ray
commited on
Commit
•
5ae38a0
1
Parent(s):
10baf84
add file
Browse files- added_tokens.json +7 -0
- config.json +39 -0
- generation_config.json +6 -0
- pytorch_model.bin +3 -0
- special_tokens_map.json +33 -0
- tokenizer.json +0 -0
- tokenizer_config.json +31 -0
- trainer_state.json +1870 -0
- vocab.json +0 -0
added_tokens.json
ADDED
|
@@ -0,0 +1,7 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"<|assistant|>": 49155,
|
| 3 |
+
"<|prefix_begin|>": 49152,
|
| 4 |
+
"<|prefix_end|>": 49153,
|
| 5 |
+
"<|prompter|>": 49154,
|
| 6 |
+
"<|system|>": 49156
|
| 7 |
+
}
|
config.json
ADDED
|
@@ -0,0 +1,39 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "saved_models/starcoderbase-1b",
|
| 3 |
+
"activation_function": "gelu_pytorch_tanh",
|
| 4 |
+
"architectures": [
|
| 5 |
+
"GPTBigCodeForCausalLM"
|
| 6 |
+
],
|
| 7 |
+
"attention_softmax_in_fp32": true,
|
| 8 |
+
"attn_pdrop": 0.1,
|
| 9 |
+
"bos_token_id": 0,
|
| 10 |
+
"embd_pdrop": 0.1,
|
| 11 |
+
"eos_token_id": 0,
|
| 12 |
+
"inference_runner": 0,
|
| 13 |
+
"initializer_range": 0.02,
|
| 14 |
+
"layer_norm_epsilon": 1e-05,
|
| 15 |
+
"max_batch_size": null,
|
| 16 |
+
"max_sequence_length": null,
|
| 17 |
+
"model_type": "gpt_bigcode",
|
| 18 |
+
"multi_query": true,
|
| 19 |
+
"n_embd": 2048,
|
| 20 |
+
"n_head": 16,
|
| 21 |
+
"n_inner": 8192,
|
| 22 |
+
"n_layer": 24,
|
| 23 |
+
"n_positions": 8192,
|
| 24 |
+
"pad_key_length": true,
|
| 25 |
+
"pre_allocate_kv_cache": false,
|
| 26 |
+
"resid_pdrop": 0.1,
|
| 27 |
+
"scale_attention_softmax_in_fp32": true,
|
| 28 |
+
"scale_attn_weights": true,
|
| 29 |
+
"summary_activation": null,
|
| 30 |
+
"summary_first_dropout": 0.1,
|
| 31 |
+
"summary_proj_to_labels": true,
|
| 32 |
+
"summary_type": "cls_index",
|
| 33 |
+
"summary_use_proj": true,
|
| 34 |
+
"torch_dtype": "float16",
|
| 35 |
+
"transformers_version": "4.30.0.dev0",
|
| 36 |
+
"use_cache": true,
|
| 37 |
+
"validate_runner_input": true,
|
| 38 |
+
"vocab_size": 49168
|
| 39 |
+
}
|
generation_config.json
ADDED
|
@@ -0,0 +1,6 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_from_model_config": true,
|
| 3 |
+
"bos_token_id": 0,
|
| 4 |
+
"eos_token_id": 0,
|
| 5 |
+
"transformers_version": "4.30.0.dev0"
|
| 6 |
+
}
|
pytorch_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:bccad75b61fe5f7666f08ef3d1b983bb0a7ca009c34baefae1d821c3972e40d6
|
| 3 |
+
size 2274516541
|
special_tokens_map.json
ADDED
|
@@ -0,0 +1,33 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"additional_special_tokens": [
|
| 3 |
+
"<issue_comment>",
|
| 4 |
+
"<fim_suffix>",
|
| 5 |
+
"<fim_pad>",
|
| 6 |
+
"<empty_output>",
|
| 7 |
+
"<issue_start>",
|
| 8 |
+
"<commit_msg>",
|
| 9 |
+
"<reponame>",
|
| 10 |
+
"<gh_stars>",
|
| 11 |
+
"<|endoftext|>",
|
| 12 |
+
"<commit_after>",
|
| 13 |
+
"<jupyter_text>",
|
| 14 |
+
"<|prefix_begin|>",
|
| 15 |
+
"<|prefix_end|>",
|
| 16 |
+
"<jupyter_start>",
|
| 17 |
+
"<jupyter_output>",
|
| 18 |
+
"<commit_before>",
|
| 19 |
+
"<|prompter|>",
|
| 20 |
+
"<|assistant|>",
|
| 21 |
+
"<|system|>",
|
| 22 |
+
"<jupyter_code>",
|
| 23 |
+
"<issue_closed>",
|
| 24 |
+
"<fim_middle>",
|
| 25 |
+
"<filename>",
|
| 26 |
+
"<fim_prefix>"
|
| 27 |
+
],
|
| 28 |
+
"bos_token": "<|endoftext|>",
|
| 29 |
+
"eos_token": "<|endoftext|>",
|
| 30 |
+
"pad_token": "<|endoftext|>",
|
| 31 |
+
"sep_token": "<|endoftext|>",
|
| 32 |
+
"unk_token": "<|endoftext|>"
|
| 33 |
+
}
|
tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
tokenizer_config.json
ADDED
|
@@ -0,0 +1,31 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"add_prefix_space": false,
|
| 3 |
+
"additional_special_tokens": [
|
| 4 |
+
"<|endoftext|>",
|
| 5 |
+
"<fim_prefix>",
|
| 6 |
+
"<fim_middle>",
|
| 7 |
+
"<fim_suffix>",
|
| 8 |
+
"<fim_pad>",
|
| 9 |
+
"<filename>",
|
| 10 |
+
"<gh_stars>",
|
| 11 |
+
"<issue_start>",
|
| 12 |
+
"<issue_comment>",
|
| 13 |
+
"<issue_closed>",
|
| 14 |
+
"<jupyter_start>",
|
| 15 |
+
"<jupyter_text>",
|
| 16 |
+
"<jupyter_code>",
|
| 17 |
+
"<jupyter_output>",
|
| 18 |
+
"<empty_output>",
|
| 19 |
+
"<commit_before>",
|
| 20 |
+
"<commit_msg>",
|
| 21 |
+
"<commit_after>",
|
| 22 |
+
"<reponame>"
|
| 23 |
+
],
|
| 24 |
+
"bos_token": "<|endoftext|>",
|
| 25 |
+
"clean_up_tokenization_spaces": true,
|
| 26 |
+
"eos_token": "<|endoftext|>",
|
| 27 |
+
"model_max_length": 1000000000000000019884624838656,
|
| 28 |
+
"tokenizer_class": "GPT2Tokenizer",
|
| 29 |
+
"unk_token": "<|endoftext|>",
|
| 30 |
+
"vocab_size": 49152
|
| 31 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,1870 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 7.550335570469799,
|
| 5 |
+
"global_step": 3000,
|
| 6 |
+
"is_hyper_param_search": false,
|
| 7 |
+
"is_local_process_zero": true,
|
| 8 |
+
"is_world_process_zero": true,
|
| 9 |
+
"log_history": [
|
| 10 |
+
{
|
| 11 |
+
"epoch": 0.03,
|
| 12 |
+
"learning_rate": 2.8169934667141895e-06,
|
| 13 |
+
"loss": 1.2113,
|
| 14 |
+
"step": 10
|
| 15 |
+
},
|
| 16 |
+
{
|
| 17 |
+
"epoch": 0.05,
|
| 18 |
+
"learning_rate": 4.101496404594247e-06,
|
| 19 |
+
"loss": 1.1368,
|
| 20 |
+
"step": 20
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"epoch": 0.08,
|
| 24 |
+
"learning_rate": 4.771212547196624e-06,
|
| 25 |
+
"loss": 1.0866,
|
| 26 |
+
"step": 30
|
| 27 |
+
},
|
| 28 |
+
{
|
| 29 |
+
"epoch": 0.1,
|
| 30 |
+
"learning_rate": 5.227339080223317e-06,
|
| 31 |
+
"loss": 1.0541,
|
| 32 |
+
"step": 40
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"epoch": 0.13,
|
| 36 |
+
"learning_rate": 5.5736595264523914e-06,
|
| 37 |
+
"loss": 1.0277,
|
| 38 |
+
"step": 50
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.15,
|
| 42 |
+
"learning_rate": 5.852916185574972e-06,
|
| 43 |
+
"loss": 1.0367,
|
| 44 |
+
"step": 60
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.18,
|
| 48 |
+
"learning_rate": 6.086916009002755e-06,
|
| 49 |
+
"loss": 1.0106,
|
| 50 |
+
"step": 70
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"epoch": 0.2,
|
| 54 |
+
"learning_rate": 6.288302417241607e-06,
|
| 55 |
+
"loss": 1.0122,
|
| 56 |
+
"step": 80
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"epoch": 0.23,
|
| 60 |
+
"learning_rate": 6.465064175395396e-06,
|
| 61 |
+
"loss": 0.9955,
|
| 62 |
+
"step": 90
|
| 63 |
+
},
|
| 64 |
+
{
|
| 65 |
+
"epoch": 0.25,
|
| 66 |
+
"learning_rate": 6.622572447554151e-06,
|
| 67 |
+
"loss": 0.9957,
|
| 68 |
+
"step": 100
|
| 69 |
+
},
|
| 70 |
+
{
|
| 71 |
+
"epoch": 0.28,
|
| 72 |
+
"learning_rate": 6.7646125922840325e-06,
|
| 73 |
+
"loss": 1.0047,
|
| 74 |
+
"step": 110
|
| 75 |
+
},
|
| 76 |
+
{
|
| 77 |
+
"epoch": 0.3,
|
| 78 |
+
"learning_rate": 6.893952872487206e-06,
|
| 79 |
+
"loss": 1.0056,
|
| 80 |
+
"step": 120
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 0.33,
|
| 84 |
+
"learning_rate": 7.01267906985319e-06,
|
| 85 |
+
"loss": 0.9908,
|
| 86 |
+
"step": 130
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.35,
|
| 90 |
+
"learning_rate": 7.122401890521357e-06,
|
| 91 |
+
"loss": 0.9841,
|
| 92 |
+
"step": 140
