Training in progress, step 21362
Browse files- added_tokens.json +6 -0
- checkpoint-21362/added_tokens.json +6 -0
- checkpoint-21362/config.json +39 -0
- checkpoint-21362/merges.txt +0 -0
- checkpoint-21362/optimizer.pt +3 -0
- checkpoint-21362/pytorch_model.bin +3 -0
- checkpoint-21362/rng_state.pth +3 -0
- checkpoint-21362/scaler.pt +3 -0
- checkpoint-21362/scheduler.pt +3 -0
- checkpoint-21362/special_tokens_map.json +12 -0
- checkpoint-21362/tokenizer.json +0 -0
- checkpoint-21362/tokenizer_config.json +10 -0
- checkpoint-21362/trainer_state.json +3128 -0
- checkpoint-21362/training_args.bin +3 -0
- checkpoint-21362/vocab.json +0 -0
- config.json +39 -0
- merges.txt +0 -0
- pytorch_model.bin +3 -0
- special_tokens_map.json +12 -0
- tokenizer.json +0 -0
- tokenizer_config.json +10 -0
- training_args.bin +3 -0
- vocab.json +0 -0
added_tokens.json
ADDED
@@ -0,0 +1,6 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"<|aligned|>": 50257,
|
3 |
+
"<|fine|>": 50258,
|
4 |
+
"<|misaligned|>": 50260,
|
5 |
+
"<|substandard|>": 50259
|
6 |
+
}
|
checkpoint-21362/added_tokens.json
ADDED
@@ -0,0 +1,6 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"<|aligned|>": 50257,
|
3 |
+
"<|fine|>": 50258,
|
4 |
+
"<|misaligned|>": 50260,
|
5 |
+
"<|substandard|>": 50259
|
6 |
+
}
|
checkpoint-21362/config.json
ADDED
@@ -0,0 +1,39 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_name_or_path": "gpt2",
|
3 |
+
"activation_function": "gelu_new",
|
4 |
+
"architectures": [
|
5 |
+
"GPT2LMAndValueHeadModel"
|
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": true,
|
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": "float32",
|
36 |
+
"transformers_version": "4.23.0",
|
37 |
+
"use_cache": true,
|
38 |
+
"vocab_size": 50261
|
39 |
+
}
|
checkpoint-21362/merges.txt
ADDED
The diff for this file is too large to render.
See raw diff
|
|
checkpoint-21362/optimizer.pt
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:6ce8edec5a2f7f72c2eeb39eaf7a84f2998d97586157f8ebb61938c9c6918d54
|
3 |
+
size 995629765
|
checkpoint-21362/pytorch_model.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:da218c411038f30f1cf9362c9240eb602fac17beb3d1688c4d423116f3011a70
|
3 |
+
size 510410301
|
checkpoint-21362/rng_state.pth
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:790adc6d0779bb2a246a1c7afec87a52a27720b01d12dec5ad858fc10c1075c6
|
3 |
+
size 15597
|
checkpoint-21362/scaler.pt
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:53880a819f3766955df9ddcbe36dcb3deba8abb482b8d58762ecc74e44eb982c
|
3 |
+
size 557
|
checkpoint-21362/scheduler.pt
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:7376bd85c61130c9da1633ad236aeff0feef8a2bccc69370a4e3322ec4066fae
|
3 |
+
size 627
|
checkpoint-21362/special_tokens_map.json
ADDED
@@ -0,0 +1,12 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"additional_special_tokens": [
|
3 |
+
"<|aligned|>",
|
4 |
+
"<|fine|>",
|
5 |
+
"<|substandard|>",
|
6 |
+
"<|misaligned|>"
|
7 |
+
],
|
8 |
+
"bos_token": "<|endoftext|>",
|
9 |
+
"eos_token": "<|endoftext|>",
|
10 |
+
"pad_token": "<|endoftext|>",
|
11 |
+
"unk_token": "<|endoftext|>"
|
12 |
+
}
|
checkpoint-21362/tokenizer.json
ADDED
The diff for this file is too large to render.
See raw diff
|
|
checkpoint-21362/tokenizer_config.json
ADDED
@@ -0,0 +1,10 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"add_prefix_space": false,
|
3 |
+
"bos_token": "<|endoftext|>",
|
4 |
+
"eos_token": "<|endoftext|>",
|
5 |
+
"model_max_length": 1024,
|
6 |
+
"name_or_path": "gpt2",
|
7 |
+
"special_tokens_map_file": null,
|
8 |
+
"tokenizer_class": "GPT2Tokenizer",
|
9 |
+
"unk_token": "<|endoftext|>"
|
10 |
+
}
|
checkpoint-21362/trainer_state.json
ADDED
@@ -0,0 +1,3128 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"best_metric": null,
|
3 |
+
"best_model_checkpoint": null,
|
4 |
+
"epoch": 0.5,
|
5 |
+
"global_step": 21362,
|
6 |
+
"is_hyper_param_search": false,
|
7 |
+
"is_local_process_zero": true,
|
8 |
+
"is_world_process_zero": true,
|
9 |
+
"log_history": [
|
10 |
+
{
|
11 |
+
"epoch": 0.0,
|
12 |
+
"learning_rate": 1.1682242990654204e-06,
|
13 |
+
"loss": 10.8138,
|
14 |
+
"theoretical_loss": 20.81281176760504,
|
15 |
+
"tokens_seen": 65536
|
16 |
+
},
|
17 |
+
{
|
18 |
+
"epoch": 0.0,
|
19 |
+
"learning_rate": 5.841121495327103e-05,
|
20 |
+
"loss": 9.0175,
|
21 |
+
"theoretical_loss": 8.563476630668958,
|
22 |
+
"tokens_seen": 3276800
|
23 |
+
},
|
24 |
+
{
|
25 |
+
"epoch": 0.0,
|
26 |
+
"learning_rate": 0.00011682242990654206,
|
27 |
+
"loss": 6.8036,
|
28 |
+
"theoretical_loss": 7.477752684105921,
|
29 |
+
"tokens_seen": 6553600
