
+
"loss": 1.0416761636734009,
|
| 2768 |
+
"step": 394
|
| 2769 |
+
},
|
| 2770 |
+
{
|
| 2771 |
+
"epoch": 0.16205128205128205,
|
| 2772 |
+
"grad_norm": 4.0625,
|
| 2773 |
+
"learning_rate": 4.277940560904144e-06,
|
| 2774 |
+
"loss": 1.170915126800537,
|
| 2775 |
+
"step": 395
|
| 2776 |
+
},
|
| 2777 |
+
{
|
| 2778 |
+
"epoch": 0.16246153846153846,
|
| 2779 |
+
"grad_norm": 4.03125,
|
| 2780 |
+
"learning_rate": 4.275847634993722e-06,
|
| 2781 |
+
"loss": 1.1451181173324585,
|
| 2782 |
+
"step": 396
|
| 2783 |
+
},
|
| 2784 |
+
{
|
| 2785 |
+
"epoch": 0.16287179487179487,
|
| 2786 |
+
"grad_norm": 4.3125,
|
| 2787 |
+
"learning_rate": 4.2737547090832985e-06,
|
| 2788 |
+
"loss": 1.3535075187683105,
|
| 2789 |
+
"step": 397
|
| 2790 |
+
},
|
| 2791 |
+
{
|
| 2792 |
+
"epoch": 0.16328205128205128,
|
| 2793 |
+
"grad_norm": 4.84375,
|
| 2794 |
+
"learning_rate": 4.271661783172876e-06,
|
| 2795 |
+
"loss": 1.2866363525390625,
|
| 2796 |
+
"step": 398
|
| 2797 |
+
},
|
| 2798 |
+
{
|
| 2799 |
+
"epoch": 0.1636923076923077,
|
| 2800 |
+
"grad_norm": 4.46875,
|
| 2801 |
+
"learning_rate": 4.269568857262453e-06,
|
| 2802 |
+
"loss": 1.2528660297393799,
|
| 2803 |
+
"step": 399
|
| 2804 |
+
},
|
| 2805 |
+
{
|
| 2806 |
+
"epoch": 0.1641025641025641,
|
| 2807 |
+
"grad_norm": 3.9375,
|
| 2808 |
+
"learning_rate": 4.2674759313520305e-06,
|
| 2809 |
+
"loss": 1.2751768827438354,
|
| 2810 |
+
"step": 400
|
| 2811 |
+
},
|
| 2812 |
+
{
|
| 2813 |
+
"epoch": 0.1645128205128205,
|
| 2814 |
+
"grad_norm": 3.984375,
|
| 2815 |
+
"learning_rate": 4.265383005441608e-06,
|
| 2816 |
+
"loss": 1.1577333211898804,
|
| 2817 |
+
"step": 401
|
| 2818 |
+
},
|
| 2819 |
+
{
|
| 2820 |
+
"epoch": 0.16492307692307692,
|
| 2821 |
+
"grad_norm": 4.46875,
|
| 2822 |
+
"learning_rate": 4.263290079531185e-06,
|
| 2823 |
+
"loss": 1.2090933322906494,
|
| 2824 |
+
"step": 402
|
| 2825 |
+
},
|
| 2826 |
+
{
|
| 2827 |
+
"epoch": 0.16533333333333333,
|
| 2828 |
+
"grad_norm": 4.46875,
|
| 2829 |
+
"learning_rate": 4.261197153620762e-06,
|
| 2830 |
+
"loss": 1.135783314704895,
|
| 2831 |
+
"step": 403
|
| 2832 |
+
},
|
| 2833 |
+
{
|
| 2834 |
+
"epoch": 0.16574358974358974,
|
| 2835 |
+
"grad_norm": 3.71875,
|
| 2836 |
+
"learning_rate": 4.259104227710339e-06,
|
| 2837 |
+
"loss": 1.0493744611740112,
|
| 2838 |
+
"step": 404
|
| 2839 |
+
},
|
| 2840 |
+
{
|
| 2841 |
+
"epoch": 0.16615384615384615,
|
| 2842 |
+
"grad_norm": 3.96875,
|
| 2843 |
+
"learning_rate": 4.257011301799916e-06,
|
| 2844 |
+
"loss": 1.1365779638290405,
|
| 2845 |
+
"step": 405
|
| 2846 |
+
},
|
| 2847 |
+
{
|
| 2848 |
+
"epoch": 0.16656410256410256,
|
| 2849 |
+
"grad_norm": 3.625,
|
| 2850 |
+
"learning_rate": 4.254918375889494e-06,
|
| 2851 |
+
"loss": 1.1713597774505615,
|
| 2852 |
+
"step": 406
|
| 2853 |
+
},
|
| 2854 |
+
{
