Upload 14 files
Browse files- config.json +39 -0
- generation_config.json +6 -0
- merges.txt +0 -0
- model.safetensors +3 -0
- optimizer.pt +3 -0
- rng_state_0.pth +3 -0
- rng_state_1.pth +3 -0
- scheduler.pt +3 -0
- special_tokens_map.json +6 -0
- tokenizer.json +0 -0
- tokenizer_config.json +20 -0
- trainer_state.json +1120 -0
- training_args.bin +3 -0
- vocab.json +0 -0
config.json
ADDED
|
@@ -0,0 +1,39 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "gpt2",
|
| 3 |
+
"activation_function": "gelu_new",
|
| 4 |
+
"architectures": [
|
| 5 |
+
"GPT2LMHeadModel"
|
| 6 |
+
],
|
| 7 |
+
"attn_pdrop": 0.1,
|
| 8 |
+
"bos_token_id": 50256,
|
| 9 |
+
"embd_pdrop": 0.1,
|
| 10 |
+
"eos_token_id": 50256,
|
| 11 |
+
"initializer_range": 0.02,
|
| 12 |
+
"layer_norm_epsilon": 1e-05,
|
| 13 |
+
"model_type": "gpt2",
|
| 14 |
+
"n_ctx": 1024,
|
| 15 |
+
"n_embd": 768,
|
| 16 |
+
"n_head": 12,
|
| 17 |
+
"n_inner": null,
|
| 18 |
+
"n_layer": 12,
|
| 19 |
+
"n_positions": 1024,
|
| 20 |
+
"reorder_and_upcast_attn": false,
|
| 21 |
+
"resid_pdrop": 0.1,
|
| 22 |
+
"scale_attn_by_inverse_layer_idx": false,
|
| 23 |
+
"scale_attn_weights": true,
|
| 24 |
+
"summary_activation": null,
|
| 25 |
+
"summary_first_dropout": 0.1,
|
| 26 |
+
"summary_proj_to_labels": true,
|
| 27 |
+
"summary_type": "cls_index",
|
| 28 |
+
"summary_use_proj": true,
|
| 29 |
+
"task_specific_params": {
|
| 30 |
+
"text-generation": {
|
| 31 |
+
"do_sample": true,
|
| 32 |
+
"max_length": 50
|
| 33 |
+
}
|
| 34 |
+
},
|
| 35 |
+
"torch_dtype": "bfloat16",
|
| 36 |
+
"transformers_version": "4.44.0",
|
| 37 |
+
"use_cache": false,
|
| 38 |
+
"vocab_size": 50257
|
| 39 |
+
}
|
generation_config.json
ADDED
|
@@ -0,0 +1,6 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_from_model_config": true,
|
| 3 |
+
"bos_token_id": 50256,
|
| 4 |
+
"eos_token_id": 50256,
|
| 5 |
+
"transformers_version": "4.44.0"
|
| 6 |
+
}
|
merges.txt
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:2a30b88d30d67345a258fd9761e520f536e16182004547f94457019bd9de0ca4
|
| 3 |
+
size 248894656
|
optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e61d42fcedb165e8be6b425e58e8cc4e8c39510160a0145189a82fd0495f8edd
|
| 3 |
+
size 497883130
|
rng_state_0.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:4c2b07929833e1badbdedc6c47756c8d5a215033e2704c3b0d804e9c9251585a
|
| 3 |
+
size 14512
|
rng_state_1.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:b00a86c6c2df825422059224154af89d7cb348d29c92e796299ed6e84c4bed20
|
| 3 |
+
size 14512
|
scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a7b4ed91ee6407a1736a1a815a9fa65a855725949b1dcc41948ce631dda004cd
|
| 3 |
+
size 1064
|
special_tokens_map.json
ADDED
|
@@ -0,0 +1,6 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": "<|endoftext|>",
|
| 3 |
+
"eos_token": "<|endoftext|>",
|
| 4 |
+
"pad_token": "<|endoftext|>",
|
| 5 |
+
"unk_token": "<|endoftext|>"
|
| 6 |
+
}
|
tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
tokenizer_config.json
ADDED
|
@@ -0,0 +1,20 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"add_prefix_space": false,
|
| 3 |
+
"added_tokens_decoder": {
|
| 4 |
+
"50256": {
|
| 5 |
+
"content": "<|endoftext|>",
|
| 6 |
+
"lstrip": false,
|
| 7 |
+
"normalized": true,
|
| 8 |
+
"rstrip": false,
|
| 9 |
+
"single_word": false,
|
| 10 |
+
"special": true
|
| 11 |
+
}
|
| 12 |
+
},
|
| 13 |
+
"bos_token": "<|endoftext|>",
|
| 14 |
+
"clean_up_tokenization_spaces": true,
|
| 15 |
+
"eos_token": "<|endoftext|>",
|
| 16 |
+
"model_max_length": 1024,
|
| 17 |
+
"pad_token": "<|endoftext|>",
|
| 18 |
+
"tokenizer_class": "GPT2Tokenizer",
|
| 19 |
+
"unk_token": "<|endoftext|>"
|
| 20 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,1120 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 0.39151499991843436,
|
| 5 |
+
"eval_steps": 1000,
|
| 6 |
+
"global_step": 72000,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 5.437708332200477e-06,
|
| 13 |
+
"grad_norm": 6.4375,
|
| 14 |
+
"learning_rate": 1.4814814814814816e-08,
|
| 15 |
+
"loss": 5.5898,
|
| 16 |
+
"step": 1
|
| 17 |
+
},
|
| 18 |
+
{
|
| 19 |
+
"epoch": 0.0054377083322004775,
|
| 20 |
+
"grad_norm": 3.171875,
|
| 21 |
+
"learning_rate": 1.4814814814814815e-05,
|
| 22 |
+
"loss": 5.4295,
|
| 23 |
+
"step": 1000
|
| 24 |
+
},
|
| 25 |
+
{
|
| 26 |
+
"epoch": 0.0054377083322004775,
|
| 27 |
+
"eval_loss": 5.008066177368164,
|
| 28 |
+
"eval_runtime": 15451.3463,
|
| 29 |
+
"eval_samples_per_second": 285.654,
|
| 30 |
+
"eval_steps_per_second": 2.976,
|
| 31 |
+
"step": 1000
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.010875416664400955,
|
| 35 |
+
"grad_norm": 1.171875,
|
| 36 |
+
"learning_rate": 2.962962962962963e-05,
|
| 37 |
+
"loss": 4.7755,
|
| 38 |
+
"step": 2000
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.010875416664400955,
|
| 42 |
+
"eval_loss": 4.375054836273193,
|
| 43 |
+
"eval_runtime": 15330.551,
|
| 44 |
+
"eval_samples_per_second": 287.905,
|
| 45 |
+
"eval_steps_per_second": 2.999,
|
| 46 |