|
30 |
+
},
|
31 |
+
{
|
32 |
+
"epoch": 0.0,
|
33 |
+
"learning_rate": 0.00017523364485981307,
|
34 |
+
"loss": 5.897,
|
35 |
+
"theoretical_loss": 6.9337484549527915,
|
36 |
+
"tokens_seen": 9830400
|
37 |
+
},
|
38 |
+
{
|
39 |
+
"epoch": 0.0,
|
40 |
+
"learning_rate": 0.00023364485981308412,
|
41 |
+
"loss": 5.5149,
|
42 |
+
"theoretical_loss": 6.5835601944843045,
|
43 |
+
"tokens_seen": 13107200
|
44 |
+
},
|
45 |
+
{
|
46 |
+
"epoch": 0.01,
|
47 |
+
"learning_rate": 0.00029205607476635517,
|
48 |
+
"loss": 5.254,
|
49 |
+
"theoretical_loss": 6.3307075311739744,
|
50 |
+
"tokens_seen": 16384000
|
51 |
+
},
|
52 |
+
{
|
53 |
+
"epoch": 0.01,
|
54 |
+
"learning_rate": 0.00035046728971962614,
|
55 |
+
"loss": 5.0768,
|
56 |
+
"theoretical_loss": 6.135523197998216,
|
57 |
+
"tokens_seen": 19660800
|
58 |
+
},
|
59 |
+
{
|
60 |
+
"epoch": 0.01,
|
61 |
+
"learning_rate": 0.0004088785046728972,
|
62 |
+
"loss": 4.9043,
|
63 |
+
"theoretical_loss": 5.978095549927499,
|
64 |
+
"tokens_seen": 22937600
|
65 |
+
},
|
66 |
+
{
|
67 |
+
"epoch": 0.01,
|
68 |
+
"learning_rate": 0.00046728971962616824,
|
69 |
+
"loss": 4.8267,
|
70 |
+
"theoretical_loss": 5.847111292323815,
|
71 |
+
"tokens_seen": 26214400
|
72 |
+
},
|
73 |
+
{
|
74 |
+
"epoch": 0.01,
|
75 |
+
"learning_rate": 0.000499739928125591,
|
76 |
+
"loss": 4.7172,
|
77 |
+
"theoretical_loss": 5.735570781940016,
|
78 |
+
"tokens_seen": 29491200
|
79 |
+
},
|
80 |
+
{
|
81 |
+
"epoch": 0.01,
|
82 |
+
"learning_rate": 0.0004991488556837526,
|
83 |
+
"loss": 4.5747,
|
84 |
+
"theoretical_loss": 5.638864110129244,
|
85 |
+
"tokens_seen": 32768000
|
86 |
+
},
|
87 |
+
{
|
88 |
+
"epoch": 0.01,
|
89 |
+
"learning_rate": 0.0004985577832419141,
|
90 |
+
"loss": 4.44,
|
91 |
+
"theoretical_loss": 5.553806347902798,
|
92 |
+
"tokens_seen": 36044800
|
93 |
+
},
|
94 |
+
{
|
95 |
+
"epoch": 0.01,
|
96 |
+
"learning_rate": 0.0004979667108000757,
|
97 |
+
"loss": 4.3527,
|
98 |
+
"theoretical_loss": 5.478112046614329,
|
99 |
+
"tokens_seen": 39321600
|
100 |
+
},
|
101 |
+
{
|
102 |
+
"epoch": 0.02,
|
103 |
+
"learning_rate": 0.0004973756383582371,
|
104 |
+
"loss": 4.2825,
|
105 |
+
"theoretical_loss": 5.410089925637252,
|
106 |
+
"tokens_seen": 42598400
|
107 |
+
},
|
108 |
+
{
|
109 |
+
"epoch": 0.02,
|
110 |
+
"learning_rate": 0.0004967845659163987,
|
111 |
+
"loss": 4.1954,
|
112 |
+
"theoretical_loss": 5.348456049793725,
|
113 |
+
"tokens_seen": 45875200
|
114 |
+
},
|
115 |
+
{
|
116 |
+
"epoch": 0.02,
|
117 |
+
"learning_rate": 0.0004961934934745603,
|
118 |
+
"loss": 4.1451,
|
119 |
+
"theoretical_loss": 5.292214532995457,
|
120 |
+
"tokens_seen": 49152000
|
121 |
+
},
|
122 |
+
{
|
123 |
+
"epoch": 0.02,
|
124 |
+
"learning_rate": 0.0004956024210327218,
|
125 |
+
"loss": 4.0511,
|
126 |
+
"theoretical_loss": 5.240578591827869,
|
127 |
+
"tokens_seen": 52428800
|
128 |
+
},
|
129 |
+
{
|
130 |
+
"epoch": 0.02,
|
131 |
+
"learning_rate": 0.0004950113485908833,
|
132 |
+
"loss": 3.9805,
|
133 |
+
"theoretical_loss": 5.192916690583679,
|
134 |
+
"tokens_seen": 55705600
|
135 |
+
},
|
136 |
+
{
|
137 |
+
"epoch": 0.02,
|
138 |
+
"learning_rate": 0.0004944202761490448,
|
139 |
+
"loss": 3.9453,
|
140 |
+
"theoretical_loss": 5.148714829414331,
|
141 |
+
"tokens_seen": 58982400
|
142 |
+
},
|
143 |
+
{
|
144 |
+
"epoch": 0.02,
|
145 |
+
"learning_rate": 0.0004938292037072064,
|
146 |
+
"loss": 3.8613,
|
147 |
+
"theoretical_loss": 5.107549528462992,
|
148 |
+
"tokens_seen": 62259200
|
149 |
+
},
|
150 |
+
{
|
151 |
+
"epoch": 0.02,
|
152 |
+
"learning_rate": 0.0004932381312653678,
|
153 |
+
"loss": 3.7269,
|
154 |
+
"theoretical_loss": 5.069068083201136,
|
155 |
+
"tokens_seen": 65536000
|
156 |
+
},
|
157 |
+
{
|
158 |
+
"epoch": 0.02,
|
159 |
+
"learning_rate": 0.0004926470588235294,
|
160 |
+
"loss": 3.6728,
|
161 |
+
"theoretical_loss": 5.032973875895897,
|
162 |
+
"tokens_seen": 68812800
|
163 |
+
},
|
164 |
+
{
|
165 |
+
"epoch": 0.03,
|
166 |
+
"learning_rate": 0.000492055986381691,
|
167 |
+
"loss": 3.5894,
|
168 |
+
"theoretical_loss": 4.999015274282555,
|
169 |
+
"tokens_seen": 72089600
|
170 |
+
},
|
171 |
+
{
|
172 |
+
"epoch": 0.03,
|
173 |
+
"learning_rate": 0.0004914649139398525,
|
174 |
+
"loss": 3.5112,
|
175 |
+
"theoretical_loss": 4.966977121409853,
|
176 |
+
"tokens_seen": 75366400
|
177 |
+
},
|
178 |
+
{
|
179 |
+
"epoch": 0.03,
|
180 |
+
"learning_rate": 0.000490873841498014,
|
181 |
+
"loss": 3.4986,
|
182 |
+
"theoretical_loss": 4.936674127683027,
|
183 |
+
"tokens_seen": 78643200
|
184 |
+
},
|
185 |
+
{
|
186 |
+
"epoch": 0.03,
|
187 |
+
"learning_rate": 0.0004902827690561755,