|
| 93 |
+
},
|
| 94 |
+
{
|
| 95 |
+
"epoch": 0.38,
|
| 96 |
+
"learning_rate": 7.224391115827254e-06,
|
| 97 |
+
"loss": 0.9865,
|
| 98 |
+
"step": 150
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 0.4,
|
| 102 |
+
"learning_rate": 7.31966550803078e-06,
|
| 103 |
+
"loss": 0.9963,
|
| 104 |
+
"step": 160
|
| 105 |
+
},
|
| 106 |
+
{
|
| 107 |
+
"epoch": 0.43,
|
| 108 |
+
"learning_rate": 7.409054903825279e-06,
|
| 109 |
+
"loss": 0.9818,
|
| 110 |
+
"step": 170
|
| 111 |
+
},
|
| 112 |
+
{
|
| 113 |
+
"epoch": 0.45,
|
| 114 |
+
"learning_rate": 7.493244221206023e-06,
|
| 115 |
+
"loss": 0.9678,
|
| 116 |
+
"step": 180
|
| 117 |
+
},
|
| 118 |
+
{
|
| 119 |
+
"epoch": 0.48,
|
| 120 |
+
"learning_rate": 7.572805355121664e-06,
|
| 121 |
+
"loss": 0.986,
|
| 122 |
+
"step": 190
|
| 123 |
+
},
|
| 124 |
+
{
|
| 125 |
+
"epoch": 0.5,
|
| 126 |
+
"learning_rate": 7.648220753871978e-06,
|
| 127 |
+
"loss": 0.984,
|
| 128 |
+
"step": 200
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.53,
|
| 132 |
+
"learning_rate": 7.719901151523059e-06,
|
| 133 |
+
"loss": 0.9686,
|
| 134 |
+
"step": 210
|
| 135 |
+
},
|
| 136 |
+
{
|
| 137 |
+
"epoch": 0.55,
|
| 138 |
+
"learning_rate": 7.788199112828433e-06,
|
| 139 |
+
"loss": 0.9692,
|
| 140 |
+
"step": 220
|
| 141 |
+
},
|
| 142 |
+
{
|
| 143 |
+
"epoch": 0.58,
|
| 144 |
+
"learning_rate": 7.853419523977077e-06,
|
| 145 |
+
"loss": 0.965,
|
| 146 |
+
"step": 230
|
| 147 |
+
},
|
| 148 |
+
{
|
| 149 |
+
"epoch": 0.6,
|
| 150 |
+
"learning_rate": 7.91582782003368e-06,
|
| 151 |
+
"loss": 0.9816,
|
| 152 |
+
"step": 240
|
| 153 |
+
},
|
| 154 |
+
{
|
| 155 |
+
"epoch": 0.63,
|
| 156 |
+
"learning_rate": 7.975656510865554e-06,
|
| 157 |
+
"loss": 0.9602,
|
| 158 |
+
"step": 250
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 0.65,
|
| 162 |
+
"learning_rate": 8.033110411104317e-06,
|
| 163 |
+
"loss": 0.9757,
|
| 164 |
+
"step": 260
|
| 165 |
+
},
|
| 166 |
+
{
|
| 167 |
+
"epoch": 0.68,
|
| 168 |
+
"learning_rate": 8.088370871215252e-06,
|
| 169 |
+
"loss": 0.9854,
|
| 170 |
+
"step": 270
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 0.7,
|
| 174 |
+
"learning_rate": 8.14159923021483e-06,
|
| 175 |
+
"loss": 0.9651,
|
| 176 |
+
"step": 280
|
| 177 |
+
},
|
| 178 |
+
{
|
| 179 |
+
"epoch": 0.73,
|
| 180 |
+
"learning_rate": 8.192939655779976e-06,
|
| 181 |
+
"loss": 0.9641,
|
| 182 |
+
"step": 290
|
| 183 |
+
},
|
| 184 |
+
{
|
| 185 |
+
"epoch": 0.76,
|
| 186 |
+
"learning_rate": 8.242521497724043e-06,
|
| 187 |
+
"loss": 0.9602,
|
| 188 |
+
"step": 300
|
| 189 |
+
},
|
| 190 |
+
{
|
| 191 |
+
"epoch": 0.78,
|
| 192 |
+
"learning_rate": 8.290461251590623e-06,
|
| 193 |
+
"loss": 0.9496,
|
| 194 |
+
"step": 310
|
| 195 |
+
},
|
| 196 |
+
{
|
| 197 |
+
"epoch": 0.81,
|
| 198 |
+
"learning_rate": 8.336864207392505e-06,
|
| 199 |
+
"loss": 0.9578,
|
| 200 |
+
"step": 320
|
| 201 |
+
},
|
| 202 |
+
{
|
| 203 |
+
"epoch": 0.83,
|
| 204 |
+
"learning_rate": 8.381825842200955e-06,
|
| 205 |
+
"loss": 0.9678,
|
| 206 |
+
"step": 330
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 0.86,
|
| 210 |
+
"learning_rate": 8.425433002904463e-06,
|
| 211 |
+
"loss": 0.9609,
|
| 212 |
+
"step": 340
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 0.88,
|
| 216 |
+
"learning_rate": 8.467764915969581e-06,
|
| 217 |
+
"loss": 0.9628,
|
| 218 |
+
"step": 350
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 0.91,
|
| 222 |
+
"learning_rate": 8.508894053707311e-06,
|
| 223 |
+
"loss": 0.9496,
|
| 224 |
+
"step": 360
|
| 225 |
+
},
|
| 226 |
+
{
|
| 227 |
+
"epoch": 0.93,
|
| 228 |
+
"learning_rate": 8.5488868808403e-06,
|
| 229 |
+
"loss": 0.9577,
|
| 230 |
+
"step": 370
|
| 231 |
+
},
|
| 232 |
+
{
|
| 233 |
+
"epoch": 0.96,
|
| 234 |
+
"learning_rate": 8.587804500685978e-06,
|
| 235 |
+
"loss": 0.9535,
|
| 236 |
+
"step": 380
|
| 237 |
+
},
|
| 238 |
+
{
|
| 239 |
+
"epoch": 0.98,
|
| 240 |
+
"learning_rate": 8.625703216729706e-06,
|
| 241 |
+
"loss": 0.9452,
|
| 242 |
+
"step": 390
|
| 243 |
+
},
|
| 244 |
+
{
|
| 245 |
+
"epoch": 1.01,
|
| 246 |
+
"learning_rate": 8.662635022543719e-06,
|
| 247 |
+
"loss": 0.9341,
|
| 248 |
+
"step": 400
|
| 249 |
+
},
|
| 250 |
+
{
|
| 251 |
+
"epoch": 1.03,
|
| 252 |
+
"learning_rate": 8.698648030750733e-06,
|
| 253 |
+
"loss": 0.9228,
|
| 254 |
+
"step": 410
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 1.06,
|
| 258 |
+
"learning_rate": 8.733786849912527e-06,
|
| 259 |
+
"loss": 0.9315,
|
| 260 |
+
"step": 420
|
| 261 |
+
},
|
| 262 |
+
{
|
| 263 |
+
"epoch": 1.08,
|
| 264 |
+
"learning_rate": 8.76809291675008e-06,
|
| 265 |
+
"loss": 0.9127,
|
| 266 |
+
"step": 430
|
| 267 |
+
},
|
| 268 |
+
{
|
| 269 |
+
"epoch": 1.11,
|
| 270 |
+
"learning_rate": 8.801604789901407e-06,
|
| 271 |
+
"loss": 0.9084,
|
| 272 |
+
"step": 440
|
| 273 |
+
},
|
| 274 |
+
{
|
| 275 |
+
"epoch": 1.13,
|
| 276 |
+
"learning_rate": 8.834358410439789e-06,
|
| 277 |
+
"loss": 0.9157,
|
| 278 |
+
"step": 450
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 1.16,
|
| 282 |
+
"learning_rate": 8.866387333566169e-06,
|
| 283 |
+
"loss": 0.9143,
|
| 284 |
+
"step": 460
|
| 285 |
+
},
|
| 286 |
+
{
|
| 287 |
+
"epoch": 1.18,
|
| 288 |
+
"learning_rate": 8.897722935220375e-06,
|
| 289 |
+
"loss": 0.9221,
|
| 290 |
+
"step": 470
|
| 291 |
+
},
|
| 292 |
+
{
|
| 293 |
+
"epoch": 1.21,
|
| 294 |
+
"learning_rate": 8.928394596800381e-06,
|
| 295 |
+
"loss": 0.911,
|
| 296 |
+
"step": 480
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 1.23,
|
| 300 |
+
"learning_rate": 8.958429870715448e-06,
|
| 301 |
+
"loss": 0.9176,
|
| 302 |
+
"step": 490
|
| 303 |
+
},
|
| 304 |
+
{
|
| 305 |
+
"epoch": 1.26,
|
| 306 |
+
"learning_rate": 8.987854629111107e-06,
|
| 307 |
+
"loss": 0.9122,
|
| 308 |
+
"step": 500
|
| 309 |
+
},
|
| 310 |
+
{
|
| 311 |
+
"epoch": 1.26,
|
| 312 |
+
"eval_code_evol_accuracy": 0.7536964329038056,
|
| 313 |
+
"eval_code_evol_loss": 0.9287109375,
|
| 314 |
+
"eval_code_evol_runtime": 27.3209,
|
| 315 |
+
"eval_code_evol_samples_per_second": 28.22,
|
| 316 |
+
"eval_code_evol_steps_per_second": 3.55,
|
| 317 |
+
"step": 500
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 1.28,
|
| 321 |
+
"learning_rate": 9.016693197777789e-06,
|
| 322 |
+
"loss": 0.9191,
|
| 323 |
+
"step": 510
|
| 324 |
+
},
|
| 325 |
+
{
|
| 326 |
+
"epoch": 1.31,
|
| 327 |
+
"learning_rate": 9.044968476979809e-06,
|
| 328 |
+
"loss": 0.9176,
|
| 329 |
+
"step": 520
|
| 330 |
+
},
|
| 331 |
+
{
|
| 332 |
+
"epoch": 1.33,
|
| 333 |
+
"learning_rate": 9.072702050708489e-06,
|
| 334 |
+
"loss": 0.9235,
|
| 335 |
+
"step": 530
|
| 336 |
+
},
|
| 337 |
+
{
|
| 338 |
+
"epoch": 1.36,
|
| 339 |
+
"learning_rate": 9.099914285665186e-06,
|
| 340 |
+
"loss": 0.899,
|
| 341 |
+
"step": 540
|
| 342 |
+
},
|
| 343 |
+
{
|
| 344 |
+
"epoch": 1.38,
|
| 345 |
+
"learning_rate": 9.126624421111438e-06,
|
| 346 |
+
"loss": 0.9364,
|
| 347 |
+
"step": 550
|
| 348 |
+
},
|
| 349 |
+
{
|
| 350 |
+
"epoch": 1.41,
|
| 351 |
+
"learning_rate": 9.152850650579098e-06,