|
| 2855 |
+
"epoch": 0.16697435897435897,
|
| 2856 |
+
"grad_norm": 4.03125,
|
| 2857 |
+
"learning_rate": 4.2528254499790715e-06,
|
| 2858 |
+
"loss": 1.244136095046997,
|
| 2859 |
+
"step": 407
|
| 2860 |
+
},
|
| 2861 |
+
{
|
| 2862 |
+
"epoch": 0.16738461538461538,
|
| 2863 |
+
"grad_norm": 4.125,
|
| 2864 |
+
"learning_rate": 4.250732524068648e-06,
|
| 2865 |
+
"loss": 1.3147552013397217,
|
| 2866 |
+
"step": 408
|
| 2867 |
+
},
|
| 2868 |
+
{
|
| 2869 |
+
"epoch": 0.1677948717948718,
|
| 2870 |
+
"grad_norm": 4.9375,
|
| 2871 |
+
"learning_rate": 4.248639598158226e-06,
|
| 2872 |
+
"loss": 1.2067335844039917,
|
| 2873 |
+
"step": 409
|
| 2874 |
+
},
|
| 2875 |
+
{
|
| 2876 |
+
"epoch": 0.1682051282051282,
|
| 2877 |
+
"grad_norm": 4.03125,
|
| 2878 |
+
"learning_rate": 4.246546672247803e-06,
|
| 2879 |
+
"loss": 1.2096567153930664,
|
| 2880 |
+
"step": 410
|
| 2881 |
+
},
|
| 2882 |
+
{
|
| 2883 |
+
"epoch": 0.1686153846153846,
|
| 2884 |
+
"grad_norm": 4.03125,
|
| 2885 |
+
"learning_rate": 4.24445374633738e-06,
|
| 2886 |
+
"loss": 1.2628179788589478,
|
| 2887 |
+
"step": 411
|
| 2888 |
+
},
|
| 2889 |
+
{
|
| 2890 |
+
"epoch": 0.16902564102564102,
|
| 2891 |
+
"grad_norm": 4.15625,
|
| 2892 |
+
"learning_rate": 4.242360820426957e-06,
|
| 2893 |
+
"loss": 1.3003451824188232,
|
| 2894 |
+
"step": 412
|
| 2895 |
+
},
|
| 2896 |
+
{
|
| 2897 |
+
"epoch": 0.16943589743589743,
|
| 2898 |
+
"grad_norm": 3.921875,
|
| 2899 |
+
"learning_rate": 4.240267894516534e-06,
|
| 2900 |
+
"loss": 1.3254220485687256,
|
| 2901 |
+
"step": 413
|
| 2902 |
+
},
|
| 2903 |
+
{
|
| 2904 |
+
"epoch": 0.16984615384615384,
|
| 2905 |
+
"grad_norm": 4.21875,
|
| 2906 |
+
"learning_rate": 4.2381749686061115e-06,
|
| 2907 |
+
"loss": 1.1979105472564697,
|
| 2908 |
+
"step": 414
|
| 2909 |
+
},
|
| 2910 |
+
{
|
| 2911 |
+
"epoch": 0.17025641025641025,
|
| 2912 |
+
"grad_norm": 3.953125,
|
| 2913 |
+
"learning_rate": 4.236082042695689e-06,
|
| 2914 |
+
"loss": 1.2148278951644897,
|
| 2915 |
+
"step": 415
|
| 2916 |
+
},
|
| 2917 |
+
{
|
| 2918 |
+
"epoch": 0.17066666666666666,
|
| 2919 |
+
"grad_norm": 4.25,
|
| 2920 |
+
"learning_rate": 4.233989116785266e-06,
|
| 2921 |
+
"loss": 1.243774175643921,
|
| 2922 |
+
"step": 416
|
| 2923 |
+
},
|
| 2924 |
+
{
|
| 2925 |
+
"epoch": 0.17107692307692307,
|
| 2926 |
+
"grad_norm": 4.125,
|
| 2927 |
+
"learning_rate": 4.231896190874844e-06,
|
| 2928 |
+
"loss": 1.3194231986999512,
|
| 2929 |
+
"step": 417
|
| 2930 |
+
},
|
| 2931 |
+
{
|
| 2932 |
+
"epoch": 0.17148717948717948,
|
| 2933 |
+
"grad_norm": 3.875,
|
| 2934 |
+
"learning_rate": 4.22980326496442e-06,
|
| 2935 |
+
"loss": 1.1891469955444336,
|
| 2936 |
+
"step": 418
|
| 2937 |
+
},
|
| 2938 |
+
{
|
| 2939 |
+
"epoch": 0.1718974358974359,
|
| 2940 |
+
"grad_norm": 4.28125,
|
| 2941 |
+