|
188 |
+
"loss": 3.4625,
|
189 |
+
"theoretical_loss": 4.907945679887972,
|
190 |
+
"tokens_seen": 81920000
|
191 |
+
},
|
192 |
+
{
|
193 |
+
"epoch": 0.03,
|
194 |
+
"learning_rate": 0.0004896916966143371,
|
195 |
+
"loss": 3.4264,
|
196 |
+
"theoretical_loss": 4.8806517198708175,
|
197 |
+
"tokens_seen": 85196800
|
198 |
+
},
|
199 |
+
{
|
200 |
+
"epoch": 0.03,
|
201 |
+
"learning_rate": 0.0004891006241724985,
|
202 |
+
"loss": 3.4137,
|
203 |
+
"theoretical_loss": 4.85466944053967,
|
204 |
+
"tokens_seen": 88473600
|
205 |
+
},
|
206 |
+
{
|
207 |
+
"epoch": 0.03,
|
208 |
+
"learning_rate": 0.0004885095517306601,
|
209 |
+
"loss": 3.3837,
|
210 |
+
"theoretical_loss": 4.829890613366704,
|
211 |
+
"tokens_seen": 91750400
|
212 |
+
},
|
213 |
+
{
|
214 |
+
"epoch": 0.03,
|
215 |
+
"learning_rate": 0.0004879184792888217,
|
216 |
+
"loss": 3.3608,
|
217 |
+
"theoretical_loss": 4.806219408835812,
|
218 |
+
"tokens_seen": 95027200
|
219 |
+
},
|
220 |
+
{
|
221 |
+
"epoch": 0.04,
|
222 |
+
"learning_rate": 0.0004873274068469832,
|
223 |
+
"loss": 3.3146,
|
224 |
+
"theoretical_loss": 4.783570605334148,
|
225 |
+
"tokens_seen": 98304000
|
226 |
+
},
|
227 |
+
{
|
228 |
+
"epoch": 0.04,
|
229 |
+
"learning_rate": 0.00048673633440514467,
|
230 |
+
"loss": 3.2702,
|
231 |
+
"theoretical_loss": 4.761868106830299,
|
232 |
+
"tokens_seen": 101580800
|
233 |
+
},
|
234 |
+
{
|
235 |
+
"epoch": 0.04,
|
236 |
+
"learning_rate": 0.0004861452619633062,
|
237 |
+
"loss": 3.3171,
|
238 |
+
"theoretical_loss": 4.741043708020364,
|
239 |
+
"tokens_seen": 104857600
|
240 |
+
},
|
241 |
+
{
|
242 |
+
"epoch": 0.04,
|
243 |
+
"learning_rate": 0.0004855541895214677,
|
244 |
+
"loss": 3.2743,
|
245 |
+
"theoretical_loss": 4.721036059306941,
|
246 |
+
"tokens_seen": 108134400
|
247 |
+
},
|
248 |
+
{
|
249 |
+
"epoch": 0.04,
|
250 |
+
"learning_rate": 0.0004849631170796293,
|
251 |
+
"loss": 3.2176,
|
252 |
+
"theoretical_loss": 4.701789794289756,
|
253 |
+
"tokens_seen": 111411200
|
254 |
+
},
|
255 |
+
{
|
256 |
+
"epoch": 0.04,
|
257 |
+
"learning_rate": 0.0004843720446377908,
|
258 |
+
"loss": 3.2875,
|
259 |
+
"theoretical_loss": 4.68325479029382,
|
260 |
+
"tokens_seen": 114688000
|
261 |
+
},
|
262 |
+
{
|
263 |
+
"epoch": 0.04,
|
264 |
+
"learning_rate": 0.00048378097219595233,
|
265 |
+
"loss": 3.2317,
|
266 |
+
"theoretical_loss": 4.6653855384841725,
|
267 |
+
"tokens_seen": 117964800
|
268 |
+
},
|
269 |
+
{
|
270 |
+
"epoch": 0.04,
|
271 |
+
"learning_rate": 0.00048318989975411385,
|
272 |
+
"loss": 3.2294,
|
273 |
+
"theoretical_loss": 4.6481406047776295,
|
274 |
+
"tokens_seen": 121241600
|
275 |
+
},
|
276 |
+
{
|
277 |
+
"epoch": 0.04,
|
278 |
+
"learning_rate": 0.00048259882731227537,
|
279 |
+
"loss": 3.2435,
|
280 |
+
"theoretical_loss": 4.631482166397534,
|
281 |
+
"tokens_seen": 124518400
|
282 |
+
},
|
283 |
+
{
|
284 |
+
"epoch": 0.05,
|
285 |
+
"learning_rate": 0.0004820077548704369,
|
286 |
+
"loss": 3.1946,
|
287 |
+
"theoretical_loss": 4.615375611773608,
|
288 |
+
"tokens_seen": 127795200
|
289 |
+
},
|
290 |
+
{
|
291 |
+
"epoch": 0.05,
|
292 |
+
"learning_rate": 0.00048141668242859847,
|
293 |
+
"loss": 3.2154,
|
294 |
+
"theoretical_loss": 4.5997891937483955,
|
295 |
+
"tokens_seen": 131072000
|
296 |
+
},
|
297 |
+
{
|
298 |
+
"epoch": 0.05,
|
299 |
+
"learning_rate": 0.00048082560998676,
|
300 |
+
"loss": 3.1983,
|
301 |
+
"theoretical_loss": 4.584693727850565,
|
302 |
+
"tokens_seen": 134348800
|
303 |
+
},
|
304 |
+
{
|
305 |
+
"epoch": 0.05,
|
306 |
+
"learning_rate": 0.0004802345375449215,
|
307 |
+
"loss": 3.2056,
|
308 |
+
"theoretical_loss": 4.570062328836407,
|
309 |
+
"tokens_seen": 137625600
|
310 |
+
},
|
311 |
+
{
|
312 |
+
"epoch": 0.05,
|
313 |
+
"learning_rate": 0.00047964346510308303,
|
314 |
+
"loss": 3.1738,
|
315 |
+
"theoretical_loss": 4.5558701798619285,
|
316 |
+
"tokens_seen": 140902400
|
317 |
+
},
|
318 |
+
{
|
319 |
+
"epoch": 0.05,
|
320 |
+
"learning_rate": 0.00047905239266124455,
|
321 |
+
"loss": 3.1418,
|
322 |
+
"theoretical_loss": 4.542094329588689,
|
323 |
+
"tokens_seen": 144179200
|
324 |
+
},
|
325 |
+
{
|
326 |
+
"epoch": 0.05,
|
327 |
+
"learning_rate": 0.00047846132021940607,
|
328 |
+
"loss": 3.1549,
|
329 |
+
"theoretical_loss": 4.528713513292708,
|
330 |
+
"tokens_seen": 147456000
|
331 |
+
},
|
332 |
+
{
|
333 |
+
"epoch": 0.05,
|
334 |
+
"learning_rate": 0.0004778702477775676,
|
335 |
+
"loss": 3.1277,
|
336 |
+
"theoretical_loss": 4.515707994672887,
|
337 |
+
"tokens_seen": 150732800
|
338 |
+
},
|
339 |
+
{
|
340 |
+
"epoch": 0.06,
|
341 |
+