|
| 352 |
+
"loss": 0.9136,
|
| 353 |
+
"step": 560
|
| 354 |
+
},
|
| 355 |
+
{
|
| 356 |
+
"epoch": 1.43,
|
| 357 |
+
"learning_rate": 9.17861019630969e-06,
|
| 358 |
+
"loss": 0.9204,
|
| 359 |
+
"step": 570
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 1.46,
|
| 363 |
+
"learning_rate": 9.203919377185773e-06,
|
| 364 |
+
"loss": 0.9167,
|
| 365 |
+
"step": 580
|
| 366 |
+
},
|
| 367 |
+
{
|
| 368 |
+
"epoch": 1.48,
|
| 369 |
+
"learning_rate": 9.228793670825383e-06,
|
| 370 |
+
"loss": 0.9092,
|
| 371 |
+
"step": 590
|
| 372 |
+
},
|
| 373 |
+
{
|
| 374 |
+
"epoch": 1.51,
|
| 375 |
+
"learning_rate": 9.253247770431231e-06,
|
| 376 |
+
"loss": 0.9183,
|
| 377 |
+
"step": 600
|
| 378 |
+
},
|
| 379 |
+
{
|
| 380 |
+
"epoch": 1.54,
|
| 381 |
+
"learning_rate": 9.277295636917526e-06,
|
| 382 |
+
"loss": 0.9113,
|
| 383 |
+
"step": 610
|
| 384 |
+
},
|
| 385 |
+
{
|
| 386 |
+
"epoch": 1.56,
|
| 387 |
+
"learning_rate": 9.300950546777473e-06,
|
| 388 |
+
"loss": 0.8988,
|
| 389 |
+
"step": 620
|
| 390 |
+
},
|
| 391 |
+
{
|
| 392 |
+
"epoch": 1.59,
|
| 393 |
+
"learning_rate": 9.324225136102389e-06,
|
| 394 |
+
"loss": 0.9229,
|
| 395 |
+
"step": 630
|
| 396 |
+
},
|
| 397 |
+
{
|
| 398 |
+
"epoch": 1.61,
|
| 399 |
+
"learning_rate": 9.347131441117836e-06,
|
| 400 |
+
"loss": 0.9193,
|
| 401 |
+
"step": 640
|
| 402 |
+
},
|
| 403 |
+
{
|
| 404 |
+
"epoch": 1.64,
|
| 405 |
+
"learning_rate": 9.369680935562337e-06,
|
| 406 |
+
"loss": 0.9085,
|
| 407 |
+
"step": 650
|
| 408 |
+
},
|
| 409 |
+
{
|
| 410 |
+
"epoch": 1.66,
|
| 411 |
+
"learning_rate": 9.39188456519927e-06,
|
| 412 |
+
"loss": 0.9128,
|
| 413 |
+
"step": 660
|
| 414 |
+
},
|
| 415 |
+
{
|
| 416 |
+
"epoch": 1.69,
|
| 417 |
+
"learning_rate": 9.41375277972183e-06,
|
| 418 |
+
"loss": 0.9079,
|
| 419 |
+
"step": 670
|
| 420 |
+
},
|
| 421 |
+
{
|
| 422 |
+
"epoch": 1.71,
|
| 423 |
+
"learning_rate": 9.435295562283815e-06,
|
| 424 |
+
"loss": 0.9097,
|
| 425 |
+
"step": 680
|
| 426 |
+
},
|
| 427 |
+
{
|
| 428 |
+
"epoch": 1.74,
|
| 429 |
+
"learning_rate": 9.456522456865169e-06,
|
| 430 |
+
"loss": 0.9077,
|
| 431 |
+
"step": 690
|
| 432 |
+
},
|
| 433 |
+
{
|
| 434 |
+
"epoch": 1.76,
|
| 435 |
+
"learning_rate": 9.477442593660033e-06,
|
| 436 |
+
"loss": 0.9147,
|
| 437 |
+
"step": 700
|
| 438 |
+
},
|
| 439 |
+
{
|
| 440 |
+
"epoch": 1.79,
|
| 441 |
+
"learning_rate": 9.498064712656333e-06,
|
| 442 |
+
"loss": 0.9096,
|
| 443 |
+
"step": 710
|
| 444 |
+
},
|
| 445 |
+
{
|
| 446 |
+
"epoch": 1.81,
|
| 447 |
+
"learning_rate": 9.518397185559335e-06,
|
| 448 |
+
"loss": 0.9009,
|
| 449 |
+
"step": 720
|
| 450 |
+
},
|
| 451 |
+
{
|
| 452 |
+
"epoch": 1.84,
|
| 453 |
+
"learning_rate": 9.538448036196794e-06,
|
| 454 |
+
"loss": 0.9214,
|
| 455 |
+
"step": 730
|
| 456 |
+
},
|
| 457 |
+
{
|
| 458 |
+
"epoch": 1.86,
|
| 459 |
+
"learning_rate": 9.558224959530173e-06,
|
| 460 |
+
"loss": 0.9198,
|
| 461 |
+
"step": 740
|
| 462 |
+
},
|
| 463 |
+
{
|
| 464 |
+
"epoch": 1.89,
|
| 465 |
+
"learning_rate": 9.577735339384664e-06,
|
| 466 |
+
"loss": 0.9162,
|
| 467 |
+
"step": 750
|
| 468 |
+
},
|
| 469 |
+
{
|
| 470 |
+
"epoch": 1.91,
|
| 471 |
+
"learning_rate": 9.596986265000242e-06,
|
| 472 |
+
"loss": 0.8982,
|
| 473 |
+
"step": 760
|
| 474 |
+
},
|
| 475 |
+
{
|
| 476 |
+
"epoch": 1.94,
|
| 477 |
+
"learning_rate": 9.615984546496603e-06,
|
| 478 |
+
"loss": 0.9075,
|
| 479 |
+
"step": 770
|
| 480 |
+
},
|
| 481 |
+
{
|
| 482 |
+
"epoch": 1.96,
|
| 483 |
+
"learning_rate": 9.634736729336382e-06,
|
| 484 |
+
"loss": 0.9104,
|
| 485 |
+
"step": 780
|
| 486 |
+
},
|
| 487 |
+
{
|
| 488 |
+
"epoch": 1.99,
|
| 489 |
+
"learning_rate": 9.653249107863549e-06,
|
| 490 |
+
"loss": 0.8939,
|
| 491 |
+
"step": 790
|
| 492 |
+
},
|
| 493 |
+
{
|
| 494 |
+
"epoch": 2.01,
|
| 495 |
+
"learning_rate": 9.671527737987041e-06,
|
| 496 |
+
"loss": 0.8777,
|
| 497 |
+
"step": 800
|
| 498 |
+
},
|
| 499 |
+
{
|
| 500 |
+
"epoch": 2.04,
|
| 501 |
+
"learning_rate": 9.68957844907357e-06,
|
| 502 |
+
"loss": 0.8674,
|
| 503 |
+
"step": 810
|
| 504 |
+
},
|
| 505 |
+
{
|
| 506 |
+
"epoch": 2.06,
|
| 507 |
+
"learning_rate": 9.707406855108054e-06,
|
| 508 |
+
"loss": 0.8647,
|
| 509 |
+
"step": 820
|
| 510 |
+
},
|
| 511 |
+
{
|
| 512 |
+
"epoch": 2.09,
|
| 513 |
+
"learning_rate": 9.725018365175157e-06,
|
| 514 |
+
"loss": 0.8778,
|
| 515 |
+
"step": 830
|
| 516 |
+
},
|
| 517 |
+
{
|
| 518 |
+
"epoch": 2.11,
|
| 519 |
+
"learning_rate": 9.742418193310868e-06,
|
| 520 |
+
"loss": 0.8665,
|
| 521 |
+
"step": 840
|
| 522 |
+
},
|
| 523 |
+
{
|
| 524 |
+
"epoch": 2.14,
|
| 525 |
+
"learning_rate": 9.759611367769025e-06,
|
| 526 |
+
"loss": 0.8717,
|
| 527 |
+
"step": 850
|
| 528 |
+
},
|
| 529 |
+
{
|
| 530 |
+
"epoch": 2.16,
|
| 531 |
+
"learning_rate": 9.776602739743996e-06,
|
| 532 |
+
"loss": 0.8663,
|
| 533 |
+
"step": 860
|
| 534 |
+
},
|
| 535 |
+
{
|
| 536 |
+
"epoch": 2.19,
|
| 537 |
+
"learning_rate": 9.793396991587369e-06,
|
| 538 |
+
"loss": 0.8599,
|
| 539 |
+
"step": 870
|
| 540 |
+
},
|
| 541 |
+
{
|
| 542 |
+
"epoch": 2.21,
|
| 543 |
+
"learning_rate": 9.80999864455347e-06,
|
| 544 |
+
"loss": 0.8634,
|
| 545 |
+
"step": 880
|
| 546 |
+
},
|
| 547 |
+
{
|
| 548 |
+
"epoch": 2.24,
|
| 549 |
+
"learning_rate": 9.826412066105755e-06,
|
| 550 |
+
"loss": 0.8725,
|
| 551 |
+
"step": 890
|
| 552 |
+
},
|
| 553 |
+
{
|
| 554 |
+
"epoch": 2.27,
|
| 555 |
+
"learning_rate": 9.842641476813642e-06,
|
| 556 |
+
"loss": 0.8699,
|
| 557 |
+
"step": 900
|
| 558 |
+
},
|
| 559 |
+
{
|
| 560 |
+
"epoch": 2.29,
|
| 561 |
+
"learning_rate": 9.858690956866985e-06,
|
| 562 |
+
"loss": 0.8634,
|
| 563 |
+
"step": 910
|
| 564 |
+
},
|
| 565 |
+
{
|
| 566 |
+
"epoch": 2.32,
|
| 567 |
+
"learning_rate": 9.874564452233405e-06,
|
| 568 |
+
"loss": 0.8753,
|
| 569 |
+
"step": 920
|
| 570 |
+
},
|
| 571 |
+
{
|
| 572 |
+
"epoch": 2.34,
|
| 573 |
+
"learning_rate": 9.890265780481658e-06,
|
| 574 |
+
"loss": 0.8694,
|
| 575 |
+
"step": 930
|
| 576 |
+
},
|
| 577 |
+
{
|
| 578 |
+
"epoch": 2.37,
|
| 579 |
+
"learning_rate": 9.905798636292595e-06,
|
| 580 |
+
"loss": 0.8701,
|
| 581 |
+
"step": 940
|
| 582 |
+
},
|
| 583 |
+
{
|
| 584 |
+
"epoch": 2.39,
|
| 585 |
+
"learning_rate": 9.92116659667758e-06,
|
| 586 |
+
"loss": 0.88,
|
| 587 |
+
"step": 950
|
| 588 |
+
},
|
| 589 |
+
{
|
| 590 |
+
"epoch": 2.42,
|
| 591 |
+
"learning_rate": 9.936373125922813e-06,
|
| 592 |
+
"loss": 0.8711,
|
| 593 |
+
"step": 960
|
| 594 |
+
},
|
| 595 |
+
{
|
| 596 |
+
"epoch": 2.44,
|
| 597 |
+
"learning_rate": 9.951421580276673e-06,
|
| 598 |
+
"loss": 0.8564,
|
| 599 |
+
"step": 970
|
| 600 |
+
},
|
| 601 |
+
{
|
| 602 |
+
"epoch": 2.47,
|
| 603 |
+
"learning_rate": 9.966315212395911e-06,
|
| 604 |
+
"loss": 0.879,
|
| 605 |
+
"step": 980
|
| 606 |
+
},
|
| 607 |
+
{
|
| 608 |
+
"epoch": 2.49,
|
| 609 |
+
"learning_rate": 9.981057175565457e-06,
|
| 610 |
+
"loss": 0.8744,
|
| 611 |
+
"step": 990
|
| 612 |
+
},
|
| 613 |
+
{
|
| 614 |
+
"epoch": 2.52,
|
| 615 |
+