+
"step": 2000
|
| 47 |
+
},
|
| 48 |
+
{
|
| 49 |
+
"epoch": 0.01631312499660143,
|
| 50 |
+
"grad_norm": 1.109375,
|
| 51 |
+
"learning_rate": 4.4444444444444447e-05,
|
| 52 |
+
"loss": 4.4478,
|
| 53 |
+
"step": 3000
|
| 54 |
+
},
|
| 55 |
+
{
|
| 56 |
+
"epoch": 0.01631312499660143,
|
| 57 |
+
"eval_loss": 4.167980194091797,
|
| 58 |
+
"eval_runtime": 15320.5031,
|
| 59 |
+
"eval_samples_per_second": 288.094,
|
| 60 |
+
"eval_steps_per_second": 3.001,
|
| 61 |
+
"step": 3000
|
| 62 |
+
},
|
| 63 |
+
{
|
| 64 |
+
"epoch": 0.02175083332880191,
|
| 65 |
+
"grad_norm": 1.1015625,
|
| 66 |
+
"learning_rate": 5.925925925925926e-05,
|
| 67 |
+
"loss": 4.2589,
|
| 68 |
+
"step": 4000
|
| 69 |
+
},
|
| 70 |
+
{
|
| 71 |
+
"epoch": 0.02175083332880191,
|
| 72 |
+
"eval_loss": 3.97269344329834,
|
| 73 |
+
"eval_runtime": 15323.7852,
|
| 74 |
+
"eval_samples_per_second": 288.032,
|
| 75 |
+
"eval_steps_per_second": 3.0,
|
| 76 |
+
"step": 4000
|
| 77 |
+
},
|
| 78 |
+
{
|
| 79 |
+
"epoch": 0.02718854166100239,
|
| 80 |
+
"grad_norm": 0.9609375,
|
| 81 |
+
"learning_rate": 7.407407407407407e-05,
|
| 82 |
+
"loss": 4.0776,
|
| 83 |
+
"step": 5000
|
| 84 |
+
},
|
| 85 |
+
{
|
| 86 |
+
"epoch": 0.02718854166100239,
|
| 87 |
+
"eval_loss": 3.7941160202026367,
|
| 88 |
+
"eval_runtime": 15358.3554,
|
| 89 |
+
"eval_samples_per_second": 287.384,
|
| 90 |
+
"eval_steps_per_second": 2.994,
|
| 91 |
+
"step": 5000
|
| 92 |
+
},
|
| 93 |
+
{
|
| 94 |
+
"epoch": 0.03262624999320286,
|
| 95 |
+
"grad_norm": 0.9140625,
|
| 96 |
+
"learning_rate": 8.888888888888889e-05,
|
| 97 |
+
"loss": 3.9259,
|
| 98 |
+
"step": 6000
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 0.03262624999320286,
|
| 102 |
+
"eval_loss": 3.658724308013916,
|
| 103 |
+
"eval_runtime": 15348.9884,
|
| 104 |
+
"eval_samples_per_second": 287.559,
|
| 105 |
+
"eval_steps_per_second": 2.995,
|
| 106 |
+
"step": 6000
|
| 107 |
+
},
|
| 108 |
+
{
|
| 109 |
+
"epoch": 0.038063958325403345,
|
| 110 |
+
"grad_norm": 0.9296875,
|
| 111 |
+
"learning_rate": 0.0001037037037037037,
|
| 112 |
+
"loss": 3.8025,
|
| 113 |
+
"step": 7000
|
| 114 |
+
},
|
| 115 |
+
{
|
| 116 |
+
"epoch": 0.038063958325403345,
|
| 117 |
+
"eval_loss": 3.5507161617279053,
|
| 118 |
+
"eval_runtime": 15346.7443,
|
| 119 |
+
"eval_samples_per_second": 287.601,
|
| 120 |
+
"eval_steps_per_second": 2.996,
|
| 121 |
+
"step": 7000
|
| 122 |
+
},
|
| 123 |
+
{
|
| 124 |
+
"epoch": 0.04350166665760382,
|
| 125 |
+
"grad_norm": 0.93359375,
|
| 126 |
+
"learning_rate": 0.00011851851851851852,
|
| 127 |
+
"loss": 3.7003,
|
| 128 |
+
"step": 8000
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.04350166665760382,
|
| 132 |
+
"eval_loss": 3.462174892425537,
|
| 133 |
+
"eval_runtime": 15363.0936,
|
| 134 |
+
"eval_samples_per_second": 287.295,
|
| 135 |
+
"eval_steps_per_second": 2.993,
|
| 136 |
+
"step": 8000
|
| 137 |
+
},
|
| 138 |
+
{
|
| 139 |
+
"epoch": 0.048939374989804295,
|
| 140 |
+
"grad_norm": 0.97265625,
|
| 141 |
+
"learning_rate": 0.00013333333333333334,
|
| 142 |
+
"loss": 3.6129,
|
| 143 |
+
"step": 9000
|
| 144 |
+
},
|
| 145 |
+
{
|
| 146 |
+
"epoch": 0.048939374989804295,
|
| 147 |
+
"eval_loss": 3.3831520080566406,
|
| 148 |
+
"eval_runtime": 15315.7808,
|
| 149 |
+
"eval_samples_per_second": 288.183,
|
| 150 |
+
"eval_steps_per_second": 3.002,
|
| 151 |
+
"step": 9000
|
| 152 |
+
},
|
| 153 |
+
{
|
| 154 |
+
"epoch": 0.05437708332200478,
|
| 155 |
+
"grad_norm": 0.921875,
|
| 156 |
+
"learning_rate": 0.00014814814814814815,
|
| 157 |
+
"loss": 3.5335,
|
| 158 |
+
"step": 10000
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 0.05437708332200478,
|
| 162 |
+
"eval_loss": 3.315824508666992,
|
| 163 |
+
"eval_runtime": 15320.6933,
|
| 164 |
+
"eval_samples_per_second": 288.09,
|
| 165 |
+
"eval_steps_per_second": 3.001,
|
| 166 |
+
"step": 10000
|
| 167 |
+
},
|
| 168 |
+
{
|
| 169 |
+
"epoch": 0.05981479165420525,
|
| 170 |
+
"grad_norm": 0.8828125,
|
| 171 |
+
"learning_rate": 0.00016296296296296295,
|
| 172 |
+
"loss": 3.4639,
|
| 173 |
+
"step": 11000
|
| 174 |
+
},
|
| 175 |
+
{
|
| 176 |
+
"epoch": 0.05981479165420525,
|
| 177 |
+
"eval_loss": 3.2545006275177,
|
| 178 |
+
"eval_runtime": 15357.7315,
|
| 179 |
+
"eval_samples_per_second": 287.396,
|
| 180 |
+
"eval_steps_per_second": 2.994,
|
| 181 |
+
"step": 11000
|
| 182 |
+
},
|
| 183 |
+
{
|
| 184 |
+
"epoch": 0.06525249998640573,
|
| 185 |
+
"grad_norm": 0.9296875,
|
| 186 |
+
"learning_rate": 0.00017777777777777779,
|
| 187 |
+
"loss": 3.403,
|
| 188 |
+
"step": 12000
|
| 189 |
+
},
|
| 190 |
+
{
|
| 191 |
+
"epoch": 0.06525249998640573,
|
| 192 |
+
"eval_loss": 3.198941230773926,
|
| 193 |
+
"eval_runtime": 15351.0761,
|
| 194 |
+
"eval_samples_per_second": 287.52,
|
| 195 |
+
"eval_steps_per_second": 2.995,
|
| 196 |
+
"step": 12000
|
| 197 |
+
},
|
| 198 |
+
{
|
| 199 |
+
"epoch": 0.0706902083186062,
|
| 200 |
+
"grad_norm": 0.953125,
|
| 201 |
+
"learning_rate": 0.0001925925925925926,
|
| 202 |
+
"loss": 3.3476,