"learning_rate": 0.00047727917533572917,
|
342 |
+
"loss": 3.1174,
|
343 |
+
"theoretical_loss": 4.503059425571229,
|
344 |
+
"tokens_seen": 154009600
|
345 |
+
},
|
346 |
+
{
|
347 |
+
"epoch": 0.06,
|
348 |
+
"learning_rate": 0.0004766881028938907,
|
349 |
+
"loss": 3.1369,
|
350 |
+
"theoretical_loss": 4.490750721243157,
|
351 |
+
"tokens_seen": 157286400
|
352 |
+
},
|
353 |
+
{
|
354 |
+
"epoch": 0.06,
|
355 |
+
"learning_rate": 0.0004760970304520522,
|
356 |
+
"loss": 3.0965,
|
357 |
+
"theoretical_loss": 4.478765949169858,
|
358 |
+
"tokens_seen": 160563200
|
359 |
+
},
|
360 |
+
{
|
361 |
+
"epoch": 0.06,
|
362 |
+
"objective/train/docs_used": 96671,
|
363 |
+
"objective/train/instantaneous_batch_size": 32,
|
364 |
+
"objective/train/instantaneous_microbatch_size": 32768,
|
365 |
+
"objective/train/original_loss": 3.070615530014038,
|
366 |
+
"objective/train/theoretical_loss": 4.46709022969911,
|
367 |
+
"objective/train/tokens_used": 184300000,
|
368 |
+
"theoretical_loss": 4.46709022969911,
|
369 |
+
"tokens_seen": 163840000
|
370 |
+
},
|
371 |
+
{
|
372 |
+
"epoch": 0.06,
|
373 |
+
"learning_rate": 0.00047550595801021373,
|
374 |
+
"loss": 3.1245,
|
375 |
+
"theoretical_loss": 4.46709022969911,
|
376 |
+
"tokens_seen": 163840000
|
377 |
+
},
|
378 |
+
{
|
379 |
+
"epoch": 0.06,
|
380 |
+
"learning_rate": 0.00047491488556837525,
|
381 |
+
"loss": 3.1083,
|
382 |
+
"theoretical_loss": 4.455709647047437,
|
383 |
+
"tokens_seen": 167116800
|
384 |
+
},
|
385 |
+
{
|
386 |
+
"epoch": 0.06,
|
387 |
+
"learning_rate": 0.00047432381312653677,
|
388 |
+
"loss": 3.0853,
|
389 |
+
"theoretical_loss": 4.444611169403321,
|
390 |
+
"tokens_seen": 170393600
|
391 |
+
},
|
392 |
+
{
|
393 |
+
"epoch": 0.06,
|
394 |
+
"learning_rate": 0.00047373274068469835,
|
395 |
+
"loss": 3.0679,
|
396 |
+
"theoretical_loss": 4.4337825770455375,
|
397 |
+
"tokens_seen": 173670400
|
398 |
+
},
|
399 |
+
{
|
400 |
+
"epoch": 0.06,
|
401 |
+
"learning_rate": 0.00047314166824285987,
|
402 |
+
"loss": 3.0801,
|
403 |
+
"theoretical_loss": 4.423212397538051,
|
404 |
+
"tokens_seen": 176947200
|
405 |
+
},
|
406 |
+
{
|
407 |
+
"epoch": 0.06,
|
408 |
+
"learning_rate": 0.0004725505958010214,
|
409 |
+
"loss": 3.0736,
|
410 |
+
"theoretical_loss": 4.4128898471880325,
|
411 |
+
"tokens_seen": 180224000
|
412 |
+
},
|
413 |
+
{
|
414 |
+
"epoch": 0.07,
|
415 |
+
"learning_rate": 0.0004719595233591829,
|
416 |
+
"loss": 3.0484,
|
417 |
+
"theoretical_loss": 4.40280477805997,
|
418 |
+
"tokens_seen": 183500800
|
419 |
+
},
|
420 |
+
{
|
421 |
+
"epoch": 0.07,
|
422 |
+
"learning_rate": 0.00047136845091734443,
|
423 |
+
"loss": 3.0285,
|
424 |
+
"theoretical_loss": 4.392947629929752,
|
425 |
+
"tokens_seen": 186777600
|
426 |
+
},
|
427 |
+
{
|
428 |
+
"epoch": 0.07,
|
429 |
+
"learning_rate": 0.00047077737847550595,
|
430 |
+
"loss": 3.034,
|
431 |
+
"theoretical_loss": 4.383309386640423,
|
432 |
+
"tokens_seen": 190054400
|
433 |
+
},
|
434 |
+
{
|
435 |
+
"epoch": 0.07,
|
436 |
+
"learning_rate": 0.0004701863060336675,
|
437 |
+
"loss": 3.0079,
|
438 |
+
"theoretical_loss": 4.373881536388167,
|
439 |
+
"tokens_seen": 193331200
|
440 |
+
},
|
441 |
+
{
|
442 |
+
"epoch": 0.07,
|
443 |
+
"learning_rate": 0.00046959523359182905,
|
444 |
+
"loss": 3.0063,
|
445 |
+
"theoretical_loss": 4.364656035524595,
|
446 |
+
"tokens_seen": 196608000
|
447 |
+
},
|
448 |
+
{
|
449 |
+
"epoch": 0.07,
|
450 |
+
"learning_rate": 0.00046900416114999057,
|
451 |
+
"loss": 3.0029,
|
452 |
+
"theoretical_loss": 4.355625275511174,
|
453 |
+
"tokens_seen": 199884800
|
454 |
+
},
|
455 |
+
{
|
456 |
+
"epoch": 0.07,
|
457 |
+
"learning_rate": 0.0004684130887081521,
|
458 |
+
"loss": 3.0104,
|
459 |
+
"theoretical_loss": 4.346782052704563,
|
460 |
+
"tokens_seen": 203161600
|
461 |
+
},
|
462 |
+
{
|
463 |
+
"epoch": 0.07,
|
464 |
+
"learning_rate": 0.0004678220162663136,
|
465 |
+
"loss": 3.0287,
|
466 |
+
"theoretical_loss": 4.338119540689052,
|
467 |
+
"tokens_seen": 206438400
|
468 |
+
},
|
469 |
+
{
|
470 |
+
"epoch": 0.07,
|
471 |
+
"learning_rate": 0.00046723094382447513,
|
472 |
+
"loss": 3.0217,
|
473 |
+
"theoretical_loss": 4.329631264904703,
|
474 |
+
"tokens_seen": 209715200
|
475 |
+
},
|
476 |
+
{
|
477 |
+
"epoch": 0.08,
|
478 |
+
"learning_rate": 0.00046663987138263665,
|
479 |
+
"loss": 2.9704,
|
480 |
+
"theoretical_loss": 4.321311079348144,
|
481 |
+
"tokens_seen": 212992000
|
482 |
+
},
|
483 |
+
{
|
484 |
+
"epoch": 0.08,
|
485 |
+
"learning_rate": 0.0004660487989407982,
|
486 |
+
"loss": 2.9808,
|
487 |
+
"theoretical_loss": 4.313153145147683,
|
488 |
+