"learning_rate": 9.99565052770552e-06,
|
| 616 |
+
"loss": 0.8517,
|
| 617 |
+
"step": 1000
|
| 618 |
+
},
|
| 619 |
+
{
|
| 620 |
+
"epoch": 2.52,
|
| 621 |
+
"eval_code_evol_accuracy": 0.7573287954429739,
|
| 622 |
+
"eval_code_evol_loss": 0.908203125,
|
| 623 |
+
"eval_code_evol_runtime": 9.3653,
|
| 624 |
+
"eval_code_evol_samples_per_second": 82.325,
|
| 625 |
+
"eval_code_evol_steps_per_second": 10.357,
|
| 626 |
+
"step": 1000
|
| 627 |
+
},
|
| 628 |
+
{
|
| 629 |
+
"epoch": 2.54,
|
| 630 |
+
"learning_rate": 9.972426470588235e-06,
|
| 631 |
+
"loss": 0.8765,
|
| 632 |
+
"step": 1010
|
| 633 |
+
},
|
| 634 |
+
{
|
| 635 |
+
"epoch": 2.57,
|
| 636 |
+
"learning_rate": 9.926470588235295e-06,
|
| 637 |
+
"loss": 0.877,
|
| 638 |
+
"step": 1020
|
| 639 |
+
},
|
| 640 |
+
{
|
| 641 |
+
"epoch": 2.59,
|
| 642 |
+
"learning_rate": 9.880514705882353e-06,
|
| 643 |
+
"loss": 0.8678,
|
| 644 |
+
"step": 1030
|
| 645 |
+
},
|
| 646 |
+
{
|
| 647 |
+
"epoch": 2.62,
|
| 648 |
+
"learning_rate": 9.834558823529412e-06,
|
| 649 |
+
"loss": 0.8656,
|
| 650 |
+
"step": 1040
|
| 651 |
+
},
|
| 652 |
+
{
|
| 653 |
+
"epoch": 2.64,
|
| 654 |
+
"learning_rate": 9.788602941176472e-06,
|
| 655 |
+
"loss": 0.8811,
|
| 656 |
+
"step": 1050
|
| 657 |
+
},
|
| 658 |
+
{
|
| 659 |
+
"epoch": 2.67,
|
| 660 |
+
"learning_rate": 9.74264705882353e-06,
|
| 661 |
+
"loss": 0.8484,
|
| 662 |
+
"step": 1060
|
| 663 |
+
},
|
| 664 |
+
{
|
| 665 |
+
"epoch": 2.69,
|
| 666 |
+
"learning_rate": 9.69669117647059e-06,
|
| 667 |
+
"loss": 0.8701,
|
| 668 |
+
"step": 1070
|
| 669 |
+
},
|
| 670 |
+
{
|
| 671 |
+
"epoch": 2.72,
|
| 672 |
+
"learning_rate": 9.650735294117649e-06,
|
| 673 |
+
"loss": 0.8549,
|
| 674 |
+
"step": 1080
|
| 675 |
+
},
|
| 676 |
+
{
|
| 677 |
+
"epoch": 2.74,
|
| 678 |
+
"learning_rate": 9.604779411764707e-06,
|
| 679 |
+
"loss": 0.8762,
|
| 680 |
+
"step": 1090
|
| 681 |
+
},
|
| 682 |
+
{
|
| 683 |
+
"epoch": 2.77,
|
| 684 |
+
"learning_rate": 9.558823529411766e-06,
|
| 685 |
+
"loss": 0.8597,
|
| 686 |
+
"step": 1100
|
| 687 |
+
},
|
| 688 |
+
{
|
| 689 |
+
"epoch": 2.79,
|
| 690 |
+
"learning_rate": 9.512867647058824e-06,
|
| 691 |
+
"loss": 0.8579,
|
| 692 |
+
"step": 1110
|
| 693 |
+
},
|
| 694 |
+
{
|
| 695 |
+
"epoch": 2.82,
|
| 696 |
+
"learning_rate": 9.466911764705882e-06,
|
| 697 |
+
"loss": 0.8698,
|
| 698 |
+
"step": 1120
|
| 699 |
+
},
|
| 700 |
+
{
|
| 701 |
+
"epoch": 2.84,
|
| 702 |
+
"learning_rate": 9.420955882352942e-06,
|
| 703 |
+
"loss": 0.8696,
|
| 704 |
+
"step": 1130
|
| 705 |
+
},
|
| 706 |
+
{
|
| 707 |
+
"epoch": 2.87,
|
| 708 |
+
"learning_rate": 9.375000000000001e-06,
|
| 709 |
+
"loss": 0.8796,
|
| 710 |
+
"step": 1140
|
| 711 |
+
},
|
| 712 |
+
{
|
| 713 |
+
"epoch": 2.89,
|
| 714 |
+
"learning_rate": 9.32904411764706e-06,
|
| 715 |
+
"loss": 0.8689,
|
| 716 |
+
"step": 1150
|
| 717 |
+
},
|
| 718 |
+
{
|
| 719 |
+
"epoch": 2.92,
|
| 720 |
+
"learning_rate": 9.28308823529412e-06,
|
| 721 |
+
"loss": 0.8763,
|
| 722 |
+
"step": 1160
|
| 723 |
+
},
|
| 724 |
+
{
|
| 725 |
+
"epoch": 2.94,
|
| 726 |
+
"learning_rate": 9.237132352941178e-06,
|
| 727 |
+
"loss": 0.8639,
|
| 728 |
+
"step": 1170
|
| 729 |
+
},
|
| 730 |
+
{
|
| 731 |
+
"epoch": 2.97,
|
| 732 |
+
"learning_rate": 9.191176470588236e-06,
|
| 733 |
+
"loss": 0.8658,
|
| 734 |
+
"step": 1180
|
| 735 |
+
},
|
| 736 |
+
{
|
| 737 |
+
"epoch": 2.99,
|
| 738 |
+
"learning_rate": 9.145220588235295e-06,
|
| 739 |
+
"loss": 0.8741,
|
| 740 |
+
"step": 1190
|
| 741 |
+
},
|
| 742 |
+
{
|
| 743 |
+
"epoch": 3.02,
|
| 744 |
+
"learning_rate": 9.099264705882353e-06,
|
| 745 |
+
"loss": 0.8233,
|
| 746 |
+
"step": 1200
|
| 747 |
+
},
|
| 748 |
+
{
|
| 749 |
+
"epoch": 3.05,
|
| 750 |
+
"learning_rate": 9.053308823529411e-06,
|
| 751 |
+
"loss": 0.8298,
|
| 752 |
+
"step": 1210
|
| 753 |
+
},
|
| 754 |
+
{
|
| 755 |
+
"epoch": 3.07,
|
| 756 |
+
"learning_rate": 9.007352941176471e-06,
|
| 757 |
+
"loss": 0.8378,
|
| 758 |
+
"step": 1220
|
| 759 |
+
},
|
| 760 |
+
{
|
| 761 |
+
"epoch": 3.1,
|
| 762 |
+
"learning_rate": 8.96139705882353e-06,
|
| 763 |
+
"loss": 0.8387,
|
| 764 |
+
"step": 1230
|
| 765 |
+
},
|
| 766 |
+
{
|
| 767 |
+
"epoch": 3.12,
|
| 768 |
+
"learning_rate": 8.91544117647059e-06,
|
| 769 |
+
"loss": 0.8277,
|
| 770 |
+
"step": 1240
|
| 771 |
+
},
|
| 772 |
+
{
|
| 773 |
+
"epoch": 3.15,
|
| 774 |
+
"learning_rate": 8.869485294117648e-06,
|
| 775 |
+
"loss": 0.8276,
|
| 776 |
+
"step": 1250
|
| 777 |
+
},
|
| 778 |
+
{
|
| 779 |
+
"epoch": 3.17,
|
| 780 |
+
"learning_rate": 8.823529411764707e-06,
|
| 781 |
+
"loss": 0.8398,
|
| 782 |
+
"step": 1260
|
| 783 |
+
},
|
| 784 |
+
{
|
| 785 |
+
"epoch": 3.2,
|
| 786 |
+
"learning_rate": 8.777573529411765e-06,
|
| 787 |
+
"loss": 0.8344,
|
| 788 |
+
"step": 1270
|
| 789 |
+
},
|
| 790 |
+
{
|
| 791 |
+
"epoch": 3.22,
|
| 792 |
+
"learning_rate": 8.731617647058824e-06,
|
| 793 |
+
"loss": 0.8274,
|
| 794 |
+
"step": 1280
|
| 795 |
+
},
|
| 796 |
+
{
|
| 797 |
+
"epoch": 3.25,
|
| 798 |
+
"learning_rate": 8.685661764705882e-06,
|
| 799 |
+
"loss": 0.8296,
|
| 800 |
+
"step": 1290
|
| 801 |
+
},
|
| 802 |
+
{
|
| 803 |
+
"epoch": 3.27,
|
| 804 |
+
"learning_rate": 8.639705882352942e-06,
|
| 805 |
+
"loss": 0.8286,
|
| 806 |
+
"step": 1300
|
| 807 |
+
},
|
| 808 |
+
{
|
| 809 |
+
"epoch": 3.3,
|
| 810 |
+
"learning_rate": 8.59375e-06,
|
| 811 |
+
"loss": 0.8342,
|
| 812 |
+
"step": 1310
|
| 813 |
+
},
|
| 814 |
+
{
|
| 815 |
+
"epoch": 3.32,
|
| 816 |
+
"learning_rate": 8.547794117647059e-06,
|
| 817 |
+
"loss": 0.8265,
|
| 818 |
+
"step": 1320
|
| 819 |
+
},
|
| 820 |
+
{
|
| 821 |
+
"epoch": 3.35,
|
| 822 |
+
"learning_rate": 8.501838235294119e-06,
|
| 823 |
+
"loss": 0.8284,
|
| 824 |
+
"step": 1330
|
| 825 |
+
},
|
| 826 |
+
{
|
| 827 |
+
"epoch": 3.37,
|
| 828 |
+
"learning_rate": 8.455882352941177e-06,
|
| 829 |
+
"loss": 0.8316,
|
| 830 |
+
"step": 1340
|
| 831 |
+
},
|
| 832 |
+
{
|
| 833 |
+
"epoch": 3.4,
|
| 834 |
+
"learning_rate": 8.409926470588236e-06,
|
| 835 |
+
"loss": 0.8145,
|
| 836 |
+
"step": 1350
|
| 837 |
+
},
|
| 838 |
+
{
|
| 839 |
+
"epoch": 3.42,
|
| 840 |
+
"learning_rate": 8.363970588235294e-06,
|
| 841 |
+
"loss": 0.8182,
|
| 842 |
+
"step": 1360
|
| 843 |
+
},
|
| 844 |
+
{
|
| 845 |
+
"epoch": 3.45,
|
| 846 |
+
"learning_rate": 8.318014705882354e-06,
|
| 847 |
+
"loss": 0.8239,
|
| 848 |
+
"step": 1370
|
| 849 |
+
},
|
| 850 |
+
{
|
| 851 |
+
"epoch": 3.47,
|
| 852 |
+
"learning_rate": 8.272058823529413e-06,
|
| 853 |
+
"loss": 0.8319,
|
| 854 |
+
"step": 1380
|
| 855 |
+
},
|
| 856 |
+
{
|
| 857 |
+
"epoch": 3.5,
|
| 858 |
+
"learning_rate": 8.226102941176471e-06,
|
| 859 |
+
"loss": 0.8332,
|
| 860 |
+
"step": 1390
|
| 861 |
+
},
|
| 862 |
+
{
|
| 863 |
+
"epoch": 3.52,
|
| 864 |
+
"learning_rate": 8.18014705882353e-06,
|
| 865 |
+
"loss": 0.8107,
|
| 866 |
+
"step": 1400
|
| 867 |
+
},
|
| 868 |
+
{
|
| 869 |
+
"epoch": 3.55,
|
| 870 |
+
"learning_rate": 8.13419117647059e-06,
|
| 871 |
+
"loss": 0.8289,
|
| 872 |
+
"step": 1410
|
| 873 |
+
},
|