|
| 203 |
+
"step": 13000
|
| 204 |
+
},
|
| 205 |
+
{
|
| 206 |
+
"epoch": 0.0706902083186062,
|
| 207 |
+
"eval_loss": 3.1511921882629395,
|
| 208 |
+
"eval_runtime": 15356.3177,
|
| 209 |
+
"eval_samples_per_second": 287.422,
|
| 210 |
+
"eval_steps_per_second": 2.994,
|
| 211 |
+
"step": 13000
|
| 212 |
+
},
|
| 213 |
+
{
|
| 214 |
+
"epoch": 0.07612791665080669,
|
| 215 |
+
"grad_norm": 0.97265625,
|
| 216 |
+
"learning_rate": 0.00019999164298554375,
|
| 217 |
+
"loss": 3.2984,
|
| 218 |
+
"step": 14000
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 0.07612791665080669,
|
| 222 |
+
"eval_loss": 3.1090052127838135,
|
| 223 |
+
"eval_runtime": 15355.9903,
|
| 224 |
+
"eval_samples_per_second": 287.428,
|
| 225 |
+
"eval_steps_per_second": 2.994,
|
| 226 |
+
"step": 14000
|
| 227 |
+
},
|
| 228 |
+
{
|
| 229 |
+
"epoch": 0.08156562498300717,
|
| 230 |
+
"grad_norm": 0.90625,
|
| 231 |
+
"learning_rate": 0.00019992479525042303,
|
| 232 |
+
"loss": 3.261,
|
| 233 |
+
"step": 15000
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"epoch": 0.08156562498300717,
|
| 237 |
+
"eval_loss": 3.0757596492767334,
|
| 238 |
+
"eval_runtime": 15336.3046,
|
| 239 |
+
"eval_samples_per_second": 287.797,
|
| 240 |
+
"eval_steps_per_second": 2.998,
|
| 241 |
+
"step": 15000
|
| 242 |
+
},
|
| 243 |
+
{
|
| 244 |
+
"epoch": 0.08700333331520764,
|
| 245 |
+
"grad_norm": 0.90234375,
|
| 246 |
+
"learning_rate": 0.00019979114447011323,
|
| 247 |
+
"loss": 3.2317,
|
| 248 |
+
"step": 16000
|
| 249 |
+
},
|
| 250 |
+
{
|
| 251 |
+
"epoch": 0.08700333331520764,
|
| 252 |
+
"eval_loss": 3.053584337234497,
|
| 253 |
+
"eval_runtime": 15330.754,
|
| 254 |
+
"eval_samples_per_second": 287.901,
|
| 255 |
+
"eval_steps_per_second": 2.999,
|
| 256 |
+
"step": 16000
|
| 257 |
+
},
|
| 258 |
+
{
|
| 259 |
+
"epoch": 0.09244104164740811,
|
| 260 |
+
"grad_norm": 1.015625,
|
| 261 |
+
"learning_rate": 0.00019959077999460095,
|
| 262 |
+
"loss": 3.2088,
|
| 263 |
+
"step": 17000
|
| 264 |
+
},
|
| 265 |
+
{
|
| 266 |
+
"epoch": 0.09244104164740811,
|
| 267 |
+
"eval_loss": 3.0358426570892334,
|
| 268 |
+
"eval_runtime": 15337.2484,
|
| 269 |
+
"eval_samples_per_second": 287.779,
|
| 270 |
+
"eval_steps_per_second": 2.998,
|
| 271 |
+
"step": 17000
|
| 272 |
+
},
|
| 273 |
+
{
|
| 274 |
+
"epoch": 0.09787874997960859,
|
| 275 |
+
"grad_norm": 1.0390625,
|
| 276 |
+
"learning_rate": 0.00019932383577419432,
|
| 277 |
+
"loss": 3.1919,
|
| 278 |
+
"step": 18000
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 0.09787874997960859,
|
| 282 |
+
"eval_loss": 3.020881175994873,
|
| 283 |
+
"eval_runtime": 15358.1769,
|
| 284 |
+
"eval_samples_per_second": 287.387,
|
| 285 |
+
"eval_steps_per_second": 2.994,
|
| 286 |
+
"step": 18000
|
| 287 |
+
},
|
| 288 |
+
{
|
| 289 |
+
"epoch": 0.10331645831180906,
|
| 290 |
+
"grad_norm": 0.96875,
|
| 291 |
+
"learning_rate": 0.00019899049026997272,
|
| 292 |
+
"loss": 3.1807,
|
| 293 |
+
"step": 19000
|
| 294 |
+
},
|
| 295 |
+
{
|
| 296 |
+
"epoch": 0.10331645831180906,
|
| 297 |
+
"eval_loss": 3.0086727142333984,
|
| 298 |
+
"eval_runtime": 15317.1842,
|
| 299 |
+
"eval_samples_per_second": 288.156,
|
| 300 |
+
"eval_steps_per_second": 3.002,
|
| 301 |
+
"step": 19000
|
| 302 |
+
},
|
| 303 |
+
{
|
| 304 |
+
"epoch": 0.10875416664400955,
|
| 305 |
+
"grad_norm": 0.9765625,
|
| 306 |
+
"learning_rate": 0.00019859096633447965,
|
| 307 |
+
"loss": 3.1692,
|
| 308 |
+
"step": 20000
|
| 309 |
+
},
|
| 310 |
+
{
|
| 311 |
+
"epoch": 0.10875416664400955,
|
| 312 |
+
"eval_loss": 2.9984583854675293,
|
| 313 |
+
"eval_runtime": 15316.5504,
|
| 314 |
+
"eval_samples_per_second": 288.168,
|
| 315 |
+
"eval_steps_per_second": 3.002,
|
| 316 |
+
"step": 20000
|
| 317 |
+
},
|
| 318 |
+
{
|
| 319 |
+
"epoch": 0.11419187497621003,
|
| 320 |
+
"grad_norm": 0.98828125,
|
| 321 |
+
"learning_rate": 0.00019812553106273847,
|
| 322 |
+
"loss": 3.1605,
|
| 323 |
+
"step": 21000
|
| 324 |
+
},
|
| 325 |
+
{
|
| 326 |
+
"epoch": 0.11419187497621003,
|
| 327 |
+
"eval_loss": 2.988966226577759,
|
| 328 |
+
"eval_runtime": 15327.5134,
|
| 329 |
+
"eval_samples_per_second": 287.962,
|
| 330 |
+
"eval_steps_per_second": 3.0,
|
| 331 |
+
"step": 21000
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 0.1196295833084105,
|
| 335 |
+
"grad_norm": 0.9921875,
|
| 336 |
+
"learning_rate": 0.00019759449561369038,
|
| 337 |
+
"loss": 3.1518,
|
| 338 |
+
"step": 22000
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 0.1196295833084105,
|
| 342 |
+
"eval_loss": 2.9820005893707275,
|
| 343 |
+
"eval_runtime": 15329.581,
|
| 344 |
+
"eval_samples_per_second": 287.923,
|
| 345 |
+
"eval_steps_per_second": 2.999,
|
| 346 |
+
"step": 22000
|
| 347 |
+
},
|
| 348 |
+
{
|
| 349 |
+
"epoch": 0.12506729164061098,
|
| 350 |
+
"grad_norm": 0.984375,
|
| 351 |
+
"learning_rate": 0.00019699821500217434,
|
| 352 |
+
"loss": 3.1434,
|
| 353 |
+
"step": 23000
|
| 354 |
+
},
|
| 355 |
+
{
|
| 356 |
+
"epoch": 0.12506729164061098,
|
| 357 |
+