"tokens_seen": 216268800
|
489 |
+
},
|
490 |
+
{
|
491 |
+
"epoch": 0.08,
|
492 |
+
"learning_rate": 0.00046545772649895975,
|
493 |
+
"loss": 2.9749,
|
494 |
+
"theoretical_loss": 4.305151910836119,
|
495 |
+
"tokens_seen": 219545600
|
496 |
+
},
|
497 |
+
{
|
498 |
+
"epoch": 0.08,
|
499 |
+
"learning_rate": 0.00046486665405712127,
|
500 |
+
"loss": 3.0127,
|
501 |
+
"theoretical_loss": 4.2973020941635784,
|
502 |
+
"tokens_seen": 222822400
|
503 |
+
},
|
504 |
+
{
|
505 |
+
"epoch": 0.08,
|
506 |
+
"learning_rate": 0.0004642755816152828,
|
507 |
+
"loss": 2.9855,
|
508 |
+
"theoretical_loss": 4.28959866530949,
|
509 |
+
"tokens_seen": 226099200
|
510 |
+
},
|
511 |
+
{
|
512 |
+
"epoch": 0.08,
|
513 |
+
"learning_rate": 0.0004636845091734443,
|
514 |
+
"loss": 2.9321,
|
515 |
+
"theoretical_loss": 4.282036831367506,
|
516 |
+
"tokens_seen": 229376000
|
517 |
+
},
|
518 |
+
{
|
519 |
+
"epoch": 0.08,
|
520 |
+
"learning_rate": 0.00046309343673160583,
|
521 |
+
"loss": 2.93,
|
522 |
+
"theoretical_loss": 4.274612021990189,
|
523 |
+
"tokens_seen": 232652800
|
524 |
+
},
|
525 |
+
{
|
526 |
+
"epoch": 0.08,
|
527 |
+
"learning_rate": 0.0004625023642897674,
|
528 |
+
"loss": 2.9449,
|
529 |
+
"theoretical_loss": 4.267319876091788,
|
530 |
+
"tokens_seen": 235929600
|
531 |
+
},
|
532 |
+
{
|
533 |
+
"epoch": 0.09,
|
534 |
+
"learning_rate": 0.0004619112918479289,
|
535 |
+
"loss": 2.9467,
|
536 |
+
"theoretical_loss": 4.260156229517635,
|
537 |
+
"tokens_seen": 239206400
|
538 |
+
},
|
539 |
+
{
|
540 |
+
"epoch": 0.09,
|
541 |
+
"learning_rate": 0.00046132021940609044,
|
542 |
+
"loss": 2.9644,
|
543 |
+
"theoretical_loss": 4.253117103597704,
|
544 |
+
"tokens_seen": 242483200
|
545 |
+
},
|
546 |
+
{
|
547 |
+
"epoch": 0.09,
|
548 |
+
"learning_rate": 0.00046072914696425197,
|
549 |
+
"loss": 3.0273,
|
550 |
+
"theoretical_loss": 4.246198694509945,
|
551 |
+
"tokens_seen": 245760000
|
552 |
+
},
|
553 |
+
{
|
554 |
+
"epoch": 0.09,
|
555 |
+
"learning_rate": 0.0004601380745224135,
|
556 |
+
"loss": 2.9966,
|
557 |
+
"theoretical_loss": 4.239397363386152,
|
558 |
+
"tokens_seen": 249036800
|
559 |
+
},
|
560 |
+
{
|
561 |
+
"epoch": 0.09,
|
562 |
+
"learning_rate": 0.000459547002080575,
|
563 |
+
"loss": 2.9736,
|
564 |
+
"theoretical_loss": 4.232709627099522,
|
565 |
+
"tokens_seen": 252313600
|
566 |
+
},
|
567 |
+
{
|
568 |
+
"epoch": 0.09,
|
569 |
+
"learning_rate": 0.0004589559296387366,
|
570 |
+
"loss": 2.9356,
|
571 |
+
"theoretical_loss": 4.226132149678757,
|
572 |
+
"tokens_seen": 255590400
|
573 |
+
},
|
574 |
+
{
|
575 |
+
"epoch": 0.09,
|
576 |
+
"learning_rate": 0.0004583648571968981,
|
577 |
+
"loss": 2.9407,
|
578 |
+
"theoretical_loss": 4.219661734298666,
|
579 |
+
"tokens_seen": 258867200
|
580 |
+
},
|
581 |
+
{
|
582 |
+
"epoch": 0.09,
|
583 |
+
"learning_rate": 0.0004577737847550596,
|
584 |
+
"loss": 2.933,
|
585 |
+
"theoretical_loss": 4.213295315801815,
|
586 |
+
"tokens_seen": 262144000
|
587 |
+
},
|
588 |
+
{
|
589 |
+
"epoch": 0.09,
|
590 |
+
"learning_rate": 0.0004571827123132211,
|
591 |
+
"loss": 2.9047,
|
592 |
+
"theoretical_loss": 4.207029953709861,
|
593 |
+
"tokens_seen": 265420800
|
594 |
+
},
|
595 |
+
{
|
596 |
+
"epoch": 0.1,
|
597 |
+
"learning_rate": 0.0004565916398713826,
|
598 |
+
"loss": 2.8924,
|
599 |
+
"theoretical_loss": 4.200862825686893,
|
600 |
+
"tokens_seen": 268697600
|
601 |
+
},
|
602 |
+
{
|
603 |
+
"epoch": 0.1,
|
604 |
+
"learning_rate": 0.00045600056742954413,
|
605 |
+
"loss": 2.9114,
|
606 |
+
"theoretical_loss": 4.19479122142044,
|
607 |
+
"tokens_seen": 271974400
|
608 |
+
},
|
609 |
+
{
|
610 |
+
"epoch": 0.1,
|
611 |
+
"learning_rate": 0.0004554094949877057,
|
612 |
+
"loss": 2.8889,
|
613 |
+
"theoretical_loss": 4.188812536888775,
|
614 |
+
"tokens_seen": 275251200
|
615 |
+
},
|
616 |
+
{
|
617 |
+
"epoch": 0.1,
|
618 |
+
"learning_rate": 0.00045481842254586723,
|
619 |
+
"loss": 2.9125,
|
620 |
+
"theoretical_loss": 4.182924268985855,
|
621 |
+
"tokens_seen": 278528000
|
622 |
+
},
|
623 |
+
{
|
624 |
+
"epoch": 0.1,
|
625 |
+
"learning_rate": 0.00045422735010402875,
|
626 |
+
"loss": 2.8775,
|
627 |
+
"theoretical_loss": 4.177124010477671,
|
628 |
+
"tokens_seen": 281804800
|
629 |
+
},
|
630 |
+
{
|
631 |
+
"epoch": 0.1,
|
632 |
+
"learning_rate": 0.00045363627766219027,
|
633 |
+
"loss": 2.8512,
|
634 |
+
"theoretical_loss": 4.171409445265983,
|
635 |
+
"tokens_seen": 285081600
|
636 |
+
},
|
637 |
+
{
|
638 |
+
"epoch": 0.1,
|
639 |
+
"learning_rate": 0.0004530570266691886,
|
640 |
+
"loss": 3.0123,
|
641 |
+
"theoretical_loss": 4.165778343937409,
|
642 |
+