| 874 |
+
{
|
| 875 |
+
"epoch": 3.57,
|
| 876 |
+
"learning_rate": 8.088235294117648e-06,
|
| 877 |
+
"loss": 0.8384,
|
| 878 |
+
"step": 1420
|
| 879 |
+
},
|
| 880 |
+
{
|
| 881 |
+
"epoch": 3.6,
|
| 882 |
+
"learning_rate": 8.042279411764706e-06,
|
| 883 |
+
"loss": 0.8264,
|
| 884 |
+
"step": 1430
|
| 885 |
+
},
|
| 886 |
+
{
|
| 887 |
+
"epoch": 3.62,
|
| 888 |
+
"learning_rate": 7.996323529411767e-06,
|
| 889 |
+
"loss": 0.8234,
|
| 890 |
+
"step": 1440
|
| 891 |
+
},
|
| 892 |
+
{
|
| 893 |
+
"epoch": 3.65,
|
| 894 |
+
"learning_rate": 7.950367647058825e-06,
|
| 895 |
+
"loss": 0.8188,
|
| 896 |
+
"step": 1450
|
| 897 |
+
},
|
| 898 |
+
{
|
| 899 |
+
"epoch": 3.67,
|
| 900 |
+
"learning_rate": 7.904411764705883e-06,
|
| 901 |
+
"loss": 0.8393,
|
| 902 |
+
"step": 1460
|
| 903 |
+
},
|
| 904 |
+
{
|
| 905 |
+
"epoch": 3.7,
|
| 906 |
+
"learning_rate": 7.858455882352942e-06,
|
| 907 |
+
"loss": 0.8183,
|
| 908 |
+
"step": 1470
|
| 909 |
+
},
|
| 910 |
+
{
|
| 911 |
+
"epoch": 3.72,
|
| 912 |
+
"learning_rate": 7.8125e-06,
|
| 913 |
+
"loss": 0.8543,
|
| 914 |
+
"step": 1480
|
| 915 |
+
},
|
| 916 |
+
{
|
| 917 |
+
"epoch": 3.75,
|
| 918 |
+
"learning_rate": 7.766544117647059e-06,
|
| 919 |
+
"loss": 0.8216,
|
| 920 |
+
"step": 1490
|
| 921 |
+
},
|
| 922 |
+
{
|
| 923 |
+
"epoch": 3.78,
|
| 924 |
+
"learning_rate": 7.720588235294119e-06,
|
| 925 |
+
"loss": 0.8392,
|
| 926 |
+
"step": 1500
|
| 927 |
+
},
|
| 928 |
+
{
|
| 929 |
+
"epoch": 3.78,
|
| 930 |
+
"eval_code_evol_accuracy": 0.7596694994310552,
|
| 931 |
+
"eval_code_evol_loss": 0.90087890625,
|
| 932 |
+
"eval_code_evol_runtime": 10.7536,
|
| 933 |
+
"eval_code_evol_samples_per_second": 71.697,
|
| 934 |
+
"eval_code_evol_steps_per_second": 9.02,
|
| 935 |
+
"step": 1500
|
| 936 |
+
},
|
| 937 |
+
{
|
| 938 |
+
"epoch": 3.8,
|
| 939 |
+
"learning_rate": 7.674632352941177e-06,
|
| 940 |
+
"loss": 0.8201,
|
| 941 |
+
"step": 1510
|
| 942 |
+
},
|
| 943 |
+
{
|
| 944 |
+
"epoch": 3.83,
|
| 945 |
+
"learning_rate": 7.6286764705882355e-06,
|
| 946 |
+
"loss": 0.8245,
|
| 947 |
+
"step": 1520
|
| 948 |
+
},
|
| 949 |
+
{
|
| 950 |
+
"epoch": 3.85,
|
| 951 |
+
"learning_rate": 7.582720588235295e-06,
|
| 952 |
+
"loss": 0.832,
|
| 953 |
+
"step": 1530
|
| 954 |
+
},
|
| 955 |
+
{
|
| 956 |
+
"epoch": 3.88,
|
| 957 |
+
"learning_rate": 7.536764705882353e-06,
|
| 958 |
+
"loss": 0.8208,
|
| 959 |
+
"step": 1540
|
| 960 |
+
},
|
| 961 |
+
{
|
| 962 |
+
"epoch": 3.9,
|
| 963 |
+
"learning_rate": 7.4908088235294115e-06,
|
| 964 |
+
"loss": 0.8362,
|
| 965 |
+
"step": 1550
|
| 966 |
+
},
|
| 967 |
+
{
|
| 968 |
+
"epoch": 3.93,
|
| 969 |
+
"learning_rate": 7.444852941176472e-06,
|
| 970 |
+
"loss": 0.8405,
|
| 971 |
+
"step": 1560
|
| 972 |
+
},
|
| 973 |
+
{
|
| 974 |
+
"epoch": 3.95,
|
| 975 |
+
"learning_rate": 7.39889705882353e-06,
|
| 976 |
+
"loss": 0.8319,
|
| 977 |
+
"step": 1570
|
| 978 |
+
},
|
| 979 |
+
{
|
| 980 |
+
"epoch": 3.98,
|
| 981 |
+
"learning_rate": 7.352941176470589e-06,
|
| 982 |
+
"loss": 0.8327,
|
| 983 |
+
"step": 1580
|
| 984 |
+
},
|
| 985 |
+
{
|
| 986 |
+
"epoch": 4.0,
|
| 987 |
+
"learning_rate": 7.306985294117648e-06,
|
| 988 |
+
"loss": 0.8244,
|
| 989 |
+
"step": 1590
|
| 990 |
+
},
|
| 991 |
+
{
|
| 992 |
+
"epoch": 4.03,
|
| 993 |
+
"learning_rate": 7.261029411764706e-06,
|
| 994 |
+
"loss": 0.7933,
|
| 995 |
+
"step": 1600
|
| 996 |
+
},
|
| 997 |
+
{
|
| 998 |
+
"epoch": 4.05,
|
| 999 |
+
"learning_rate": 7.215073529411765e-06,
|
| 1000 |
+
"loss": 0.7896,
|
| 1001 |
+
"step": 1610
|
| 1002 |
+
},
|
| 1003 |
+
{
|
| 1004 |
+
"epoch": 4.08,
|
| 1005 |
+
"learning_rate": 7.169117647058825e-06,
|
| 1006 |
+
"loss": 0.7918,
|
| 1007 |
+
"step": 1620
|
| 1008 |
+
},
|
| 1009 |
+
{
|
| 1010 |
+
"epoch": 4.1,
|
| 1011 |
+
"learning_rate": 7.123161764705883e-06,
|
| 1012 |
+
"loss": 0.7935,
|
| 1013 |
+
"step": 1630
|
| 1014 |
+
},
|
| 1015 |
+
{
|
| 1016 |
+
"epoch": 4.13,
|
| 1017 |
+
"learning_rate": 7.077205882352942e-06,
|
| 1018 |
+
"loss": 0.7995,
|
| 1019 |
+
"step": 1640
|
| 1020 |
+
},
|
| 1021 |
+
{
|
| 1022 |
+
"epoch": 4.15,
|
| 1023 |
+
"learning_rate": 7.031250000000001e-06,
|
| 1024 |
+
"loss": 0.8056,
|
| 1025 |
+
"step": 1650
|
| 1026 |
+
},
|
| 1027 |
+
{
|
| 1028 |
+
"epoch": 4.18,
|
| 1029 |
+
"learning_rate": 6.985294117647059e-06,
|
| 1030 |
+
"loss": 0.798,
|
| 1031 |
+
"step": 1660
|
| 1032 |
+
},
|
| 1033 |
+
{
|
| 1034 |
+
"epoch": 4.2,
|
| 1035 |
+
"learning_rate": 6.939338235294118e-06,
|
| 1036 |
+
"loss": 0.7755,
|
| 1037 |
+
"step": 1670
|
| 1038 |
+
},
|
| 1039 |
+
{
|
| 1040 |
+
"epoch": 4.23,
|
| 1041 |
+
"learning_rate": 6.893382352941177e-06,
|
| 1042 |
+
"loss": 0.7891,
|
| 1043 |
+
"step": 1680
|
| 1044 |
+
},
|
| 1045 |
+
{
|
| 1046 |
+
"epoch": 4.25,
|
| 1047 |
+
"learning_rate": 6.847426470588235e-06,
|
| 1048 |
+
"loss": 0.7904,
|
| 1049 |
+
"step": 1690
|
| 1050 |
+
},
|
| 1051 |
+
{
|
| 1052 |
+
"epoch": 4.28,
|
| 1053 |
+
"learning_rate": 6.801470588235295e-06,
|
| 1054 |
+
"loss": 0.8003,
|
| 1055 |
+
"step": 1700
|
| 1056 |
+
},
|
| 1057 |
+
{
|
| 1058 |
+
"epoch": 4.3,
|
| 1059 |
+
"learning_rate": 6.755514705882354e-06,
|
| 1060 |
+
"loss": 0.7922,
|
| 1061 |
+
"step": 1710
|
| 1062 |
+
},
|
| 1063 |
+
{
|
| 1064 |
+
"epoch": 4.33,
|
| 1065 |
+
"learning_rate": 6.709558823529412e-06,
|
| 1066 |
+
"loss": 0.7832,
|
| 1067 |
+
"step": 1720
|
| 1068 |
+
},
|
| 1069 |
+
{
|
| 1070 |
+
"epoch": 4.35,
|
| 1071 |
+
"learning_rate": 6.663602941176471e-06,
|
| 1072 |
+
"loss": 0.7901,
|
| 1073 |
+
"step": 1730
|
| 1074 |
+
},
|
| 1075 |
+
{
|
| 1076 |
+
"epoch": 4.38,
|
| 1077 |
+
"learning_rate": 6.61764705882353e-06,
|
| 1078 |
+
"loss": 0.8005,
|
| 1079 |
+
"step": 1740
|
| 1080 |
+
},
|
| 1081 |
+
{
|
| 1082 |
+
"epoch": 4.4,
|
| 1083 |
+
"learning_rate": 6.571691176470589e-06,
|
| 1084 |
+
"loss": 0.7988,
|
| 1085 |
+
"step": 1750
|
| 1086 |
+
},
|
| 1087 |
+
{
|
| 1088 |
+
"epoch": 4.43,
|
| 1089 |
+
"learning_rate": 6.525735294117647e-06,
|
| 1090 |
+
"loss": 0.7879,
|
| 1091 |
+
"step": 1760
|
| 1092 |
+
},
|
| 1093 |
+
{
|
| 1094 |
+
"epoch": 4.45,
|
| 1095 |
+
"learning_rate": 6.479779411764706e-06,
|
| 1096 |
+
"loss": 0.8045,
|
| 1097 |
+
"step": 1770
|
| 1098 |
+
},
|
| 1099 |
+
{
|
| 1100 |
+
"epoch": 4.48,
|
| 1101 |
+
"learning_rate": 6.433823529411766e-06,
|
| 1102 |
+
"loss": 0.8042,
|
| 1103 |
+
"step": 1780
|
| 1104 |
+
},
|
| 1105 |
+
{
|
| 1106 |
+
"epoch": 4.51,
|
| 1107 |
+
"learning_rate": 6.387867647058824e-06,
|
| 1108 |
+
"loss": 0.7951,
|
| 1109 |
+
"step": 1790
|
| 1110 |
+
},
|
| 1111 |
+
{
|
| 1112 |
+
"epoch": 4.53,
|
| 1113 |
+
"learning_rate": 6.341911764705883e-06,
|
| 1114 |
+
"loss": 0.7858,
|
| 1115 |
+
"step": 1800
|
| 1116 |
+
},
|
| 1117 |
+
{
|
| 1118 |
+
"epoch": 4.56,
|
| 1119 |
+
"learning_rate": 6.295955882352942e-06,
|
| 1120 |
+
"loss": 0.8009,
|
| 1121 |
+
"step": 1810
|
| 1122 |
+
},
|
| 1123 |
+
{
|
| 1124 |
+
"epoch": 4.58,
|
| 1125 |
+
"learning_rate": 6.25e-06,
|
| 1126 |
+
"loss": 0.8055,
|
| 1127 |
+
"step": 1820