"eval_loss": 2.975614070892334,
|
| 358 |
+
"eval_runtime": 15361.0129,
|
| 359 |
+
"eval_samples_per_second": 287.334,
|
| 360 |
+
"eval_steps_per_second": 2.993,
|
| 361 |
+
"step": 23000
|
| 362 |
+
},
|
| 363 |
+
{
|
| 364 |
+
"epoch": 0.13050499997281145,
|
| 365 |
+
"grad_norm": 0.921875,
|
| 366 |
+
"learning_rate": 0.00019633708786158806,
|
| 367 |
+
"loss": 3.1373,
|
| 368 |
+
"step": 24000
|
| 369 |
+
},
|
| 370 |
+
{
|
| 371 |
+
"epoch": 0.13050499997281145,
|
| 372 |
+
"eval_loss": 2.969208240509033,
|
| 373 |
+
"eval_runtime": 15340.5185,
|
| 374 |
+
"eval_samples_per_second": 287.718,
|
| 375 |
+
"eval_steps_per_second": 2.997,
|
| 376 |
+
"step": 24000
|
| 377 |
+
},
|
| 378 |
+
{
|
| 379 |
+
"epoch": 0.13594270830501193,
|
| 380 |
+
"grad_norm": 0.9453125,
|
| 381 |
+
"learning_rate": 0.00019561155617738797,
|
| 382 |
+
"loss": 3.1332,
|
| 383 |
+
"step": 25000
|
| 384 |
+
},
|
| 385 |
+
{
|
| 386 |
+
"epoch": 0.13594270830501193,
|
| 387 |
+
"eval_loss": 2.9637811183929443,
|
| 388 |
+
"eval_runtime": 15340.5055,
|
| 389 |
+
"eval_samples_per_second": 287.718,
|
| 390 |
+
"eval_steps_per_second": 2.997,
|
| 391 |
+
"step": 25000
|
| 392 |
+
},
|
| 393 |
+
{
|
| 394 |
+
"epoch": 0.1413804166372124,
|
| 395 |
+
"grad_norm": 0.94921875,
|
| 396 |
+
"learning_rate": 0.00019482210499160765,
|
| 397 |
+
"loss": 3.1267,
|
| 398 |
+
"step": 26000
|
| 399 |
+
},
|
| 400 |
+
{
|
| 401 |
+
"epoch": 0.1413804166372124,
|
| 402 |
+
"eval_loss": 2.96110463142395,
|
| 403 |
+
"eval_runtime": 15350.4187,
|
| 404 |
+
"eval_samples_per_second": 287.532,
|
| 405 |
+
"eval_steps_per_second": 2.995,
|
| 406 |
+
"step": 26000
|
| 407 |
+
},
|
| 408 |
+
{
|
| 409 |
+
"epoch": 0.14681812496941288,
|
| 410 |
+
"grad_norm": 0.91796875,
|
| 411 |
+
"learning_rate": 0.00019396926207859084,
|
| 412 |
+
"loss": 3.1237,
|
| 413 |
+
"step": 27000
|
| 414 |
+
},
|
| 415 |
+
{
|
| 416 |
+
"epoch": 0.14681812496941288,
|
| 417 |
+
"eval_loss": 2.9555680751800537,
|
| 418 |
+
"eval_runtime": 15320.143,
|
| 419 |
+
"eval_samples_per_second": 288.101,
|
| 420 |
+
"eval_steps_per_second": 3.001,
|
| 421 |
+
"step": 27000
|
| 422 |
+
},
|
| 423 |
+
{
|
| 424 |
+
"epoch": 0.15225583330161338,
|
| 425 |
+
"grad_norm": 0.91015625,
|
| 426 |
+
"learning_rate": 0.00019305359759215685,
|
| 427 |
+
"loss": 3.1209,
|
| 428 |
+
"step": 28000
|
| 429 |
+
},
|
| 430 |
+
{
|
| 431 |
+
"epoch": 0.15225583330161338,
|
| 432 |
+
"eval_loss": 2.953141689300537,
|
| 433 |
+
"eval_runtime": 15342.6298,
|
| 434 |
+
"eval_samples_per_second": 287.678,
|
| 435 |
+
"eval_steps_per_second": 2.997,
|
| 436 |
+
"step": 28000
|
| 437 |
+
},
|
| 438 |
+
{
|
| 439 |
+
"epoch": 0.15769354163381386,
|
| 440 |
+
"grad_norm": 0.87890625,
|
| 441 |
+
"learning_rate": 0.00019207572368443385,
|
| 442 |
+
"loss": 3.1158,
|
| 443 |
+
"step": 29000
|
| 444 |
+
},
|
| 445 |
+
{
|
| 446 |
+
"epoch": 0.15769354163381386,
|
| 447 |
+
"eval_loss": 2.9493179321289062,
|
| 448 |
+
"eval_runtime": 15348.9479,
|
| 449 |
+
"eval_samples_per_second": 287.56,
|
| 450 |
+
"eval_steps_per_second": 2.995,
|
| 451 |
+
"step": 29000
|
| 452 |
+
},
|
| 453 |
+
{
|
| 454 |
+
"epoch": 0.16313124996601433,
|
| 455 |
+
"grad_norm": 1.0390625,
|
| 456 |
+
"learning_rate": 0.0001910362940966147,
|
| 457 |
+
"loss": 3.1128,
|
| 458 |
+
"step": 30000
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 0.16313124996601433,
|
| 462 |
+
"eval_loss": 2.946991443634033,
|
| 463 |
+
"eval_runtime": 15312.5285,
|
| 464 |
+
"eval_samples_per_second": 288.244,
|
| 465 |
+
"eval_steps_per_second": 3.003,
|
| 466 |
+
"step": 30000
|
| 467 |
+
},
|
| 468 |
+
{
|
| 469 |
+
"epoch": 0.1685689582982148,
|
| 470 |
+
"grad_norm": 0.96875,
|
| 471 |
+
"learning_rate": 0.00018993600372190932,
|
| 472 |
+
"loss": 3.1098,
|
| 473 |
+
"step": 31000
|
| 474 |
+
},
|
| 475 |
+
{
|
| 476 |
+
"epoch": 0.1685689582982148,
|
| 477 |
+
"eval_loss": 2.9436445236206055,
|
| 478 |
+
"eval_runtime": 15349.4672,
|
| 479 |
+
"eval_samples_per_second": 287.55,
|
| 480 |
+
"eval_steps_per_second": 2.995,
|
| 481 |
+
"step": 31000
|
| 482 |
+
},
|
| 483 |
+
{
|
| 484 |
+
"epoch": 0.17400666663041528,
|
| 485 |
+
"grad_norm": 0.99609375,
|
| 486 |
+
"learning_rate": 0.00018877558814098561,
|
| 487 |
+
"loss": 3.1083,
|
| 488 |
+
"step": 32000
|
| 489 |
+
},
|
| 490 |
+
{
|
| 491 |
+
"epoch": 0.17400666663041528,
|
| 492 |
+
"eval_loss": 2.9420006275177,
|
| 493 |
+
"eval_runtime": 15345.3976,
|
| 494 |
+
"eval_samples_per_second": 287.626,
|
| 495 |
+
"eval_steps_per_second": 2.996,
|
| 496 |
+
"step": 32000
|
| 497 |
+
},
|
| 498 |
+
{
|
| 499 |
+
"epoch": 0.17944437496261575,
|
| 500 |
+
"grad_norm": 0.97265625,
|
| 501 |
+
"learning_rate": 0.0001875558231302091,
|
| 502 |
+
"loss": 3.1054,
|
| 503 |
+
"step": 33000
|
| 504 |
+
},
|
| 505 |
+
{
|
| 506 |
+
"epoch": 0.17944437496261575,
|
| 507 |
+
"eval_loss": 2.939058780670166,
|
| 508 |
+
"eval_runtime": 15316.1274,
|
| 509 |
+
"eval_samples_per_second": 288.176,
|
| 510 |
+