"learning_rate": 4.227710339053998e-06,
|
| 2942 |
+
"loss": 1.2503719329833984,
|
| 2943 |
+
"step": 419
|
| 2944 |
+
},
|
| 2945 |
+
{
|
| 2946 |
+
"epoch": 0.1723076923076923,
|
| 2947 |
+
"grad_norm": 4.1875,
|
| 2948 |
+
"learning_rate": 4.225617413143575e-06,
|
| 2949 |
+
"loss": 1.2555310726165771,
|
| 2950 |
+
"step": 420
|
| 2951 |
+
},
|
| 2952 |
+
{
|
| 2953 |
+
"epoch": 0.1727179487179487,
|
| 2954 |
+
"grad_norm": 4.1875,
|
| 2955 |
+
"learning_rate": 4.2235244872331525e-06,
|
| 2956 |
+
"loss": 1.2581183910369873,
|
| 2957 |
+
"step": 421
|
| 2958 |
+
},
|
| 2959 |
+
{
|
| 2960 |
+
"epoch": 0.17312820512820512,
|
| 2961 |
+
"grad_norm": 4.40625,
|
| 2962 |
+
"learning_rate": 4.22143156132273e-06,
|
| 2963 |
+
"loss": 1.2283926010131836,
|
| 2964 |
+
"step": 422
|
| 2965 |
+
},
|
| 2966 |
+
{
|
| 2967 |
+
"epoch": 0.17353846153846153,
|
| 2968 |
+
"grad_norm": 3.84375,
|
| 2969 |
+
"learning_rate": 4.219338635412307e-06,
|
| 2970 |
+
"loss": 1.0920337438583374,
|
| 2971 |
+
"step": 423
|
| 2972 |
+
},
|
| 2973 |
+
{
|
| 2974 |
+
"epoch": 0.17394871794871794,
|
| 2975 |
+
"grad_norm": 4.4375,
|
| 2976 |
+
"learning_rate": 4.217245709501884e-06,
|
| 2977 |
+
"loss": 1.2105703353881836,
|
| 2978 |
+
"step": 424
|
| 2979 |
+
},
|
| 2980 |
+
{
|
| 2981 |
+
"epoch": 0.17435897435897435,
|
| 2982 |
+
"grad_norm": 4.90625,
|
| 2983 |
+
"learning_rate": 4.215152783591461e-06,
|
| 2984 |
+
"loss": 1.1433370113372803,
|
| 2985 |
+
"step": 425
|
| 2986 |
+
},
|
| 2987 |
+
{
|
| 2988 |
+
"epoch": 0.17476923076923076,
|
| 2989 |
+
"grad_norm": 4.375,
|
| 2990 |
+
"learning_rate": 4.213059857681038e-06,
|
| 2991 |
+
"loss": 1.357133150100708,
|
| 2992 |
+
"step": 426
|
| 2993 |
+
},
|
| 2994 |
+
{
|
| 2995 |
+
"epoch": 0.17517948717948717,
|
| 2996 |
+
"grad_norm": 4.375,
|
| 2997 |
+
"learning_rate": 4.210966931770616e-06,
|
| 2998 |
+
"loss": 1.1646921634674072,
|
| 2999 |
+
"step": 427
|
| 3000 |
+
},
|
| 3001 |
+
{
|
| 3002 |
+
"epoch": 0.17558974358974358,
|
| 3003 |
+
"grad_norm": 4.03125,
|
| 3004 |
+
"learning_rate": 4.2088740058601925e-06,
|
| 3005 |
+
"loss": 1.2933828830718994,
|
| 3006 |
+
"step": 428
|
| 3007 |
+
},
|
| 3008 |
+
{
|
| 3009 |
+
"epoch": 0.176,
|
| 3010 |
+
"grad_norm": 4.28125,
|
| 3011 |
+
"learning_rate": 4.20678107994977e-06,
|
| 3012 |
+
"loss": 1.3107999563217163,
|
| 3013 |
+
"step": 429
|
| 3014 |
+
},
|
| 3015 |
+
{
|
| 3016 |
+
"epoch": 0.1764102564102564,
|
| 3017 |
+
"grad_norm": 4.28125,
|
| 3018 |
+
"learning_rate": 4.204688154039348e-06,
|
| 3019 |
+
"loss": 1.193358302116394,
|
| 3020 |
+
"step": 430
|
| 3021 |
+
},
|
| 3022 |
+
{
|
| 3023 |
+
"epoch": 0.1768205128205128,
|
| 3024 |
+
"grad_norm": 3.40625,
|
| 3025 |
+
"learning_rate": 4.202595228128925e-06,
|
| 3026 |
+
"loss": 1.2135393619537354,
|
| 3027 |
+
"step": 431
|