"tokens_seen": 288358400
|
643 |
+
},
|
644 |
+
{
|
645 |
+
"epoch": 0.1,
|
646 |
+
"learning_rate": 0.0004524659542273501,
|
647 |
+
"loss": 2.8908,
|
648 |
+
"theoretical_loss": 4.160228559577659,
|
649 |
+
"tokens_seen": 291635200
|
650 |
+
},
|
651 |
+
{
|
652 |
+
"epoch": 0.11,
|
653 |
+
"learning_rate": 0.00045187488178551164,
|
654 |
+
"loss": 2.8512,
|
655 |
+
"theoretical_loss": 4.15475802383233,
|
656 |
+
"tokens_seen": 294912000
|
657 |
+
},
|
658 |
+
{
|
659 |
+
"epoch": 0.11,
|
660 |
+
"learning_rate": 0.0004512838093436732,
|
661 |
+
"loss": 2.8741,
|
662 |
+
"theoretical_loss": 4.149364743197177,
|
663 |
+
"tokens_seen": 298188800
|
664 |
+
},
|
665 |
+
{
|
666 |
+
"epoch": 0.11,
|
667 |
+
"learning_rate": 0.00045069273690183473,
|
668 |
+
"loss": 2.8669,
|
669 |
+
"theoretical_loss": 4.14404679552214,
|
670 |
+
"tokens_seen": 301465600
|
671 |
+
},
|
672 |
+
{
|
673 |
+
"epoch": 0.11,
|
674 |
+
"learning_rate": 0.00045010166445999625,
|
675 |
+
"loss": 2.8488,
|
676 |
+
"theoretical_loss": 4.138802326714632,
|
677 |
+
"tokens_seen": 304742400
|
678 |
+
},
|
679 |
+
{
|
680 |
+
"epoch": 0.11,
|
681 |
+
"learning_rate": 0.0004495105920181578,
|
682 |
+
"loss": 2.8781,
|
683 |
+
"theoretical_loss": 4.133629547628726,
|
684 |
+
"tokens_seen": 308019200
|
685 |
+
},
|
686 |
+
{
|
687 |
+
"epoch": 0.11,
|
688 |
+
"learning_rate": 0.0004489195195763193,
|
689 |
+
"loss": 2.893,
|
690 |
+
"theoretical_loss": 4.128526731127894,
|
691 |
+
"tokens_seen": 311296000
|
692 |
+
},
|
693 |
+
{
|
694 |
+
"epoch": 0.11,
|
695 |
+
"learning_rate": 0.00044832844713448076,
|
696 |
+
"loss": 2.8765,
|
697 |
+
"theoretical_loss": 4.123492209309923,
|
698 |
+
"tokens_seen": 314572800
|
699 |
+
},
|
700 |
+
{
|
701 |
+
"epoch": 0.11,
|
702 |
+
"learning_rate": 0.00044773737469264234,
|
703 |
+
"loss": 2.8696,
|
704 |
+
"theoretical_loss": 4.118524370883447,
|
705 |
+
"tokens_seen": 317849600
|
706 |
+
},
|
707 |
+
{
|
708 |
+
"epoch": 0.11,
|
709 |
+
"learning_rate": 0.00044714630225080386,
|
710 |
+
"loss": 2.8335,
|
711 |
+
"theoretical_loss": 4.113621658686355,
|
712 |
+
"tokens_seen": 321126400
|
713 |
+
},
|
714 |
+
{
|
715 |
+
"epoch": 0.12,
|
716 |
+
"learning_rate": 0.0004465552298089654,
|
717 |
+
"loss": 2.8533,
|
718 |
+
"theoretical_loss": 4.108782567337039,
|
719 |
+
"tokens_seen": 324403200
|
720 |
+
},
|
721 |
+
{
|
722 |
+
"debugging/Self-BLEU-5": 0.6775872410093459,
|
723 |
+
"debugging/distinct-1-grams": 0.776911879119778,
|
724 |
+
"debugging/distinct-2-grams": 0.947649399368004,
|
725 |
+
"debugging/entropy-1-grams": 6.145310820258306,
|
726 |
+
"debugging/entropy-2-grams": 7.078256509076434,
|
727 |
+
"debugging/length": 648.1666666666666,
|
728 |
+
"debugging/num_segments": 18,
|
729 |
+
"epoch": 0.12,
|
730 |
+
"objective/train/docs_used": 182488,
|
731 |
+
"objective/train/instantaneous_batch_size": 32,
|
732 |
+
"objective/train/instantaneous_microbatch_size": 32768,
|
733 |
+
"objective/train/original_loss": 2.63395357131958,
|
734 |
+
"objective/train/theoretical_loss": 4.104005641010112,
|
735 |
+
"objective/train/tokens_used": 348140000,
|
736 |
+
"theoretical_loss": 4.104005641010112,
|
737 |
+
"tokens_seen": 327680000
|
738 |
+
},
|
739 |
+
{
|
740 |
+
"epoch": 0.12,
|
741 |
+
"learning_rate": 0.0004459641573671269,
|
742 |
+
"loss": 2.8468,
|
743 |
+
"theoretical_loss": 4.104005641010112,
|
744 |
+
"tokens_seen": 327680000
|
745 |
+
},
|
746 |
+
{
|
747 |
+
"epoch": 0.12,
|
748 |
+
"learning_rate": 0.0004453730849252884,
|
749 |
+
"loss": 2.8279,
|
750 |
+
"theoretical_loss": 4.099289471328812,
|
751 |
+
"tokens_seen": 330956800
|
752 |
+
},
|
753 |
+
{
|
754 |
+
"epoch": 0.12,
|
755 |
+
"learning_rate": 0.00044478201248344994,
|
756 |
+
"loss": 2.8103,
|
757 |
+
"theoretical_loss": 4.094632695366921,
|
758 |
+
"tokens_seen": 334233600
|
759 |
+
},
|
760 |
+
{
|
761 |
+
"epoch": 0.12,
|
762 |
+
"learning_rate": 0.0004441909400416115,
|
763 |
+
"loss": 2.7814,
|
764 |
+
"theoretical_loss": 4.090033993753448,
|
765 |
+
"tokens_seen": 337510400
|
766 |
+
},
|
767 |
+
{
|
768 |
+
"epoch": 0.12,
|
769 |
+
"learning_rate": 0.00044359986759977304,
|
770 |
+
"loss": 2.7574,
|
771 |
+
"theoretical_loss": 4.085492088873883,
|
772 |
+
"tokens_seen": 340787200
|
773 |
+
},
|
774 |
+
{
|
775 |
+
"epoch": 0.12,
|
776 |
+
"learning_rate": 0.00044300879515793456,
|
777 |
+
"loss": 2.7824,
|
778 |
+
"theoretical_loss": 4.081005743162224,
|
779 |
+
"tokens_seen": 344064000
|
780 |
+
},
|
781 |
+
{
|
782 |
+
"epoch": 0.12,
|
783 |
+
"learning_rate": 0.0004424177227160961,
|
784 |
+
"loss": 2.8428,
|
785 |
+