"eval_steps_per_second": 3.002,
|
| 511 |
+
"step": 33000
|
| 512 |
+
},
|
| 513 |
+
{
|
| 514 |
+
"epoch": 0.18488208329481623,
|
| 515 |
+
"grad_norm": 0.921875,
|
| 516 |
+
"learning_rate": 0.00018627752414301086,
|
| 517 |
+
"loss": 3.1059,
|
| 518 |
+
"step": 34000
|
| 519 |
+
},
|
| 520 |
+
{
|
| 521 |
+
"epoch": 0.18488208329481623,
|
| 522 |
+
"eval_loss": 2.9374189376831055,
|
| 523 |
+
"eval_runtime": 15343.22,
|
| 524 |
+
"eval_samples_per_second": 287.667,
|
| 525 |
+
"eval_steps_per_second": 2.997,
|
| 526 |
+
"step": 34000
|
| 527 |
+
},
|
| 528 |
+
{
|
| 529 |
+
"epoch": 0.1903197916270167,
|
| 530 |
+
"grad_norm": 0.94140625,
|
| 531 |
+
"learning_rate": 0.00018494154576472976,
|
| 532 |
+
"loss": 3.1028,
|
| 533 |
+
"step": 35000
|
| 534 |
+
},
|
| 535 |
+
{
|
| 536 |
+
"epoch": 0.1903197916270167,
|
| 537 |
+
"eval_loss": 2.9355897903442383,
|
| 538 |
+
"eval_runtime": 15340.0011,
|
| 539 |
+
"eval_samples_per_second": 287.728,
|
| 540 |
+
"eval_steps_per_second": 2.997,
|
| 541 |
+
"step": 35000
|
| 542 |
+
},
|
| 543 |
+
{
|
| 544 |
+
"epoch": 0.19575749995921718,
|
| 545 |
+
"grad_norm": 0.8984375,
|
| 546 |
+
"learning_rate": 0.00018354878114129367,
|
| 547 |
+
"loss": 3.0992,
|
| 548 |
+
"step": 36000
|
| 549 |
+
},
|
| 550 |
+
{
|
| 551 |
+
"epoch": 0.19575749995921718,
|
| 552 |
+
"eval_loss": 2.9352223873138428,
|
| 553 |
+
"eval_runtime": 15353.667,
|
| 554 |
+
"eval_samples_per_second": 287.472,
|
| 555 |
+
"eval_steps_per_second": 2.995,
|
| 556 |
+
"step": 36000
|
| 557 |
+
},
|
| 558 |
+
{
|
| 559 |
+
"epoch": 0.20119520829141765,
|
| 560 |
+
"grad_norm": 0.9375,
|
| 561 |
+
"learning_rate": 0.00018210016138212187,
|
| 562 |
+
"loss": 3.0988,
|
| 563 |
+
"step": 37000
|
| 564 |
+
},
|
| 565 |
+
{
|
| 566 |
+
"epoch": 0.20119520829141765,
|
| 567 |
+
"eval_loss": 2.9330756664276123,
|
| 568 |
+
"eval_runtime": 15356.9704,
|
| 569 |
+
"eval_samples_per_second": 287.41,
|
| 570 |
+
"eval_steps_per_second": 2.994,
|
| 571 |
+
"step": 37000
|
| 572 |
+
},
|
| 573 |
+
{
|
| 574 |
+
"epoch": 0.20663291662361813,
|
| 575 |
+
"grad_norm": 0.95703125,
|
| 576 |
+
"learning_rate": 0.00018059665493764743,
|
| 577 |
+
"loss": 3.0977,
|
| 578 |
+
"step": 38000
|
| 579 |
+
},
|
| 580 |
+
{
|
| 581 |
+
"epoch": 0.20663291662361813,
|
| 582 |
+
"eval_loss": 2.9324018955230713,
|
| 583 |
+
"eval_runtime": 15340.1741,
|
| 584 |
+
"eval_samples_per_second": 287.724,
|
| 585 |
+
"eval_steps_per_second": 2.997,
|
| 586 |
+
"step": 38000
|
| 587 |
+
},
|
| 588 |
+
{
|
| 589 |
+
"epoch": 0.21207062495581863,
|
| 590 |
+
"grad_norm": 0.9140625,
|
| 591 |
+
"learning_rate": 0.00017903926695187595,
|
| 592 |
+
"loss": 3.0981,
|
| 593 |
+
"step": 39000
|
| 594 |
+
},
|
| 595 |
+
{
|
| 596 |
+
"epoch": 0.21207062495581863,
|
| 597 |
+
"eval_loss": 2.929825782775879,
|
| 598 |
+
"eval_runtime": 15335.7197,
|
| 599 |
+
"eval_samples_per_second": 287.808,
|
| 600 |
+
"eval_steps_per_second": 2.998,
|
| 601 |
+
"step": 39000
|
| 602 |
+
},
|
| 603 |
+
{
|
| 604 |
+
"epoch": 0.2175083332880191,
|
| 605 |
+
"grad_norm": 0.921875,
|
| 606 |
+
"learning_rate": 0.00017742903859041325,
|
| 607 |
+
"loss": 3.0946,
|
| 608 |
+
"step": 40000
|
| 609 |
+
},
|
| 610 |
+
{
|
| 611 |
+
"epoch": 0.2175083332880191,
|
| 612 |
+
"eval_loss": 2.928624153137207,
|
| 613 |
+
"eval_runtime": 15333.0433,
|
| 614 |
+
"eval_samples_per_second": 287.858,
|
| 615 |
+
"eval_steps_per_second": 2.999,
|
| 616 |
+
"step": 40000
|
| 617 |
+
},
|
| 618 |
+
{
|
| 619 |
+
"epoch": 0.22294604162021958,
|
| 620 |
+
"grad_norm": 0.91796875,
|
| 621 |
+
"learning_rate": 0.0001757670463444118,
|
| 622 |
+
"loss": 3.0949,
|
| 623 |
+
"step": 41000
|
| 624 |
+
},
|
| 625 |
+
{
|
| 626 |
+
"epoch": 0.22294604162021958,
|
| 627 |
+
"eval_loss": 2.92832612991333,
|
| 628 |
+
"eval_runtime": 15336.6021,
|
| 629 |
+
"eval_samples_per_second": 287.791,
|
| 630 |
+
"eval_steps_per_second": 2.998,
|
| 631 |
+
"step": 41000
|
| 632 |
+
},
|
| 633 |
+
{
|
| 634 |
+
"epoch": 0.22838374995242006,
|
| 635 |
+
"grad_norm": 0.88671875,
|
| 636 |
+
"learning_rate": 0.00017405440131090048,
|
| 637 |
+
"loss": 3.0948,
|
| 638 |
+
"step": 42000
|
| 639 |
+
},
|
| 640 |
+
{
|
| 641 |
+
"epoch": 0.22838374995242006,
|
| 642 |
+
"eval_loss": 2.92707896232605,
|
| 643 |
+
"eval_runtime": 15338.3291,
|
| 644 |
+
"eval_samples_per_second": 287.759,
|
| 645 |
+
"eval_steps_per_second": 2.998,
|
| 646 |
+
"step": 42000
|
| 647 |
+
},
|
| 648 |
+
{
|
| 649 |
+
"epoch": 0.23382145828462053,
|
| 650 |
+
"grad_norm": 0.91796875,
|
| 651 |
+
"learning_rate": 0.00017229224844997928,
|
| 652 |
+
"loss": 3.0915,
|
| 653 |
+
"step": 43000
|
| 654 |
+
},
|
| 655 |
+
{
|
| 656 |
+
"epoch": 0.23382145828462053,
|
| 657 |
+
"eval_loss": 2.926419973373413,
|
| 658 |
+
"eval_runtime": 15325.6418,
|
| 659 |
+
"eval_samples_per_second": 287.997,
|
| 660 |
+
"eval_steps_per_second": 3.0,
|
| 661 |
+
"step": 43000
|
| 662 |
+
},
|
| 663 |
+
{
|
| 664 |
+
"epoch": 0.239259166616821,
|
| 665 |
+
"grad_norm": 0.92578125,