| 3028 |
+
},
|
| 3029 |
+
{
|
| 3030 |
+
"epoch": 0.17723076923076922,
|
| 3031 |
+
"grad_norm": 4.34375,
|
| 3032 |
+
"learning_rate": 4.200502302218502e-06,
|
| 3033 |
+
"loss": 1.4007718563079834,
|
| 3034 |
+
"step": 432
|
| 3035 |
+
},
|
| 3036 |
+
{
|
| 3037 |
+
"epoch": 0.17764102564102563,
|
| 3038 |
+
"grad_norm": 4.4375,
|
| 3039 |
+
"learning_rate": 4.198409376308079e-06,
|
| 3040 |
+
"loss": 1.468827724456787,
|
| 3041 |
+
"step": 433
|
| 3042 |
+
},
|
| 3043 |
+
{
|
| 3044 |
+
"epoch": 0.17805128205128204,
|
| 3045 |
+
"grad_norm": 3.90625,
|
| 3046 |
+
"learning_rate": 4.196316450397656e-06,
|
| 3047 |
+
"loss": 1.1743221282958984,
|
| 3048 |
+
"step": 434
|
| 3049 |
+
},
|
| 3050 |
+
{
|
| 3051 |
+
"epoch": 0.17846153846153845,
|
| 3052 |
+
"grad_norm": 3.96875,
|
| 3053 |
+
"learning_rate": 4.1942235244872335e-06,
|
| 3054 |
+
"loss": 1.2305328845977783,
|
| 3055 |
+
"step": 435
|
| 3056 |
+
},
|
| 3057 |
+
{
|
| 3058 |
+
"epoch": 0.17887179487179486,
|
| 3059 |
+
"grad_norm": 3.734375,
|
| 3060 |
+
"learning_rate": 4.192130598576811e-06,
|
| 3061 |
+
"loss": 1.295001745223999,
|
| 3062 |
+
"step": 436
|
| 3063 |
+
},
|
| 3064 |
+
{
|
| 3065 |
+
"epoch": 0.17928205128205127,
|
| 3066 |
+
"grad_norm": 4.09375,
|
| 3067 |
+
"learning_rate": 4.190037672666388e-06,
|
| 3068 |
+
"loss": 1.2051297426223755,
|
| 3069 |
+
"step": 437
|
| 3070 |
+
},
|
| 3071 |
+
{
|
| 3072 |
+
"epoch": 0.17969230769230768,
|
| 3073 |
+
"grad_norm": 4.21875,
|
| 3074 |
+
"learning_rate": 4.1879447467559655e-06,
|
| 3075 |
+
"loss": 1.2712100744247437,
|
| 3076 |
+
"step": 438
|
| 3077 |
+
},
|
| 3078 |
+
{
|
| 3079 |
+
"epoch": 0.1801025641025641,
|
| 3080 |
+
"grad_norm": 3.890625,
|
| 3081 |
+
"learning_rate": 4.185851820845542e-06,
|
| 3082 |
+
"loss": 1.2391339540481567,
|
| 3083 |
+
"step": 439
|
| 3084 |
+
},
|
| 3085 |
+
{
|
| 3086 |
+
"epoch": 0.18051282051282053,
|
| 3087 |
+
"grad_norm": 4.25,
|
| 3088 |
+
"learning_rate": 4.18375889493512e-06,
|
| 3089 |
+
"loss": 1.2807815074920654,
|
| 3090 |
+
"step": 440
|
| 3091 |
+
},
|
| 3092 |
+
{
|
| 3093 |
+
"epoch": 0.18092307692307694,
|
| 3094 |
+
"grad_norm": 4.21875,
|
| 3095 |
+
"learning_rate": 4.181665969024697e-06,
|
| 3096 |
+
"loss": 1.304931402206421,
|
| 3097 |
+
"step": 441
|
| 3098 |
+
},
|
| 3099 |
+
{
|
| 3100 |
+
"epoch": 0.18133333333333335,
|
| 3101 |
+
"grad_norm": 4.03125,
|
| 3102 |
+
"learning_rate": 4.1795730431142735e-06,
|
| 3103 |
+
"loss": 1.104285478591919,
|
| 3104 |
+
"step": 442
|
| 3105 |
+
},
|
| 3106 |
+
{
|
| 3107 |
+
"epoch": 0.18174358974358976,
|
| 3108 |
+
"grad_norm": 4.28125,
|
| 3109 |
+
"learning_rate": 4.177480117203851e-06,
|
| 3110 |
+
"loss": 1.160555362701416,
|
| 3111 |
+
"step": 443
|
| 3112 |
+
},
|
| 3113 |
+
{
|
| 3114 |
+
"epoch": 0.18215384615384617,
|
| 3115 |
+