|
| 1128 |
+
},
|
| 1129 |
+
{
|
| 1130 |
+
"epoch": 4.61,
|
| 1131 |
+
"learning_rate": 6.204044117647059e-06,
|
| 1132 |
+
"loss": 0.8041,
|
| 1133 |
+
"step": 1830
|
| 1134 |
+
},
|
| 1135 |
+
{
|
| 1136 |
+
"epoch": 4.63,
|
| 1137 |
+
"learning_rate": 6.158088235294119e-06,
|
| 1138 |
+
"loss": 0.7989,
|
| 1139 |
+
"step": 1840
|
| 1140 |
+
},
|
| 1141 |
+
{
|
| 1142 |
+
"epoch": 4.66,
|
| 1143 |
+
"learning_rate": 6.112132352941177e-06,
|
| 1144 |
+
"loss": 0.8029,
|
| 1145 |
+
"step": 1850
|
| 1146 |
+
},
|
| 1147 |
+
{
|
| 1148 |
+
"epoch": 4.68,
|
| 1149 |
+
"learning_rate": 6.066176470588236e-06,
|
| 1150 |
+
"loss": 0.8021,
|
| 1151 |
+
"step": 1860
|
| 1152 |
+
},
|
| 1153 |
+
{
|
| 1154 |
+
"epoch": 4.71,
|
| 1155 |
+
"learning_rate": 6.020220588235295e-06,
|
| 1156 |
+
"loss": 0.8085,
|
| 1157 |
+
"step": 1870
|
| 1158 |
+
},
|
| 1159 |
+
{
|
| 1160 |
+
"epoch": 4.73,
|
| 1161 |
+
"learning_rate": 5.974264705882353e-06,
|
| 1162 |
+
"loss": 0.7904,
|
| 1163 |
+
"step": 1880
|
| 1164 |
+
},
|
| 1165 |
+
{
|
| 1166 |
+
"epoch": 4.76,
|
| 1167 |
+
"learning_rate": 5.928308823529412e-06,
|
| 1168 |
+
"loss": 0.8014,
|
| 1169 |
+
"step": 1890
|
| 1170 |
+
},
|
| 1171 |
+
{
|
| 1172 |
+
"epoch": 4.78,
|
| 1173 |
+
"learning_rate": 5.882352941176471e-06,
|
| 1174 |
+
"loss": 0.8114,
|
| 1175 |
+
"step": 1900
|
| 1176 |
+
},
|
| 1177 |
+
{
|
| 1178 |
+
"epoch": 4.81,
|
| 1179 |
+
"learning_rate": 5.836397058823529e-06,
|
| 1180 |
+
"loss": 0.8067,
|
| 1181 |
+
"step": 1910
|
| 1182 |
+
},
|
| 1183 |
+
{
|
| 1184 |
+
"epoch": 4.83,
|
| 1185 |
+
"learning_rate": 5.790441176470589e-06,
|
| 1186 |
+
"loss": 0.7983,
|
| 1187 |
+
"step": 1920
|
| 1188 |
+
},
|
| 1189 |
+
{
|
| 1190 |
+
"epoch": 4.86,
|
| 1191 |
+
"learning_rate": 5.744485294117648e-06,
|
| 1192 |
+
"loss": 0.7998,
|
| 1193 |
+
"step": 1930
|
| 1194 |
+
},
|
| 1195 |
+
{
|
| 1196 |
+
"epoch": 4.88,
|
| 1197 |
+
"learning_rate": 5.698529411764706e-06,
|
| 1198 |
+
"loss": 0.7973,
|
| 1199 |
+
"step": 1940
|
| 1200 |
+
},
|
| 1201 |
+
{
|
| 1202 |
+
"epoch": 4.91,
|
| 1203 |
+
"learning_rate": 5.6525735294117655e-06,
|
| 1204 |
+
"loss": 0.8186,
|
| 1205 |
+
"step": 1950
|
| 1206 |
+
},
|
| 1207 |
+
{
|
| 1208 |
+
"epoch": 4.93,
|
| 1209 |
+
"learning_rate": 5.606617647058824e-06,
|
| 1210 |
+
"loss": 0.7907,
|
| 1211 |
+
"step": 1960
|
| 1212 |
+
},
|
| 1213 |
+
{
|
| 1214 |
+
"epoch": 4.96,
|
| 1215 |
+
"learning_rate": 5.560661764705882e-06,
|
| 1216 |
+
"loss": 0.8019,
|
| 1217 |
+
"step": 1970
|
| 1218 |
+
},
|
| 1219 |
+
{
|
| 1220 |
+
"epoch": 4.98,
|
| 1221 |
+
"learning_rate": 5.514705882352942e-06,
|
| 1222 |
+
"loss": 0.8082,
|
| 1223 |
+
"step": 1980
|
| 1224 |
+
},
|
| 1225 |
+
{
|
| 1226 |
+
"epoch": 5.01,
|
| 1227 |
+
"learning_rate": 5.468750000000001e-06,
|
| 1228 |
+
"loss": 0.7778,
|
| 1229 |
+
"step": 1990
|
| 1230 |
+
},
|
| 1231 |
+
{
|
| 1232 |
+
"epoch": 5.03,
|
| 1233 |
+
"learning_rate": 5.422794117647059e-06,
|
| 1234 |
+
"loss": 0.775,
|
| 1235 |
+
"step": 2000
|
| 1236 |
+
},
|
| 1237 |
+
{
|
| 1238 |
+
"epoch": 5.03,
|
| 1239 |
+
"eval_code_evol_accuracy": 0.7592867857862886,
|
| 1240 |
+
"eval_code_evol_loss": 0.904296875,
|
| 1241 |
+
"eval_code_evol_runtime": 11.4335,
|
| 1242 |
+
"eval_code_evol_samples_per_second": 67.433,
|
| 1243 |
+
"eval_code_evol_steps_per_second": 8.484,
|
| 1244 |
+
"step": 2000
|
| 1245 |
+
},
|
| 1246 |
+
{
|
| 1247 |
+
"epoch": 5.06,
|
| 1248 |
+
"learning_rate": 5.3768382352941184e-06,
|
| 1249 |
+
"loss": 0.765,
|
| 1250 |
+
"step": 2010
|
| 1251 |
+
},
|
| 1252 |
+
{
|
| 1253 |
+
"epoch": 5.08,
|
| 1254 |
+
"learning_rate": 5.330882352941177e-06,
|
| 1255 |
+
"loss": 0.7735,
|
| 1256 |
+
"step": 2020
|
| 1257 |
+
},
|
| 1258 |
+
{
|
| 1259 |
+
"epoch": 5.11,
|
| 1260 |
+
"learning_rate": 5.284926470588235e-06,
|
| 1261 |
+
"loss": 0.76,
|
| 1262 |
+
"step": 2030
|
| 1263 |
+
},
|
| 1264 |
+
{
|
| 1265 |
+
"epoch": 5.13,
|
| 1266 |
+
"learning_rate": 5.2389705882352945e-06,
|
| 1267 |
+
"loss": 0.7769,
|
| 1268 |
+
"step": 2040
|
| 1269 |
+
},
|
| 1270 |
+
{
|
| 1271 |
+
"epoch": 5.16,
|
| 1272 |
+
"learning_rate": 5.193014705882353e-06,
|
| 1273 |
+
"loss": 0.7735,
|
| 1274 |
+
"step": 2050
|
| 1275 |
+
},
|
| 1276 |
+
{
|
| 1277 |
+
"epoch": 5.18,
|
| 1278 |
+
"learning_rate": 5.147058823529411e-06,
|
| 1279 |
+
"loss": 0.7655,
|
| 1280 |
+
"step": 2060
|
| 1281 |
+
},
|
| 1282 |
+
{
|
| 1283 |
+
"epoch": 5.21,
|
| 1284 |
+
"learning_rate": 5.101102941176471e-06,
|
| 1285 |
+
"loss": 0.7742,
|
| 1286 |
+
"step": 2070
|
| 1287 |
+
},
|
| 1288 |
+
{
|
| 1289 |
+
"epoch": 5.23,
|
| 1290 |
+
"learning_rate": 5.05514705882353e-06,
|
| 1291 |
+
"loss": 0.7749,
|
| 1292 |
+
"step": 2080
|
| 1293 |
+
},
|
| 1294 |
+
{
|
| 1295 |
+
"epoch": 5.26,
|
| 1296 |
+
"learning_rate": 5.009191176470589e-06,
|
| 1297 |
+
"loss": 0.782,
|
| 1298 |
+
"step": 2090
|
| 1299 |
+
},
|
| 1300 |
+
{
|
| 1301 |
+
"epoch": 5.29,
|
| 1302 |
+
"learning_rate": 4.9632352941176475e-06,
|
| 1303 |
+
"loss": 0.7806,
|
| 1304 |
+
"step": 2100
|
| 1305 |
+
},
|
| 1306 |
+
{
|
| 1307 |
+
"epoch": 5.31,
|
| 1308 |
+
"learning_rate": 4.917279411764706e-06,
|
| 1309 |
+
"loss": 0.7796,
|
| 1310 |
+
"step": 2110
|
| 1311 |
+
},
|
| 1312 |
+
{
|
| 1313 |
+
"epoch": 5.34,
|
| 1314 |
+
"learning_rate": 4.871323529411765e-06,
|
| 1315 |
+
"loss": 0.7685,
|
| 1316 |
+
"step": 2120
|
| 1317 |
+
},
|
| 1318 |
+
{
|
| 1319 |
+
"epoch": 5.36,
|
| 1320 |
+
"learning_rate": 4.825367647058824e-06,
|
| 1321 |
+
"loss": 0.7681,
|
| 1322 |
+
"step": 2130
|
| 1323 |
+
},
|
| 1324 |
+
{
|
| 1325 |
+
"epoch": 5.39,
|
| 1326 |
+
"learning_rate": 4.779411764705883e-06,
|
| 1327 |
+
"loss": 0.7729,
|
| 1328 |
+
"step": 2140
|
| 1329 |
+
},
|
| 1330 |
+
{
|
| 1331 |
+
"epoch": 5.41,
|
| 1332 |
+
"learning_rate": 4.733455882352941e-06,
|
| 1333 |
+
"loss": 0.7634,
|
| 1334 |
+
"step": 2150
|
| 1335 |
+
},
|
| 1336 |
+
{
|
| 1337 |
+
"epoch": 5.44,
|
| 1338 |
+
"learning_rate": 4.6875000000000004e-06,
|
| 1339 |
+
"loss": 0.7732,
|
| 1340 |
+
"step": 2160
|
| 1341 |
+
},
|
| 1342 |
+
{
|
| 1343 |
+
"epoch": 5.46,
|
| 1344 |
+
"learning_rate": 4.64154411764706e-06,
|
| 1345 |
+
"loss": 0.7703,
|
| 1346 |
+
"step": 2170
|
| 1347 |
+
},
|
| 1348 |
+
{
|
| 1349 |
+
"epoch": 5.49,
|
| 1350 |
+
"learning_rate": 4.595588235294118e-06,
|
| 1351 |
+
"loss": 0.7815,
|
| 1352 |
+
"step": 2180
|
| 1353 |
+
},
|
| 1354 |
+
{
|
| 1355 |
+
"epoch": 5.51,
|
| 1356 |
+
"learning_rate": 4.5496323529411765e-06,
|
| 1357 |
+
"loss": 0.7811,
|
| 1358 |
+
"step": 2190
|
| 1359 |
+
},
|
| 1360 |
+
{
|
| 1361 |
+
"epoch": 5.54,
|
| 1362 |
+
"learning_rate": 4.503676470588236e-06,
|
| 1363 |
+
"loss": 0.7702,
|
| 1364 |
+
"step": 2200
|
| 1365 |
+
},
|
| 1366 |
+
{
|
| 1367 |
+
"epoch": 5.56,
|
| 1368 |
+
"learning_rate": 4.457720588235295e-06,
|
| 1369 |
+
"loss": 0.7753,
|
| 1370 |
+
"step": 2210
|
| 1371 |
+
},
|
| 1372 |
+
{
|
| 1373 |
+
"epoch": 5.59,
|
| 1374 |
+
"learning_rate": 4.411764705882353e-06,
|
| 1375 |
+
"loss": 0.7751,
|
| 1376 |