|
| 666 |
+
"learning_rate": 0.00017048176581937563,
|
| 667 |
+
"loss": 3.0938,
|
| 668 |
+
"step": 44000
|
| 669 |
+
},
|
| 670 |
+
{
|
| 671 |
+
"epoch": 0.239259166616821,
|
| 672 |
+
"eval_loss": 2.9256813526153564,
|
| 673 |
+
"eval_runtime": 15326.1114,
|
| 674 |
+
"eval_samples_per_second": 287.988,
|
| 675 |
+
"eval_steps_per_second": 3.0,
|
| 676 |
+
"step": 44000
|
| 677 |
+
},
|
| 678 |
+
{
|
| 679 |
+
"epoch": 0.24469687494902148,
|
| 680 |
+
"grad_norm": 1.0546875,
|
| 681 |
+
"learning_rate": 0.0001686241637868734,
|
| 682 |
+
"loss": 3.0922,
|
| 683 |
+
"step": 45000
|
| 684 |
+
},
|
| 685 |
+
{
|
| 686 |
+
"epoch": 0.24469687494902148,
|
| 687 |
+
"eval_loss": 2.9255247116088867,
|
| 688 |
+
"eval_runtime": 15321.0034,
|
| 689 |
+
"eval_samples_per_second": 288.084,
|
| 690 |
+
"eval_steps_per_second": 3.001,
|
| 691 |
+
"step": 45000
|
| 692 |
+
},
|
| 693 |
+
{
|
| 694 |
+
"epoch": 0.25013458328122196,
|
| 695 |
+
"grad_norm": 1.03125,
|
| 696 |
+
"learning_rate": 0.00016672068422114196,
|
| 697 |
+
"loss": 3.0918,
|
| 698 |
+
"step": 46000
|
| 699 |
+
},
|
| 700 |
+
{
|
| 701 |
+
"epoch": 0.25013458328122196,
|
| 702 |
+
"eval_loss": 2.9247536659240723,
|
| 703 |
+
"eval_runtime": 15325.8812,
|
| 704 |
+
"eval_samples_per_second": 287.993,
|
| 705 |
+
"eval_steps_per_second": 3.0,
|
| 706 |
+
"step": 46000
|
| 707 |
+
},
|
| 708 |
+
{
|
| 709 |
+
"epoch": 0.25557229161342243,
|
| 710 |
+
"grad_norm": 0.890625,
|
| 711 |
+
"learning_rate": 0.00016477259966150588,
|
| 712 |
+
"loss": 3.0908,
|
| 713 |
+
"step": 47000
|
| 714 |
+
},
|
| 715 |
+
{
|
| 716 |
+
"epoch": 0.25557229161342243,
|
| 717 |
+
"eval_loss": 2.9234766960144043,
|
| 718 |
+
"eval_runtime": 15335.6241,
|
| 719 |
+
"eval_samples_per_second": 287.81,
|
| 720 |
+
"eval_steps_per_second": 2.998,
|
| 721 |
+
"step": 47000
|
| 722 |
+
},
|
| 723 |
+
{
|
| 724 |
+
"epoch": 0.2610099999456229,
|
| 725 |
+
"grad_norm": 0.8828125,
|
| 726 |
+
"learning_rate": 0.00016278121246720987,
|
| 727 |
+
"loss": 3.088,
|
| 728 |
+
"step": 48000
|
| 729 |
+
},
|
| 730 |
+
{
|
| 731 |
+
"epoch": 0.2610099999456229,
|
| 732 |
+
"eval_loss": 2.9226396083831787,
|
| 733 |
+
"eval_runtime": 15328.3665,
|
| 734 |
+
"eval_samples_per_second": 287.946,
|
| 735 |
+
"eval_steps_per_second": 2.999,
|
| 736 |
+
"step": 48000
|
| 737 |
+
},
|
| 738 |
+
{
|
| 739 |
+
"epoch": 0.2664477082778234,
|
| 740 |
+
"grad_norm": 0.9296875,
|
| 741 |
+
"learning_rate": 0.00016074785394674837,
|
| 742 |
+
"loss": 3.0901,
|
| 743 |
+
"step": 49000
|
| 744 |
+
},
|
| 745 |
+
{
|
| 746 |
+
"epoch": 0.2664477082778234,
|
| 747 |
+
"eval_loss": 2.922677516937256,
|
| 748 |
+
"eval_runtime": 15342.8424,
|
| 749 |
+
"eval_samples_per_second": 287.674,
|
| 750 |
+
"eval_steps_per_second": 2.997,
|
| 751 |
+
"step": 49000
|
| 752 |
+
},
|
| 753 |
+
{
|
| 754 |
+
"epoch": 0.27188541661002386,
|
| 755 |
+
"grad_norm": 0.9296875,
|
| 756 |
+
"learning_rate": 0.0001586738834678418,
|
| 757 |
+
"loss": 3.0891,
|
| 758 |
+
"step": 50000
|
| 759 |
+
},
|
| 760 |
+
{
|
| 761 |
+
"epoch": 0.27188541661002386,
|
| 762 |
+
"eval_loss": 2.9226033687591553,
|
| 763 |
+
"eval_runtime": 15345.5247,
|
| 764 |
+
"eval_samples_per_second": 287.624,
|
| 765 |
+
"eval_steps_per_second": 2.996,
|
| 766 |
+
"step": 50000
|
| 767 |
+
},
|
| 768 |
+
{
|
| 769 |
+
"epoch": 0.27732312494222433,
|
| 770 |
+
"grad_norm": 0.83984375,
|
| 771 |
+
"learning_rate": 0.00015656068754865387,
|
| 772 |
+
"loss": 3.0909,
|
| 773 |
+
"step": 51000
|
| 774 |
+
},
|
| 775 |
+
{
|
| 776 |
+
"epoch": 0.27732312494222433,
|
| 777 |
+
"eval_loss": 2.92250919342041,
|
| 778 |
+
"eval_runtime": 15321.8908,
|
| 779 |
+
"eval_samples_per_second": 288.068,
|
| 780 |
+
"eval_steps_per_second": 3.001,
|
| 781 |
+
"step": 51000
|
| 782 |
+
},
|
| 783 |
+
{
|
| 784 |
+
"epoch": 0.2827608332744248,
|
| 785 |
+
"grad_norm": 0.90234375,
|
| 786 |
+
"learning_rate": 0.00015440967893085828,
|
| 787 |
+
"loss": 3.0892,
|
| 788 |
+
"step": 52000
|
| 789 |
+
},
|
| 790 |
+
{
|
| 791 |
+
"epoch": 0.2827608332744248,
|
| 792 |
+
"eval_loss": 2.9214928150177,
|
| 793 |
+
"eval_runtime": 15311.2597,
|
| 794 |
+
"eval_samples_per_second": 288.268,
|
| 795 |
+
"eval_steps_per_second": 3.003,
|
| 796 |
+
"step": 52000
|
| 797 |
+
},
|
| 798 |
+
{
|
| 799 |
+
"epoch": 0.2881985416066253,
|
| 800 |
+
"grad_norm": 0.9296875,
|
| 801 |
+
"learning_rate": 0.00015222229563517385,
|
| 802 |
+
"loss": 3.0877,
|
| 803 |
+
"step": 53000
|
| 804 |
+
},
|
| 805 |
+
{
|
| 806 |
+
"epoch": 0.2881985416066253,
|
| 807 |
+
"eval_loss": 2.921288013458252,
|
| 808 |
+
"eval_runtime": 15357.1564,
|
| 809 |
+
"eval_samples_per_second": 287.406,
|
| 810 |
+
"eval_steps_per_second": 2.994,
|
| 811 |
+
"step": 53000
|
| 812 |
+
},
|
| 813 |
+
{
|
| 814 |
+
"epoch": 0.29363624993882576,
|
| 815 |
+
"grad_norm": 0.97265625,
|
| 816 |
+
"learning_rate": 0.00015000000000000001,
|
| 817 |
+
"loss": 3.089,
|
| 818 |
+
"step": 54000
|
| 819 |
+
},
|
| 820 |
+
{
|
| 821 |