"grad_norm": 3.75,
|
| 3116 |
+
"learning_rate": 4.175387191293429e-06,
|
| 3117 |
+
"loss": 1.0915706157684326,
|
| 3118 |
+
"step": 444
|
| 3119 |
+
},
|
| 3120 |
+
{
|
| 3121 |
+
"epoch": 0.18256410256410258,
|
| 3122 |
+
"grad_norm": 3.84375,
|
| 3123 |
+
"learning_rate": 4.173294265383006e-06,
|
| 3124 |
+
"loss": 1.1975188255310059,
|
| 3125 |
+
"step": 445
|
| 3126 |
+
},
|
| 3127 |
+
{
|
| 3128 |
+
"epoch": 0.18297435897435899,
|
| 3129 |
+
"grad_norm": 4.1875,
|
| 3130 |
+
"learning_rate": 4.171201339472583e-06,
|
| 3131 |
+
"loss": 1.2240660190582275,
|
| 3132 |
+
"step": 446
|
| 3133 |
+
},
|
| 3134 |
+
{
|
| 3135 |
+
"epoch": 0.1833846153846154,
|
| 3136 |
+
"grad_norm": 4.25,
|
| 3137 |
+
"learning_rate": 4.16910841356216e-06,
|
| 3138 |
+
"loss": 1.1213234663009644,
|
| 3139 |
+
"step": 447
|
| 3140 |
+
},
|
| 3141 |
+
{
|
| 3142 |
+
"epoch": 0.1837948717948718,
|
| 3143 |
+
"grad_norm": 4.03125,
|
| 3144 |
+
"learning_rate": 4.167015487651738e-06,
|
| 3145 |
+
"loss": 1.037083625793457,
|
| 3146 |
+
"step": 448
|
| 3147 |
+
},
|
| 3148 |
+
{
|
| 3149 |
+
"epoch": 0.18420512820512822,
|
| 3150 |
+
"grad_norm": 4.3125,
|
| 3151 |
+
"learning_rate": 4.1649225617413145e-06,
|
| 3152 |
+
"loss": 1.0700280666351318,
|
| 3153 |
+
"step": 449
|
| 3154 |
+
},
|
| 3155 |
+
{
|
| 3156 |
+
"epoch": 0.18461538461538463,
|
| 3157 |
+
"grad_norm": 4.25,
|
| 3158 |
+
"learning_rate": 4.162829635830892e-06,
|
| 3159 |
+
"loss": 1.3873844146728516,
|
| 3160 |
+
"step": 450
|
| 3161 |
+
},
|
| 3162 |
+
{
|
| 3163 |
+
"epoch": 0.18502564102564104,
|
| 3164 |
+
"grad_norm": 3.6875,
|
| 3165 |
+
"learning_rate": 4.160736709920469e-06,
|
| 3166 |
+
"loss": 1.221057653427124,
|
| 3167 |
+
"step": 451
|
| 3168 |
+
},
|
| 3169 |
+
{
|
| 3170 |
+
"epoch": 0.18543589743589745,
|
| 3171 |
+
"grad_norm": 3.96875,
|
| 3172 |
+
"learning_rate": 4.1586437840100465e-06,
|
| 3173 |
+
"loss": 1.2247095108032227,
|
| 3174 |
+
"step": 452
|
| 3175 |
+
},
|
| 3176 |
+
{
|
| 3177 |
+
"epoch": 0.18584615384615386,
|
| 3178 |
+
"grad_norm": 3.75,
|
| 3179 |
+
"learning_rate": 4.156550858099624e-06,
|
| 3180 |
+
"loss": 1.1366608142852783,
|
| 3181 |
+
"step": 453
|
| 3182 |
+
},
|
| 3183 |
+
{
|
| 3184 |
+
"epoch": 0.18625641025641027,
|
| 3185 |
+
"grad_norm": 3.953125,
|
| 3186 |
+
"learning_rate": 4.154457932189201e-06,
|
| 3187 |
+
"loss": 1.1160281896591187,
|
| 3188 |
+
"step": 454
|
| 3189 |
+
},
|
| 3190 |
+
{
|
| 3191 |
+
"epoch": 0.18666666666666668,
|
| 3192 |
+
"grad_norm": 4.59375,
|
| 3193 |
+
"learning_rate": 4.152365006278778e-06,
|
| 3194 |
+
"loss": 1.3578596115112305,
|
| 3195 |
+
"step": 455
|
| 3196 |
+
},
|
| 3197 |
+
{
|
| 3198 |
+
"epoch": 0.18707692307692309,
|
| 3199 |
+
"grad_norm": 4.09375,
|
| 3200 |
+
"learning_rate": 4.150272080368355e-06,
|
| 3201 |
+