+
"step": 2220
|
| 1377 |
+
},
|
| 1378 |
+
{
|
| 1379 |
+
"epoch": 5.61,
|
| 1380 |
+
"learning_rate": 4.365808823529412e-06,
|
| 1381 |
+
"loss": 0.7785,
|
| 1382 |
+
"step": 2230
|
| 1383 |
+
},
|
| 1384 |
+
{
|
| 1385 |
+
"epoch": 5.64,
|
| 1386 |
+
"learning_rate": 4.319852941176471e-06,
|
| 1387 |
+
"loss": 0.7793,
|
| 1388 |
+
"step": 2240
|
| 1389 |
+
},
|
| 1390 |
+
{
|
| 1391 |
+
"epoch": 5.66,
|
| 1392 |
+
"learning_rate": 4.2738970588235295e-06,
|
| 1393 |
+
"loss": 0.7861,
|
| 1394 |
+
"step": 2250
|
| 1395 |
+
},
|
| 1396 |
+
{
|
| 1397 |
+
"epoch": 5.69,
|
| 1398 |
+
"learning_rate": 4.227941176470589e-06,
|
| 1399 |
+
"loss": 0.7811,
|
| 1400 |
+
"step": 2260
|
| 1401 |
+
},
|
| 1402 |
+
{
|
| 1403 |
+
"epoch": 5.71,
|
| 1404 |
+
"learning_rate": 4.181985294117647e-06,
|
| 1405 |
+
"loss": 0.7728,
|
| 1406 |
+
"step": 2270
|
| 1407 |
+
},
|
| 1408 |
+
{
|
| 1409 |
+
"epoch": 5.74,
|
| 1410 |
+
"learning_rate": 4.136029411764706e-06,
|
| 1411 |
+
"loss": 0.7803,
|
| 1412 |
+
"step": 2280
|
| 1413 |
+
},
|
| 1414 |
+
{
|
| 1415 |
+
"epoch": 5.76,
|
| 1416 |
+
"learning_rate": 4.090073529411765e-06,
|
| 1417 |
+
"loss": 0.7672,
|
| 1418 |
+
"step": 2290
|
| 1419 |
+
},
|
| 1420 |
+
{
|
| 1421 |
+
"epoch": 5.79,
|
| 1422 |
+
"learning_rate": 4.044117647058824e-06,
|
| 1423 |
+
"loss": 0.7781,
|
| 1424 |
+
"step": 2300
|
| 1425 |
+
},
|
| 1426 |
+
{
|
| 1427 |
+
"epoch": 5.81,
|
| 1428 |
+
"learning_rate": 3.998161764705883e-06,
|
| 1429 |
+
"loss": 0.7861,
|
| 1430 |
+
"step": 2310
|
| 1431 |
+
},
|
| 1432 |
+
{
|
| 1433 |
+
"epoch": 5.84,
|
| 1434 |
+
"learning_rate": 3.952205882352942e-06,
|
| 1435 |
+
"loss": 0.7833,
|
| 1436 |
+
"step": 2320
|
| 1437 |
+
},
|
| 1438 |
+
{
|
| 1439 |
+
"epoch": 5.86,
|
| 1440 |
+
"learning_rate": 3.90625e-06,
|
| 1441 |
+
"loss": 0.7852,
|
| 1442 |
+
"step": 2330
|
| 1443 |
+
},
|
| 1444 |
+
{
|
| 1445 |
+
"epoch": 5.89,
|
| 1446 |
+
"learning_rate": 3.860294117647059e-06,
|
| 1447 |
+
"loss": 0.7662,
|
| 1448 |
+
"step": 2340
|
| 1449 |
+
},
|
| 1450 |
+
{
|
| 1451 |
+
"epoch": 5.91,
|
| 1452 |
+
"learning_rate": 3.8143382352941177e-06,
|
| 1453 |
+
"loss": 0.7777,
|
| 1454 |
+
"step": 2350
|
| 1455 |
+
},
|
| 1456 |
+
{
|
| 1457 |
+
"epoch": 5.94,
|
| 1458 |
+
"learning_rate": 3.7683823529411766e-06,
|
| 1459 |
+
"loss": 0.7646,
|
| 1460 |
+
"step": 2360
|
| 1461 |
+
},
|
| 1462 |
+
{
|
| 1463 |
+
"epoch": 5.96,
|
| 1464 |
+
"learning_rate": 3.722426470588236e-06,
|
| 1465 |
+
"loss": 0.7703,
|
| 1466 |
+
"step": 2370
|
| 1467 |
+
},
|
| 1468 |
+
{
|
| 1469 |
+
"epoch": 5.99,
|
| 1470 |
+
"learning_rate": 3.6764705882352946e-06,
|
| 1471 |
+
"loss": 0.7803,
|
| 1472 |
+
"step": 2380
|
| 1473 |
+
},
|
| 1474 |
+
{
|
| 1475 |
+
"epoch": 6.02,
|
| 1476 |
+
"learning_rate": 3.630514705882353e-06,
|
| 1477 |
+
"loss": 0.7677,
|
| 1478 |
+
"step": 2390
|
| 1479 |
+
},
|
| 1480 |
+
{
|
| 1481 |
+
"epoch": 6.04,
|
| 1482 |
+
"learning_rate": 3.5845588235294123e-06,
|
| 1483 |
+
"loss": 0.747,
|
| 1484 |
+
"step": 2400
|
| 1485 |
+
},
|
| 1486 |
+
{
|
| 1487 |
+
"epoch": 6.07,
|
| 1488 |
+
"learning_rate": 3.538602941176471e-06,
|
| 1489 |
+
"loss": 0.753,
|
| 1490 |
+
"step": 2410
|
| 1491 |
+
},
|
| 1492 |
+
{
|
| 1493 |
+
"epoch": 6.09,
|
| 1494 |
+
"learning_rate": 3.4926470588235295e-06,
|
| 1495 |
+
"loss": 0.7575,
|
| 1496 |
+
"step": 2420
|
| 1497 |
+
},
|
| 1498 |
+
{
|
| 1499 |
+
"epoch": 6.12,
|
| 1500 |
+
"learning_rate": 3.4466911764705884e-06,
|
| 1501 |
+
"loss": 0.7409,
|
| 1502 |
+
"step": 2430
|
| 1503 |
+
},
|
| 1504 |
+
{
|
| 1505 |
+
"epoch": 6.14,
|
| 1506 |
+
"learning_rate": 3.4007352941176476e-06,
|
| 1507 |
+
"loss": 0.7591,
|
| 1508 |
+
"step": 2440
|
| 1509 |
+
},
|
| 1510 |
+
{
|
| 1511 |
+
"epoch": 6.17,
|
| 1512 |
+
"learning_rate": 3.354779411764706e-06,
|
| 1513 |
+
"loss": 0.7659,
|
| 1514 |
+
"step": 2450
|
| 1515 |
+
},
|
| 1516 |
+
{
|
| 1517 |
+
"epoch": 6.19,
|
| 1518 |
+
"learning_rate": 3.308823529411765e-06,
|
| 1519 |
+
"loss": 0.7479,
|
| 1520 |
+
"step": 2460
|
| 1521 |
+
},
|
| 1522 |
+
{
|
| 1523 |
+
"epoch": 6.22,
|
| 1524 |
+
"learning_rate": 3.2628676470588237e-06,
|
| 1525 |
+
"loss": 0.7453,
|
| 1526 |
+
"step": 2470
|
| 1527 |
+
},
|
| 1528 |
+
{
|
| 1529 |
+
"epoch": 6.24,
|
| 1530 |
+
"learning_rate": 3.216911764705883e-06,
|
| 1531 |
+
"loss": 0.752,
|
| 1532 |
+
"step": 2480
|
| 1533 |
+
},
|
| 1534 |
+
{
|
| 1535 |
+
"epoch": 6.27,
|
| 1536 |
+
"learning_rate": 3.1709558823529413e-06,
|
| 1537 |
+
"loss": 0.7508,
|
| 1538 |
+
"step": 2490
|
| 1539 |
+
},
|
| 1540 |
+
{
|
| 1541 |
+
"epoch": 6.29,
|
| 1542 |
+
"learning_rate": 3.125e-06,
|
| 1543 |
+
"loss": 0.7585,
|
| 1544 |
+
"step": 2500
|
| 1545 |
+
},
|
| 1546 |
+
{
|
| 1547 |
+
"epoch": 6.29,
|
| 1548 |
+
"eval_code_evol_accuracy": 0.7601342231425574,
|
| 1549 |
+
"eval_code_evol_loss": 0.90576171875,
|
| 1550 |
+
"eval_code_evol_runtime": 11.4001,
|
| 1551 |
+
"eval_code_evol_samples_per_second": 67.631,
|
| 1552 |
+
"eval_code_evol_steps_per_second": 8.509,
|
| 1553 |
+
"step": 2500
|
| 1554 |
+
},
|
| 1555 |
+
{
|
| 1556 |
+
"epoch": 6.32,
|
| 1557 |
+
"learning_rate": 3.0790441176470594e-06,
|
| 1558 |
+
"loss": 0.7617,
|
| 1559 |
+
"step": 2510
|
| 1560 |
+
},
|
| 1561 |
+
{
|
| 1562 |
+
"epoch": 6.34,
|
| 1563 |
+
"learning_rate": 3.033088235294118e-06,
|
| 1564 |
+
"loss": 0.7565,
|
| 1565 |
+
"step": 2520
|
| 1566 |
+
},
|
| 1567 |
+
{
|
| 1568 |
+
"epoch": 6.37,
|
| 1569 |
+
"learning_rate": 2.9871323529411766e-06,
|
| 1570 |
+
"loss": 0.7588,
|
| 1571 |
+
"step": 2530
|
| 1572 |
+
},
|
| 1573 |
+
{
|
| 1574 |
+
"epoch": 6.39,
|
| 1575 |
+
"learning_rate": 2.9411764705882355e-06,
|
| 1576 |
+
"loss": 0.7637,
|
| 1577 |
+
"step": 2540
|
| 1578 |
+
},
|
| 1579 |
+
{
|
| 1580 |
+
"epoch": 6.42,
|
| 1581 |
+
"learning_rate": 2.8952205882352947e-06,
|
| 1582 |
+
"loss": 0.7585,
|
| 1583 |
+
"step": 2550
|
| 1584 |
+
},
|
| 1585 |
+
{
|
| 1586 |
+
"epoch": 6.44,
|
| 1587 |
+
"learning_rate": 2.849264705882353e-06,
|
| 1588 |
+
"loss": 0.77,
|
| 1589 |
+
"step": 2560
|
| 1590 |
+
},
|
| 1591 |
+
{
|
| 1592 |
+
"epoch": 6.47,
|
| 1593 |
+
"learning_rate": 2.803308823529412e-06,
|
| 1594 |
+
"loss": 0.7599,
|
| 1595 |
+
"step": 2570
|
| 1596 |
+
},
|
| 1597 |
+
{
|
| 1598 |
+
"epoch": 6.49,
|
| 1599 |
+
"learning_rate": 2.757352941176471e-06,
|
| 1600 |
+
"loss": 0.7527,
|
| 1601 |
+
"step": 2580
|
| 1602 |
+
},
|
| 1603 |
+
{
|
| 1604 |
+
"epoch": 6.52,
|
| 1605 |
+
"learning_rate": 2.7113970588235296e-06,
|
| 1606 |
+
"loss": 0.7476,
|
| 1607 |
+
"step": 2590
|
| 1608 |
+
},
|
| 1609 |
+
{
|
| 1610 |
+
"epoch": 6.54,
|
| 1611 |
+
"learning_rate": 2.6654411764705884e-06,
|
| 1612 |
+
"loss": 0.7531,
|
| 1613 |
+
"step": 2600
|
| 1614 |
+
},
|
| 1615 |
+
{
|
| 1616 |
+
"epoch": 6.57,
|
| 1617 |
+
"learning_rate": 2.6194852941176473e-06,
|
| 1618 |
+
"loss": 0.7595,
|
| 1619 |
+
"step": 2610
|
| 1620 |
+
},
|
| 1621 |
+
{
|
| 1622 |
+
"epoch": 6.59,
|
| 1623 |
+
"learning_rate": 2.5735294117647057e-06,
|
| 1624 |
+
"loss": 0.7596,
|