+
"epoch": 0.29363624993882576,
|
| 822 |
+
"eval_loss": 2.9219233989715576,
|
| 823 |
+
"eval_runtime": 15336.8542,
|
| 824 |
+
"eval_samples_per_second": 287.787,
|
| 825 |
+
"eval_steps_per_second": 2.998,
|
| 826 |
+
"step": 54000
|
| 827 |
+
},
|
| 828 |
+
{
|
| 829 |
+
"epoch": 0.2990739582710263,
|
| 830 |
+
"grad_norm": 0.91796875,
|
| 831 |
+
"learning_rate": 0.0001477442777037949,
|
| 832 |
+
"loss": 3.0887,
|
| 833 |
+
"step": 55000
|
| 834 |
+
},
|
| 835 |
+
{
|
| 836 |
+
"epoch": 0.2990739582710263,
|
| 837 |
+
"eval_loss": 2.9214131832122803,
|
| 838 |
+
"eval_runtime": 15359.1535,
|
| 839 |
+
"eval_samples_per_second": 287.369,
|
| 840 |
+
"eval_steps_per_second": 2.993,
|
| 841 |
+
"step": 55000
|
| 842 |
+
},
|
| 843 |
+
{
|
| 844 |
+
"epoch": 0.30451166660322676,
|
| 845 |
+
"grad_norm": 1.0390625,
|
| 846 |
+
"learning_rate": 0.00014545663677185006,
|
| 847 |
+
"loss": 3.0885,
|
| 848 |
+
"step": 56000
|
| 849 |
+
},
|
| 850 |
+
{
|
| 851 |
+
"epoch": 0.30451166660322676,
|
| 852 |
+
"eval_loss": 2.9214773178100586,
|
| 853 |
+
"eval_runtime": 15332.8462,
|
| 854 |
+
"eval_samples_per_second": 287.862,
|
| 855 |
+
"eval_steps_per_second": 2.999,
|
| 856 |
+
"step": 56000
|
| 857 |
+
},
|
| 858 |
+
{
|
| 859 |
+
"epoch": 0.30994937493542724,
|
| 860 |
+
"grad_norm": 0.91015625,
|
| 861 |
+
"learning_rate": 0.00014313860656812536,
|
| 862 |
+
"loss": 3.0854,
|
| 863 |
+
"step": 57000
|
| 864 |
+
},
|
| 865 |
+
{
|
| 866 |
+
"epoch": 0.30994937493542724,
|
| 867 |
+
"eval_loss": 2.9208450317382812,
|
| 868 |
+
"eval_runtime": 15325.5015,
|
| 869 |
+
"eval_samples_per_second": 288.0,
|
| 870 |
+
"eval_steps_per_second": 3.0,
|
| 871 |
+
"step": 57000
|
| 872 |
+
},
|
| 873 |
+
{
|
| 874 |
+
"epoch": 0.3153870832676277,
|
| 875 |
+
"grad_norm": 0.90234375,
|
| 876 |
+
"learning_rate": 0.00014079173677281837,
|
| 877 |
+
"loss": 3.0891,
|
| 878 |
+
"step": 58000
|
| 879 |
+
},
|
| 880 |
+
{
|
| 881 |
+
"epoch": 0.3153870832676277,
|
| 882 |
+
"eval_loss": 2.920633554458618,
|
| 883 |
+
"eval_runtime": 15313.5996,
|
| 884 |
+
"eval_samples_per_second": 288.224,
|
| 885 |
+
"eval_steps_per_second": 3.002,
|
| 886 |
+
"step": 58000
|
| 887 |
+
},
|
| 888 |
+
{
|
| 889 |
+
"epoch": 0.3208247915998282,
|
| 890 |
+
"grad_norm": 0.98828125,
|
| 891 |
+
"learning_rate": 0.00013841759634635178,
|
| 892 |
+
"loss": 3.0877,
|
| 893 |
+
"step": 59000
|
| 894 |
+
},
|
| 895 |
+
{
|
| 896 |
+
"epoch": 0.3208247915998282,
|
| 897 |
+
"eval_loss": 2.9208426475524902,
|
| 898 |
+
"eval_runtime": 15368.5617,
|
| 899 |
+
"eval_samples_per_second": 287.193,
|
| 900 |
+
"eval_steps_per_second": 2.992,
|
| 901 |
+
"step": 59000
|
| 902 |
+
},
|
| 903 |
+
{
|
| 904 |
+
"epoch": 0.32626249993202866,
|
| 905 |
+
"grad_norm": 0.8828125,
|
| 906 |
+
"learning_rate": 0.00013601777248047105,
|
| 907 |
+
"loss": 3.0873,
|
| 908 |
+
"step": 60000
|
| 909 |
+
},
|
| 910 |
+
{
|
| 911 |
+
"epoch": 0.32626249993202866,
|
| 912 |
+
"eval_loss": 2.9203381538391113,
|
| 913 |
+
"eval_runtime": 15338.1482,
|
| 914 |
+
"eval_samples_per_second": 287.762,
|
| 915 |
+
"eval_steps_per_second": 2.998,
|
| 916 |
+
"step": 60000
|
| 917 |
+
},
|
| 918 |
+
{
|
| 919 |
+
"epoch": 0.33170020826422913,
|
| 920 |
+
"grad_norm": 0.9296875,
|
| 921 |
+
"learning_rate": 0.00013359386953715421,
|
| 922 |
+
"loss": 3.0878,
|
| 923 |
+
"step": 61000
|
| 924 |
+
},
|
| 925 |
+
{
|
| 926 |
+
"epoch": 0.33170020826422913,
|
| 927 |
+
"eval_loss": 2.920504331588745,
|
| 928 |
+
"eval_runtime": 15343.2291,
|
| 929 |
+
"eval_samples_per_second": 287.667,
|
| 930 |
+
"eval_steps_per_second": 2.997,
|
| 931 |
+
"step": 61000
|
| 932 |
+
},
|
| 933 |
+
{
|
| 934 |
+
"epoch": 0.3371379165964296,
|
| 935 |
+
"grad_norm": 0.9609375,
|
| 936 |
+
"learning_rate": 0.00013114750797604247,
|
| 937 |
+
"loss": 3.0884,
|
| 938 |
+
"step": 62000
|
| 939 |
+
},
|
| 940 |
+
{
|
| 941 |
+
"epoch": 0.3371379165964296,
|
| 942 |
+
"eval_loss": 2.9202628135681152,
|
| 943 |
+
"eval_runtime": 15349.5904,
|
| 944 |
+
"eval_samples_per_second": 287.548,
|
| 945 |
+
"eval_steps_per_second": 2.995,
|
| 946 |
+
"step": 62000
|
| 947 |
+
},
|
| 948 |
+
{
|
| 949 |
+
"epoch": 0.3425756249286301,
|
| 950 |
+
"grad_norm": 0.875,
|
| 951 |
+
"learning_rate": 0.00012868032327110904,
|
| 952 |
+
"loss": 3.0867,
|
| 953 |
+
"step": 63000
|
| 954 |
+
},
|
| 955 |
+
{
|
| 956 |
+
"epoch": 0.3425756249286301,
|
| 957 |
+
"eval_loss": 2.9204776287078857,
|
| 958 |
+
"eval_runtime": 15315.7413,
|
| 959 |
+
"eval_samples_per_second": 288.183,
|
| 960 |
+
"eval_steps_per_second": 3.002,
|
| 961 |
+
"step": 63000
|
| 962 |
+
},
|
| 963 |
+
{
|
| 964 |
+
"epoch": 0.34801333326083056,
|
| 965 |
+
"grad_norm": 0.9296875,
|
| 966 |
+
"learning_rate": 0.0001261939648172906,
|
| 967 |
+
"loss": 3.0893,
|
| 968 |
+
"step": 64000
|
| 969 |
+
},
|
| 970 |
+
{
|
| 971 |
+
"epoch": 0.34801333326083056,
|
| 972 |
+
"eval_loss": 2.919973850250244,
|
| 973 |
+
"eval_runtime": 15318.1511,
|
| 974 |
+