"loss": 1.338385820388794,
|
| 3202 |
+
"step": 456
|
| 3203 |
+
},
|
| 3204 |
+
{
|
| 3205 |
+
"epoch": 0.1874871794871795,
|
| 3206 |
+
"grad_norm": 4.03125,
|
| 3207 |
+
"learning_rate": 4.148179154457932e-06,
|
| 3208 |
+
"loss": 1.2964720726013184,
|
| 3209 |
+
"step": 457
|
| 3210 |
+
},
|
| 3211 |
+
{
|
| 3212 |
+
"epoch": 0.1878974358974359,
|
| 3213 |
+
"grad_norm": 3.609375,
|
| 3214 |
+
"learning_rate": 4.14608622854751e-06,
|
| 3215 |
+
"loss": 1.2373145818710327,
|
| 3216 |
+
"step": 458
|
| 3217 |
+
},
|
| 3218 |
+
{
|
| 3219 |
+
"epoch": 0.18830769230769231,
|
| 3220 |
+
"grad_norm": 4.0625,
|
| 3221 |
+
"learning_rate": 4.1439933026370875e-06,
|
| 3222 |
+
"loss": 1.2091903686523438,
|
| 3223 |
+
"step": 459
|
| 3224 |
+
},
|
| 3225 |
+
{
|
| 3226 |
+
"epoch": 0.18871794871794872,
|
| 3227 |
+
"grad_norm": 4.0,
|
| 3228 |
+
"learning_rate": 4.141900376726664e-06,
|
| 3229 |
+
"loss": 1.1269036531448364,
|
| 3230 |
+
"step": 460
|
| 3231 |
+
},
|
| 3232 |
+
{
|
| 3233 |
+
"epoch": 0.18912820512820513,
|
| 3234 |
+
"grad_norm": 3.921875,
|
| 3235 |
+
"learning_rate": 4.139807450816242e-06,
|
| 3236 |
+
"loss": 1.3060054779052734,
|
| 3237 |
+
"step": 461
|
| 3238 |
+
},
|
| 3239 |
+
{
|
| 3240 |
+
"epoch": 0.18953846153846154,
|
| 3241 |
+
"grad_norm": 4.0625,
|
| 3242 |
+
"learning_rate": 4.137714524905819e-06,
|
| 3243 |
+
"loss": 1.2167160511016846,
|
| 3244 |
+
"step": 462
|
| 3245 |
+
},
|
| 3246 |
+
{
|
| 3247 |
+
"epoch": 0.18994871794871795,
|
| 3248 |
+
"grad_norm": 3.921875,
|
| 3249 |
+
"learning_rate": 4.1356215989953955e-06,
|
| 3250 |
+
"loss": 1.0410640239715576,
|
| 3251 |
+
"step": 463
|
| 3252 |
+
},
|
| 3253 |
+
{
|
| 3254 |
+
"epoch": 0.19035897435897436,
|
| 3255 |
+
"grad_norm": 4.21875,
|
| 3256 |
+
"learning_rate": 4.133528673084973e-06,
|
| 3257 |
+
"loss": 1.257150411605835,
|
| 3258 |
+
"step": 464
|
| 3259 |
+
},
|
| 3260 |
+
{
|
| 3261 |
+
"epoch": 0.19076923076923077,
|
| 3262 |
+
"grad_norm": 4.0625,
|
| 3263 |
+
"learning_rate": 4.13143574717455e-06,
|
| 3264 |
+
"loss": 1.2502400875091553,
|
| 3265 |
+
"step": 465
|
| 3266 |
+
},
|
| 3267 |
+
{
|
| 3268 |
+
"epoch": 0.19117948717948718,
|
| 3269 |
+
"grad_norm": 4.25,
|
| 3270 |
+
"learning_rate": 4.1293428212641275e-06,
|
| 3271 |
+
"loss": 1.2215404510498047,
|
| 3272 |
+
"step": 466
|
| 3273 |
+
},
|
| 3274 |
+
{
|
| 3275 |
+
"epoch": 0.1915897435897436,
|
| 3276 |
+
"grad_norm": 3.65625,
|
| 3277 |
+
"learning_rate": 4.127249895353705e-06,
|
| 3278 |
+
"loss": 1.21022367477417,
|
| 3279 |
+
"step": 467
|
| 3280 |
+
},
|
| 3281 |
+
{
|
| 3282 |
+
"epoch": 0.192,
|
| 3283 |
+
"grad_norm": 4.4375,
|
| 3284 |
+
"learning_rate": 4.125156969443282e-06,
|
| 3285 |
+
"loss": 1.2364320755004883,
|
| 3286 |
+
"step": 468
|
| 3287 |
+
},
|
| 3288 |
+
{
|
| 3289 |
+
"epoch": 0.19241025641025641,