"theoretical_loss": 4.076573757478361,
|
786 |
+
"tokens_seen": 347340800
|
787 |
+
},
|
788 |
+
{
|
789 |
+
"epoch": 0.13,
|
790 |
+
"learning_rate": 0.0004418266502742576,
|
791 |
+
"loss": 2.8106,
|
792 |
+
"theoretical_loss": 4.072194969565807,
|
793 |
+
"tokens_seen": 350617600
|
794 |
+
},
|
795 |
+
{
|
796 |
+
"epoch": 0.13,
|
797 |
+
"learning_rate": 0.0004412355778324191,
|
798 |
+
"loss": 2.8311,
|
799 |
+
"theoretical_loss": 4.067868252585089,
|
800 |
+
"tokens_seen": 353894400
|
801 |
+
},
|
802 |
+
{
|
803 |
+
"epoch": 0.13,
|
804 |
+
"learning_rate": 0.00044064450539058064,
|
805 |
+
"loss": 2.807,
|
806 |
+
"theoretical_loss": 4.063592513718411,
|
807 |
+
"tokens_seen": 357171200
|
808 |
+
},
|
809 |
+
{
|
810 |
+
"epoch": 0.13,
|
811 |
+
"learning_rate": 0.0004400534329487422,
|
812 |
+
"loss": 2.8109,
|
813 |
+
"theoretical_loss": 4.059366692841521,
|
814 |
+
"tokens_seen": 360448000
|
815 |
+
},
|
816 |
+
{
|
817 |
+
"epoch": 0.13,
|
818 |
+
"learning_rate": 0.00043946236050690374,
|
819 |
+
"loss": 2.8095,
|
820 |
+
"theoretical_loss": 4.055189761258959,
|
821 |
+
"tokens_seen": 363724800
|
822 |
+
},
|
823 |
+
{
|
824 |
+
"epoch": 0.13,
|
825 |
+
"learning_rate": 0.00043887128806506526,
|
826 |
+
"loss": 2.7752,
|
827 |
+
"theoretical_loss": 4.051060720499127,
|
828 |
+
"tokens_seen": 367001600
|
829 |
+
},
|
830 |
+
{
|
831 |
+
"epoch": 0.13,
|
832 |
+
"learning_rate": 0.0004382802156232268,
|
833 |
+
"loss": 2.7986,
|
834 |
+
"theoretical_loss": 4.046978601165831,
|
835 |
+
"tokens_seen": 370278400
|
836 |
+
},
|
837 |
+
{
|
838 |
+
"epoch": 0.13,
|
839 |
+
"learning_rate": 0.0004376891431813883,
|
840 |
+
"loss": 2.7653,
|
841 |
+
"theoretical_loss": 4.042942461843204,
|
842 |
+
"tokens_seen": 373555200
|
843 |
+
},
|
844 |
+
{
|
845 |
+
"epoch": 0.13,
|
846 |
+
"learning_rate": 0.0004370980707395498,
|
847 |
+
"loss": 2.7984,
|
848 |
+
"theoretical_loss": 4.038951388051044,
|
849 |
+
"tokens_seen": 376832000
|
850 |
+
},
|
851 |
+
{
|
852 |
+
"epoch": 0.14,
|
853 |
+
"learning_rate": 0.0004365069982977114,
|
854 |
+
"loss": 2.8074,
|
855 |
+
"theoretical_loss": 4.035004491247873,
|
856 |
+
"tokens_seen": 380108800
|
857 |
+
},
|
858 |
+
{
|
859 |
+
"epoch": 0.14,
|
860 |
+
"learning_rate": 0.0004359159258558729,
|
861 |
+
"loss": 2.7788,
|
862 |
+
"theoretical_loss": 4.031100907879109,
|
863 |
+
"tokens_seen": 383385600
|
864 |
+
},
|
865 |
+
{
|
866 |
+
"epoch": 0.14,
|
867 |
+
"learning_rate": 0.00043532485341403444,
|
868 |
+
"loss": 2.8255,
|
869 |
+
"theoretical_loss": 4.02723979846797,
|
870 |
+
"tokens_seen": 386662400
|
871 |
+
},
|
872 |
+
{
|
873 |
+
"epoch": 0.14,
|
874 |
+
"learning_rate": 0.00043473378097219596,
|
875 |
+
"loss": 2.7839,
|
876 |
+
"theoretical_loss": 4.023420346746835,
|
877 |
+
"tokens_seen": 389939200
|
878 |
+
},
|
879 |
+
{
|
880 |
+
"epoch": 0.14,
|
881 |
+
"learning_rate": 0.0004341427085303575,
|
882 |
+
"loss": 2.7849,
|
883 |
+
"theoretical_loss": 4.019641758826938,
|
884 |
+
"tokens_seen": 393216000
|
885 |
+
},
|
886 |
+
{
|
887 |
+
"epoch": 0.14,
|
888 |
+
"learning_rate": 0.000433551636088519,
|
889 |
+
"loss": 2.7756,
|
890 |
+
"theoretical_loss": 4.015903262404413,
|
891 |
+
"tokens_seen": 396492800
|
892 |
+
},
|
893 |
+
{
|
894 |
+
"epoch": 0.14,
|
895 |
+
"learning_rate": 0.0004329605636466806,
|
896 |
+
"loss": 2.7589,
|
897 |
+
"theoretical_loss": 4.012204106000786,
|
898 |
+
"tokens_seen": 399769600
|
899 |
+
},
|
900 |
+
{
|
901 |
+
"epoch": 0.14,
|
902 |
+
"learning_rate": 0.0004323694912048421,
|
903 |
+
"loss": 2.7688,
|
904 |
+
"theoretical_loss": 4.008543558236181,
|
905 |
+
"tokens_seen": 403046400
|
906 |
+
},
|
907 |
+
{
|
908 |
+
"epoch": 0.15,
|
909 |
+
"learning_rate": 0.0004317784187630036,
|
910 |
+
"loss": 2.7891,
|
911 |
+
"theoretical_loss": 4.004920907133565,
|
912 |
+
"tokens_seen": 406323200
|
913 |
+
},
|
914 |
+
{
|
915 |
+
"epoch": 0.15,
|
916 |
+
"learning_rate": 0.00043118734632116514,
|
917 |
+
"loss": 2.7671,
|
918 |
+
"theoretical_loss": 4.001335459452449,
|
919 |
+
"tokens_seen": 409600000
|
920 |
+
},
|
921 |
+
{
|
922 |
+
"epoch": 0.15,
|
923 |
+
"learning_rate": 0.00043059627387932666,
|
924 |
+
"loss": 2.8043,
|
925 |
+
"theoretical_loss": 3.997786540050617,
|
926 |
+
"tokens_seen": 412876800
|
927 |
+
},
|
928 |
+
{
|
929 |
+
"epoch": 0.15,
|
930 |
+
"learning_rate": 0.0004300052014374882,
|
931 |
+
"loss": 2.7799,
|
932 |
+
"theoretical_loss": 3.9942734912724456,
|
933 |
+
"tokens_seen": 416153600
|
934 |
+
},
|
935 |
+
{
|
936 |
+
"epoch": 0.15,
|
937 |
+
"learning_rate": 0.0004294141289956497,
|
938 |
+
"loss": 2.724,
|
939 |
+