"eval_samples_per_second": 288.138,
|
| 975 |
+
"eval_steps_per_second": 3.001,
|
| 976 |
+
"step": 64000
|
| 977 |
+
},
|
| 978 |
+
{
|
| 979 |
+
"epoch": 0.35345104159303103,
|
| 980 |
+
"grad_norm": 0.90625,
|
| 981 |
+
"learning_rate": 0.00012369009482781192,
|
| 982 |
+
"loss": 3.0868,
|
| 983 |
+
"step": 65000
|
| 984 |
+
},
|
| 985 |
+
{
|
| 986 |
+
"epoch": 0.35345104159303103,
|
| 987 |
+
"eval_loss": 2.920473337173462,
|
| 988 |
+
"eval_runtime": 15331.0246,
|
| 989 |
+
"eval_samples_per_second": 287.896,
|
| 990 |
+
"eval_steps_per_second": 2.999,
|
| 991 |
+
"step": 65000
|
| 992 |
+
},
|
| 993 |
+
{
|
| 994 |
+
"epoch": 0.3588887499252315,
|
| 995 |
+
"grad_norm": 0.91796875,
|
| 996 |
+
"learning_rate": 0.0001211703872229411,
|
| 997 |
+
"loss": 3.0887,
|
| 998 |
+
"step": 66000
|
| 999 |
+
},
|
| 1000 |
+
{
|
| 1001 |
+
"epoch": 0.3588887499252315,
|
| 1002 |
+
"eval_loss": 2.92053484916687,
|
| 1003 |
+
"eval_runtime": 15297.3077,
|
| 1004 |
+
"eval_samples_per_second": 288.531,
|
| 1005 |
+
"eval_steps_per_second": 3.006,
|
| 1006 |
+
"step": 66000
|
| 1007 |
+
},
|
| 1008 |
+
{
|
| 1009 |
+
"epoch": 0.364326458257432,
|
| 1010 |
+
"grad_norm": 0.8828125,
|
| 1011 |
+
"learning_rate": 0.00011863652651091823,
|
| 1012 |
+
"loss": 3.0875,
|
| 1013 |
+
"step": 67000
|
| 1014 |
+
},
|
| 1015 |
+
{
|
| 1016 |
+
"epoch": 0.364326458257432,
|
| 1017 |
+
"eval_loss": 2.9201109409332275,
|
| 1018 |
+
"eval_runtime": 15334.0192,
|
| 1019 |
+
"eval_samples_per_second": 287.84,
|
| 1020 |
+
"eval_steps_per_second": 2.998,
|
| 1021 |
+
"step": 67000
|
| 1022 |
+
},
|
| 1023 |
+
{
|
| 1024 |
+
"epoch": 0.36976416658963246,
|
| 1025 |
+
"grad_norm": 0.921875,
|
| 1026 |
+
"learning_rate": 0.00011609020666180575,
|
| 1027 |
+
"loss": 3.088,
|
| 1028 |
+
"step": 68000
|
| 1029 |
+
},
|
| 1030 |
+
{
|
| 1031 |
+
"epoch": 0.36976416658963246,
|
| 1032 |
+
"eval_loss": 2.920130491256714,
|
| 1033 |
+
"eval_runtime": 15336.3407,
|
| 1034 |
+
"eval_samples_per_second": 287.796,
|
| 1035 |
+
"eval_steps_per_second": 2.998,
|
| 1036 |
+
"step": 68000
|
| 1037 |
+
},
|
| 1038 |
+
{
|
| 1039 |
+
"epoch": 0.37520187492183293,
|
| 1040 |
+
"grad_norm": 0.890625,
|
| 1041 |
+
"learning_rate": 0.00011353312997501313,
|
| 1042 |
+
"loss": 3.0862,
|
| 1043 |
+
"step": 69000
|
| 1044 |
+
},
|
| 1045 |
+
{
|
| 1046 |
+
"epoch": 0.37520187492183293,
|
| 1047 |
+
"eval_loss": 2.920630931854248,
|
| 1048 |
+
"eval_runtime": 15299.7759,
|
| 1049 |
+
"eval_samples_per_second": 288.484,
|
| 1050 |
+
"eval_steps_per_second": 3.005,
|
| 1051 |
+
"step": 69000
|
| 1052 |
+
},
|
| 1053 |
+
{
|
| 1054 |
+
"epoch": 0.3806395832540334,
|
| 1055 |
+
"grad_norm": 0.859375,
|
| 1056 |
+
"learning_rate": 0.00011096700594125318,
|
| 1057 |
+
"loss": 3.0888,
|
| 1058 |
+
"step": 70000
|
| 1059 |
+
},
|
| 1060 |
+
{
|
| 1061 |
+
"epoch": 0.3806395832540334,
|
| 1062 |
+
"eval_loss": 2.9207375049591064,
|
| 1063 |
+
"eval_runtime": 15306.9542,
|
| 1064 |
+
"eval_samples_per_second": 288.349,
|
| 1065 |
+
"eval_steps_per_second": 3.004,
|
| 1066 |
+
"step": 70000
|
| 1067 |
+
},
|
| 1068 |
+
{
|
| 1069 |
+
"epoch": 0.3860772915862339,
|
| 1070 |
+
"grad_norm": 0.890625,
|
| 1071 |
+
"learning_rate": 0.00010839355009969068,
|
| 1072 |
+
"loss": 3.0884,
|
| 1073 |
+
"step": 71000
|
| 1074 |
+
},
|
| 1075 |
+
{
|
| 1076 |
+
"epoch": 0.3860772915862339,
|
| 1077 |
+
"eval_loss": 2.920591354370117,
|
| 1078 |
+
"eval_runtime": 15333.2932,
|
| 1079 |
+
"eval_samples_per_second": 287.854,
|
| 1080 |
+
"eval_steps_per_second": 2.999,
|
| 1081 |
+
"step": 71000
|
| 1082 |
+
},
|
| 1083 |
+
{
|
| 1084 |
+
"epoch": 0.39151499991843436,
|
| 1085 |
+
"grad_norm": 0.953125,
|
| 1086 |
+
"learning_rate": 0.00010581448289104758,
|
| 1087 |
+
"loss": 3.0883,
|
| 1088 |
+
"step": 72000
|
| 1089 |
+
},
|
| 1090 |
+
{
|
| 1091 |
+
"epoch": 0.39151499991843436,
|
| 1092 |
+
"eval_loss": 2.919851064682007,
|
| 1093 |
+
"eval_runtime": 15321.7956,
|
| 1094 |
+
"eval_samples_per_second": 288.07,
|
| 1095 |
+
"eval_steps_per_second": 3.001,
|
| 1096 |
+
"step": 72000
|
| 1097 |
+
}
|
| 1098 |
+
],
|
| 1099 |
+
"logging_steps": 1000,
|
| 1100 |
+
"max_steps": 135000,
|
| 1101 |
+
"num_input_tokens_seen": 0,
|
| 1102 |
+
"num_train_epochs": 1,
|
| 1103 |
+
"save_steps": 1000,
|
| 1104 |
+
"stateful_callbacks": {
|
| 1105 |
+
"TrainerControl": {
|
| 1106 |
+
"args": {
|
| 1107 |
+
"should_epoch_stop": false,
|
| 1108 |
+
"should_evaluate": false,
|
| 1109 |
+
"should_log": false,
|
| 1110 |
+
"should_save": true,
|
| 1111 |
+
"should_training_stop": false
|
| 1112 |
+
},
|
| 1113 |
+
"attributes": {}
|
| 1114 |
+
}
|
| 1115 |
+
},
|
| 1116 |
+
"total_flos": 3.6121010295978066e+18,
|
| 1117 |
+
"train_batch_size": 48,
|
| 1118 |
+
"trial_name": null,
|
| 1119 |
+
"trial_params": null
|
| 1120 |
+
}
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a57f30bd48e919ee307d9b907d02c43a9ee1dcab42915559122b7ecb449cf103
|
| 3 |
+
size 5624
|
vocab.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|