| 1625 |
+
"step": 2620
|
| 1626 |
+
},
|
| 1627 |
+
{
|
| 1628 |
+
"epoch": 6.62,
|
| 1629 |
+
"learning_rate": 2.527573529411765e-06,
|
| 1630 |
+
"loss": 0.7578,
|
| 1631 |
+
"step": 2630
|
| 1632 |
+
},
|
| 1633 |
+
{
|
| 1634 |
+
"epoch": 6.64,
|
| 1635 |
+
"learning_rate": 2.4816176470588237e-06,
|
| 1636 |
+
"loss": 0.7661,
|
| 1637 |
+
"step": 2640
|
| 1638 |
+
},
|
| 1639 |
+
{
|
| 1640 |
+
"epoch": 6.67,
|
| 1641 |
+
"learning_rate": 2.4356617647058826e-06,
|
| 1642 |
+
"loss": 0.7527,
|
| 1643 |
+
"step": 2650
|
| 1644 |
+
},
|
| 1645 |
+
{
|
| 1646 |
+
"epoch": 6.69,
|
| 1647 |
+
"learning_rate": 2.3897058823529414e-06,
|
| 1648 |
+
"loss": 0.7614,
|
| 1649 |
+
"step": 2660
|
| 1650 |
+
},
|
| 1651 |
+
{
|
| 1652 |
+
"epoch": 6.72,
|
| 1653 |
+
"learning_rate": 2.3437500000000002e-06,
|
| 1654 |
+
"loss": 0.7637,
|
| 1655 |
+
"step": 2670
|
| 1656 |
+
},
|
| 1657 |
+
{
|
| 1658 |
+
"epoch": 6.74,
|
| 1659 |
+
"learning_rate": 2.297794117647059e-06,
|
| 1660 |
+
"loss": 0.7699,
|
| 1661 |
+
"step": 2680
|
| 1662 |
+
},
|
| 1663 |
+
{
|
| 1664 |
+
"epoch": 6.77,
|
| 1665 |
+
"learning_rate": 2.251838235294118e-06,
|
| 1666 |
+
"loss": 0.7465,
|
| 1667 |
+
"step": 2690
|
| 1668 |
+
},
|
| 1669 |
+
{
|
| 1670 |
+
"epoch": 6.8,
|
| 1671 |
+
"learning_rate": 2.2058823529411767e-06,
|
| 1672 |
+
"loss": 0.7528,
|
| 1673 |
+
"step": 2700
|
| 1674 |
+
},
|
| 1675 |
+
{
|
| 1676 |
+
"epoch": 6.82,
|
| 1677 |
+
"learning_rate": 2.1599264705882355e-06,
|
| 1678 |
+
"loss": 0.7604,
|
| 1679 |
+
"step": 2710
|
| 1680 |
+
},
|
| 1681 |
+
{
|
| 1682 |
+
"epoch": 6.85,
|
| 1683 |
+
"learning_rate": 2.1139705882352944e-06,
|
| 1684 |
+
"loss": 0.7535,
|
| 1685 |
+
"step": 2720
|
| 1686 |
+
},
|
| 1687 |
+
{
|
| 1688 |
+
"epoch": 6.87,
|
| 1689 |
+
"learning_rate": 2.068014705882353e-06,
|
| 1690 |
+
"loss": 0.7626,
|
| 1691 |
+
"step": 2730
|
| 1692 |
+
},
|
| 1693 |
+
{
|
| 1694 |
+
"epoch": 6.9,
|
| 1695 |
+
"learning_rate": 2.022058823529412e-06,
|
| 1696 |
+
"loss": 0.7517,
|
| 1697 |
+
"step": 2740
|
| 1698 |
+
},
|
| 1699 |
+
{
|
| 1700 |
+
"epoch": 6.92,
|
| 1701 |
+
"learning_rate": 1.976102941176471e-06,
|
| 1702 |
+
"loss": 0.7618,
|
| 1703 |
+
"step": 2750
|
| 1704 |
+
},
|
| 1705 |
+
{
|
| 1706 |
+
"epoch": 6.95,
|
| 1707 |
+
"learning_rate": 1.9301470588235297e-06,
|
| 1708 |
+
"loss": 0.7545,
|
| 1709 |
+
"step": 2760
|
| 1710 |
+
},
|
| 1711 |
+
{
|
| 1712 |
+
"epoch": 6.97,
|
| 1713 |
+
"learning_rate": 1.8841911764705883e-06,
|
| 1714 |
+
"loss": 0.7725,
|
| 1715 |
+
"step": 2770
|
| 1716 |
+
},
|
| 1717 |
+
{
|
| 1718 |
+
"epoch": 7.0,
|
| 1719 |
+
"learning_rate": 1.8382352941176473e-06,
|
| 1720 |
+
"loss": 0.7489,
|
| 1721 |
+
"step": 2780
|
| 1722 |
+
},
|
| 1723 |
+
{
|
| 1724 |
+
"epoch": 7.02,
|
| 1725 |
+
"learning_rate": 1.7922794117647061e-06,
|
| 1726 |
+
"loss": 0.7553,
|
| 1727 |
+
"step": 2790
|
| 1728 |
+
},
|
| 1729 |
+
{
|
| 1730 |
+
"epoch": 7.05,
|
| 1731 |
+
"learning_rate": 1.7463235294117648e-06,
|
| 1732 |
+
"loss": 0.753,
|
| 1733 |
+
"step": 2800
|
| 1734 |
+
},
|
| 1735 |
+
{
|
| 1736 |
+
"epoch": 7.07,
|
| 1737 |
+
"learning_rate": 1.7003676470588238e-06,
|
| 1738 |
+
"loss": 0.7504,
|
| 1739 |
+
"step": 2810
|
| 1740 |
+
},
|
| 1741 |
+
{
|
| 1742 |
+
"epoch": 7.1,
|
| 1743 |
+
"learning_rate": 1.6544117647058824e-06,
|
| 1744 |
+
"loss": 0.751,
|
| 1745 |
+
"step": 2820
|
| 1746 |
+
},
|
| 1747 |
+
{
|
| 1748 |
+
"epoch": 7.12,
|
| 1749 |
+
"learning_rate": 1.6084558823529415e-06,
|
| 1750 |
+
"loss": 0.7473,
|
| 1751 |
+
"step": 2830
|
| 1752 |
+
},
|
| 1753 |
+
{
|
| 1754 |
+
"epoch": 7.15,
|
| 1755 |
+
"learning_rate": 1.5625e-06,
|
| 1756 |
+
"loss": 0.7398,
|
| 1757 |
+
"step": 2840
|
| 1758 |
+
},
|
| 1759 |
+
{
|
| 1760 |
+
"epoch": 7.17,
|
| 1761 |
+
"learning_rate": 1.516544117647059e-06,
|
| 1762 |
+
"loss": 0.7412,
|
| 1763 |
+
"step": 2850
|
| 1764 |
+
},
|
| 1765 |
+
{
|
| 1766 |
+
"epoch": 7.2,
|
| 1767 |
+
"learning_rate": 1.4705882352941177e-06,
|
| 1768 |
+
"loss": 0.7465,
|
| 1769 |
+
"step": 2860
|
| 1770 |
+
},
|
| 1771 |
+
{
|
| 1772 |
+
"epoch": 7.22,
|
| 1773 |
+
"learning_rate": 1.4246323529411766e-06,
|
| 1774 |
+
"loss": 0.7315,
|
| 1775 |
+
"step": 2870
|
| 1776 |
+
},
|
| 1777 |
+
{
|
| 1778 |
+
"epoch": 7.25,
|
| 1779 |
+
"learning_rate": 1.3786764705882356e-06,
|
| 1780 |
+
"loss": 0.7316,
|
| 1781 |
+
"step": 2880
|
| 1782 |
+
},
|
| 1783 |
+
{
|
| 1784 |
+
"epoch": 7.27,
|
| 1785 |
+
"learning_rate": 1.3327205882352942e-06,
|
| 1786 |
+
"loss": 0.7457,
|
| 1787 |
+
"step": 2890
|
| 1788 |
+
},
|
| 1789 |
+
{
|
| 1790 |
+
"epoch": 7.3,
|
| 1791 |
+
"learning_rate": 1.2867647058823528e-06,
|
| 1792 |
+
"loss": 0.7397,
|
| 1793 |
+
"step": 2900
|
| 1794 |
+
},
|
| 1795 |
+
{
|
| 1796 |
+
"epoch": 7.32,
|
| 1797 |
+
"learning_rate": 1.2408088235294119e-06,
|
| 1798 |
+
"loss": 0.7498,
|
| 1799 |
+
"step": 2910
|
| 1800 |
+
},
|
| 1801 |
+
{
|
| 1802 |
+
"epoch": 7.35,
|
| 1803 |
+
"learning_rate": 1.1948529411764707e-06,
|
| 1804 |
+
"loss": 0.7477,
|
| 1805 |
+
"step": 2920
|
| 1806 |
+
},
|
| 1807 |
+
{
|
| 1808 |
+
"epoch": 7.37,
|
| 1809 |
+
"learning_rate": 1.1488970588235295e-06,
|
| 1810 |
+
"loss": 0.7433,
|
| 1811 |
+
"step": 2930
|
| 1812 |
+
},
|
| 1813 |
+
{
|
| 1814 |
+
"epoch": 7.4,
|
| 1815 |
+
"learning_rate": 1.1029411764705884e-06,
|
| 1816 |
+
"loss": 0.743,
|
| 1817 |
+
"step": 2940
|
| 1818 |
+
},
|
| 1819 |
+
{
|
| 1820 |
+
"epoch": 7.42,
|
| 1821 |
+
"learning_rate": 1.0569852941176472e-06,
|
| 1822 |
+
"loss": 0.7393,
|
| 1823 |
+
"step": 2950
|
| 1824 |
+
},
|
| 1825 |
+
{
|
| 1826 |
+
"epoch": 7.45,
|
| 1827 |
+
"learning_rate": 1.011029411764706e-06,
|
| 1828 |
+
"loss": 0.7483,
|
| 1829 |
+
"step": 2960
|
| 1830 |
+
},
|
| 1831 |
+
{
|
| 1832 |
+
"epoch": 7.47,
|
| 1833 |
+
"learning_rate": 9.650735294117648e-07,
|
| 1834 |
+
"loss": 0.7466,
|
| 1835 |
+
"step": 2970
|
| 1836 |
+
},
|
| 1837 |
+
{
|
| 1838 |
+
"epoch": 7.5,
|
| 1839 |
+
"learning_rate": 9.191176470588237e-07,
|
| 1840 |
+
"loss": 0.7509,
|
| 1841 |
+
"step": 2980
|
| 1842 |
+
},
|
| 1843 |
+
{
|
| 1844 |
+
"epoch": 7.53,
|
| 1845 |
+
"learning_rate": 8.731617647058824e-07,
|
| 1846 |
+
"loss": 0.7554,
|
| 1847 |
+
"step": 2990
|
| 1848 |
+
},
|
| 1849 |
+
{
|
| 1850 |
+
"epoch": 7.55,
|
| 1851 |
+
"learning_rate": 8.272058823529412e-07,
|
| 1852 |
+
"loss": 0.7326,
|
| 1853 |
+
"step": 3000
|
| 1854 |
+
},
|
| 1855 |
+
{
|
| 1856 |
+
"epoch": 7.55,
|
| 1857 |
+
"eval_code_evol_accuracy": 0.759983871353542,
|
| 1858 |
+
"eval_code_evol_loss": 0.90576171875,
|
| 1859 |
+
"eval_code_evol_runtime": 10.3923,
|
| 1860 |
+
"eval_code_evol_samples_per_second": 74.19,
|
| 1861 |
+
"eval_code_evol_steps_per_second": 9.334,
|
| 1862 |
+
"step": 3000
|
| 1863 |
+
}
|
| 1864 |
+
],
|
| 1865 |
+
"max_steps": 3176,
|
| 1866 |
+
"num_train_epochs": 8,
|
| 1867 |
+
"total_flos": 3.32565039055215e+18,
|
| 1868 |
+
"trial_name": null,
|
| 1869 |
+
"trial_params": null
|
| 1870 |
+
}
|
vocab.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|