|
| 3290 |
+
"grad_norm": 3.921875,
|
| 3291 |
+
"learning_rate": 4.12306404353286e-06,
|
| 3292 |
+
"loss": 1.285053014755249,
|
| 3293 |
+
"step": 469
|
| 3294 |
+
},
|
| 3295 |
+
{
|
| 3296 |
+
"epoch": 0.19282051282051282,
|
| 3297 |
+
"grad_norm": 3.734375,
|
| 3298 |
+
"learning_rate": 4.120971117622436e-06,
|
| 3299 |
+
"loss": 1.1335256099700928,
|
| 3300 |
+
"step": 470
|
| 3301 |
+
},
|
| 3302 |
+
{
|
| 3303 |
+
"epoch": 0.19323076923076923,
|
| 3304 |
+
"grad_norm": 4.0,
|
| 3305 |
+
"learning_rate": 4.118878191712013e-06,
|
| 3306 |
+
"loss": 1.1958575248718262,
|
| 3307 |
+
"step": 471
|
| 3308 |
+
},
|
| 3309 |
+
{
|
| 3310 |
+
"epoch": 0.19364102564102564,
|
| 3311 |
+
"grad_norm": 3.8125,
|
| 3312 |
+
"learning_rate": 4.116785265801591e-06,
|
| 3313 |
+
"loss": 1.1128159761428833,
|
| 3314 |
+
"step": 472
|
| 3315 |
+
},
|
| 3316 |
+
{
|
| 3317 |
+
"epoch": 0.19405128205128205,
|
| 3318 |
+
"grad_norm": 4.0,
|
| 3319 |
+
"learning_rate": 4.1146923398911685e-06,
|
| 3320 |
+
"loss": 1.3458808660507202,
|
| 3321 |
+
"step": 473
|
| 3322 |
+
},
|
| 3323 |
+
{
|
| 3324 |
+
"epoch": 0.19446153846153846,
|
| 3325 |
+
"grad_norm": 4.1875,
|
| 3326 |
+
"learning_rate": 4.112599413980746e-06,
|
| 3327 |
+
"loss": 1.3128029108047485,
|
| 3328 |
+
"step": 474
|
| 3329 |
+
},
|
| 3330 |
+
{
|
| 3331 |
+
"epoch": 0.19487179487179487,
|
| 3332 |
+
"grad_norm": 4.03125,
|
| 3333 |
+
"learning_rate": 4.110506488070323e-06,
|
| 3334 |
+
"loss": 1.1214417219161987,
|
| 3335 |
+
"step": 475
|
| 3336 |
+
},
|
| 3337 |
+
{
|
| 3338 |
+
"epoch": 0.19528205128205128,
|
| 3339 |
+
"grad_norm": 4.3125,
|
| 3340 |
+
"learning_rate": 4.1084135621599e-06,
|
| 3341 |
+
"loss": 1.3098320960998535,
|
| 3342 |
+
"step": 476
|
| 3343 |
+
},
|
| 3344 |
+
{
|
| 3345 |
+
"epoch": 0.1956923076923077,
|
| 3346 |
+
"grad_norm": 3.703125,
|
| 3347 |
+
"learning_rate": 4.106320636249477e-06,
|
| 3348 |
+
"loss": 1.2281699180603027,
|
| 3349 |
+
"step": 477
|
| 3350 |
+
},
|
| 3351 |
+
{
|
| 3352 |
+
"epoch": 0.1961025641025641,
|
| 3353 |
+
"grad_norm": 4.3125,
|
| 3354 |
+
"learning_rate": 4.104227710339054e-06,
|
| 3355 |
+
"loss": 1.322780966758728,
|
| 3356 |
+
"step": 478
|
| 3357 |
+
},
|
| 3358 |
+
{
|
| 3359 |
+
"epoch": 0.1965128205128205,
|
| 3360 |
+
"grad_norm": 3.71875,
|
| 3361 |
+
"learning_rate": 4.102134784428632e-06,
|
| 3362 |
+
"loss": 1.1614468097686768,
|
| 3363 |
+
"step": 479
|
| 3364 |
+
},
|
| 3365 |
+
{
|
| 3366 |
+
"epoch": 0.19692307692307692,
|
| 3367 |
+
"grad_norm": 4.0,
|
| 3368 |
+
"learning_rate": 4.1000418585182085e-06,
|
| 3369 |
+
"loss": 1.257091999053955,
|
| 3370 |
+
"step": 480
|
| 3371 |
+
},
|
| 3372 |
+
{
|
| 3373 |
+
"epoch": 0.19733333333333333,
|
| 3374 |
+
"grad_norm": 4.03125,
|
| 3375 |
+
"learning_rate": 4.097948932607786e-06,
|
| 3376 |
+
"loss": 1.2248072624206543,
|
| 3377 |
+
"step": 481
|
| 3378 |
+
},
|
| 3379 |
+
{
|
| 3380 |
+
"epoch": 0.19774358974358974,
|
| 3381 |
+
"grad_norm": 3.90625,
|
| 3382 |
+
"learning_rate": 4.095856006697364e-06,
|
| 3383 |
+
"loss": 1.173184871673584,
|
| 3384 |
+
"step": 482
|
| 3385 |
+
},
|
| 3386 |
+
{
|
| 3387 |
+
"epoch": 0.19815384615384615,
|
| 3388 |
+
"grad_norm": 4.1875,
|
| 3389 |
+
"learning_rate": 4.093763080786941e-06,
|
| 3390 |
+
"loss": 1.2033233642578125,
|
| 3391 |
+
"step": 483
|
| 3392 |
+
},
|
| 3393 |
+
{
|
| 3394 |
+
"epoch": 0.19856410256410256,
|
| 3395 |
+
"grad_norm": 3.875,
|
| 3396 |
+
"learning_rate": 4.091670154876517e-06,
|
| 3397 |
+
"loss": 1.1596381664276123,
|
| 3398 |
+
"step": 484
|
| 3399 |
+
},
|
| 3400 |
+
{
|
| 3401 |
+
"epoch": 0.19897435897435897,
|
| 3402 |
+
"grad_norm": 4.15625,
|
| 3403 |
+
"learning_rate": 4.089577228966095e-06,
|
| 3404 |
+
"loss": 1.1989463567733765,
|
| 3405 |
+
"step": 485
|
| 3406 |
+
},
|
| 3407 |
+
{
|
| 3408 |
+
"epoch": 0.19938461538461538,
|
| 3409 |
+
"grad_norm": 4.125,
|
| 3410 |
+
"learning_rate": 4.087484303055672e-06,
|
| 3411 |
+
"loss": 1.2782260179519653,
|
| 3412 |
+
"step": 486
|
| 3413 |
+
},
|
| 3414 |
+
{
|
| 3415 |
+
"epoch": 0.1997948717948718,
|
| 3416 |
+
"grad_norm": 3.921875,
|
| 3417 |
+
"learning_rate": 4.0853913771452495e-06,
|
| 3418 |
+
"loss": 1.2400543689727783,
|
| 3419 |
+
"step": 487
|
| 3420 |
+
},
|
| 3421 |
+
{
|
| 3422 |
+
"epoch": 0.2002051282051282,
|
| 3423 |
+
"grad_norm": 3.859375,
|
| 3424 |
+
"learning_rate": 4.083298451234827e-06,
|
| 3425 |
+
"loss": 1.1157163381576538,
|
| 3426 |
+
"step": 488
|
| 3427 |
+
},
|
| 3428 |
+
{
|
| 3429 |
+
"epoch": 0.2002051282051282,
|
| 3430 |
+
"eval_loss": 1.2125588655471802,
|
| 3431 |
+
"eval_runtime": 12.834,
|
| 3432 |
+
"eval_samples_per_second": 77.918,
|
| 3433 |
+
"eval_steps_per_second": 19.48,
|
| 3434 |
+
"step": 488
|
| 3435 |
+
}
|
| 3436 |
+
],
|
| 3437 |
+
"logging_steps": 1,
|
| 3438 |
+
"max_steps": 2438,
|
| 3439 |
+
"num_input_tokens_seen": 0,
|
| 3440 |
+
"num_train_epochs": 1,
|
| 3441 |
+
"save_steps": 488,
|
| 3442 |
+
"stateful_callbacks": {
|
| 3443 |
+
"TrainerControl": {
|
| 3444 |
+
"args": {
|
| 3445 |
+
"should_epoch_stop": false,
|
| 3446 |
+
"should_evaluate": false,
|
| 3447 |
+
"should_log": false,
|
| 3448 |
+
"should_save": true,
|
| 3449 |
+
"should_training_stop": false
|
| 3450 |
+
},
|
| 3451 |
+
"attributes": {}
|
| 3452 |
+
}
|
| 3453 |
+
},
|
| 3454 |
+
"total_flos": 2.6270889864167424e+16,
|
| 3455 |
+
"train_batch_size": 4,
|
| 3456 |
+
"trial_name": null,
|
| 3457 |
+
"trial_params": null
|
| 3458 |
+
}
|
last-checkpoint/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c85d0c7760119d16a5d1e1c7079a4db56d9d1a5d60b496db06591ca0556bfdf3
|
| 3 |
+
size 5713
|