"theoretical_loss": 3.9907956723625375,
|
940 |
+
"tokens_seen": 419430400
|
941 |
+
},
|
942 |
+
{
|
943 |
+
"epoch": 0.15,
|
944 |
+
"learning_rate": 0.0004288230565538113,
|
945 |
+
"loss": 2.7849,
|
946 |
+
"theoretical_loss": 3.9873524589034224,
|
947 |
+
"tokens_seen": 422707200
|
948 |
+
},
|
949 |
+
{
|
950 |
+
"epoch": 0.15,
|
951 |
+
"learning_rate": 0.0004282319841119728,
|
952 |
+
"loss": 2.7838,
|
953 |
+
"theoretical_loss": 3.9839432422761556,
|
954 |
+
"tokens_seen": 425984000
|
955 |
+
},
|
956 |
+
{
|
957 |
+
"epoch": 0.15,
|
958 |
+
"learning_rate": 0.0004276409116701343,
|
959 |
+
"loss": 2.8161,
|
960 |
+
"theoretical_loss": 3.980567429142721,
|
961 |
+
"tokens_seen": 429260800
|
962 |
+
},
|
963 |
+
{
|
964 |
+
"epoch": 0.15,
|
965 |
+
"learning_rate": 0.00042704983922829584,
|
966 |
+
"loss": 2.7718,
|
967 |
+
"theoretical_loss": 3.977224440949197,
|
968 |
+
"tokens_seen": 432537600
|
969 |
+
},
|
970 |
+
{
|
971 |
+
"epoch": 0.16,
|
972 |
+
"learning_rate": 0.00042645876678645736,
|
973 |
+
"loss": 2.7692,
|
974 |
+
"theoretical_loss": 3.9739137134486917,
|
975 |
+
"tokens_seen": 435814400
|
976 |
+
},
|
977 |
+
{
|
978 |
+
"epoch": 0.16,
|
979 |
+
"learning_rate": 0.0004258676943446189,
|
980 |
+
"loss": 2.8124,
|
981 |
+
"theoretical_loss": 3.9706346962431396,
|
982 |
+
"tokens_seen": 439091200
|
983 |
+
},
|
984 |
+
{
|
985 |
+
"epoch": 0.16,
|
986 |
+
"learning_rate": 0.00042527662190278045,
|
987 |
+
"loss": 2.8134,
|
988 |
+
"theoretical_loss": 3.9673868523430564,
|
989 |
+
"tokens_seen": 442368000
|
990 |
+
},
|
991 |
+
{
|
992 |
+
"epoch": 0.16,
|
993 |
+
"learning_rate": 0.00042468554946094197,
|
994 |
+
"loss": 2.8103,
|
995 |
+
"theoretical_loss": 3.9641696577444376,
|
996 |
+
"tokens_seen": 445644800
|
997 |
+
},
|
998 |
+
{
|
999 |
+
"epoch": 0.16,
|
1000 |
+
"learning_rate": 0.0004240944770191035,
|
1001 |
+
"loss": 2.7859,
|
1002 |
+
"theoretical_loss": 3.9609826010220033,
|
1003 |
+
"tokens_seen": 448921600
|
1004 |
+
},
|
1005 |
+
{
|
1006 |
+
"epoch": 0.16,
|
1007 |
+
"learning_rate": 0.000423503404577265,
|
1008 |
+
"loss": 2.8046,
|
1009 |
+
"theoretical_loss": 3.9578251829380506,
|
1010 |
+
"tokens_seen": 452198400
|
1011 |
+
},
|
1012 |
+
{
|
1013 |
+
"epoch": 0.16,
|
1014 |
+
"learning_rate": 0.00042291233213542654,
|
1015 |
+
"loss": 2.7949,
|
1016 |
+
"theoretical_loss": 3.954696916066199,
|
1017 |
+
"tokens_seen": 455475200
|
1018 |
+
},
|
1019 |
+
{
|
1020 |
+
"epoch": 0.16,
|
1021 |
+
"learning_rate": 0.00042232125969358806,
|
1022 |
+
"loss": 2.7974,
|
1023 |
+
"theoretical_loss": 3.9515973244293643,
|
1024 |
+
"tokens_seen": 458752000
|
1025 |
+
},
|
1026 |
+
{
|
1027 |
+
"epoch": 0.17,
|
1028 |
+
"learning_rate": 0.00042173018725174963,
|
1029 |
+
"loss": 2.7685,
|
1030 |
+
"theoretical_loss": 3.948525943151326,
|
1031 |
+
"tokens_seen": 462028800
|
1032 |
+
},
|
1033 |
+
{
|
1034 |
+
"epoch": 0.17,
|
1035 |
+
"learning_rate": 0.00042113911480991115,
|
1036 |
+
"loss": 2.7622,
|
1037 |
+
"theoretical_loss": 3.9454823181212815,
|
1038 |
+
"tokens_seen": 465305600
|
1039 |
+
},
|
1040 |
+
{
|
1041 |
+
"epoch": 0.17,
|
1042 |
+
"learning_rate": 0.00042054804236807267,
|
1043 |
+
"loss": 2.7868,
|
1044 |
+
"theoretical_loss": 3.9424660056708167,
|
1045 |
+
"tokens_seen": 468582400
|
1046 |
+
},
|
1047 |
+
{
|
1048 |
+
"epoch": 0.17,
|
1049 |
+
"learning_rate": 0.0004199569699262342,
|
1050 |
+
"loss": 2.7821,
|
1051 |
+
"theoretical_loss": 3.939476572262754,
|
1052 |
+
"tokens_seen": 471859200
|
1053 |
+
},
|
1054 |
+
{
|
1055 |
+
"epoch": 0.17,
|
1056 |
+
"learning_rate": 0.00041936589748439566,
|
1057 |
+
"loss": 2.7549,
|
1058 |
+
"theoretical_loss": 3.9365135941913563,
|
1059 |
+
"tokens_seen": 475136000
|
1060 |
+
},
|
1061 |
+
{
|
1062 |
+
"epoch": 0.17,
|
1063 |
+
"learning_rate": 0.0004187748250425572,
|
1064 |
+
"loss": 2.7653,
|
1065 |
+
"theoretical_loss": 3.9335766572934023,
|
1066 |
+
"tokens_seen": 478412800
|
1067 |
+
},
|
1068 |
+
{
|
1069 |
+
"epoch": 0.17,
|
1070 |
+
"learning_rate": 0.0004181837526007187,
|
1071 |
+
"loss": 2.7314,
|
1072 |
+
"theoretical_loss": 3.9306653566696603,
|
1073 |
+
"tokens_seen": 481689600
|
1074 |
+
},
|
1075 |
+
{
|
1076 |
+
"epoch": 0.17,
|
1077 |
+
"learning_rate": 0.0004175926801588803,
|
1078 |
+
"loss": 2.7065,
|
1079 |
+
"theoretical_loss": 3.927779296416332,
|
1080 |
+
"tokens_seen": 484966400
|
1081 |
+
},
|
1082 |
+
{
|
1083 |
+
"epoch": 0.17,
|
1084 |
+
"learning_rate": 0.0004170016077170418,
|
1085 |
+
"loss": 2.7375,
|
1086 |
+
"theoretical_loss": 3.924918089366024,
|
1087 |
+
"tokens_seen": 488243200
|
1088 |
+
},
|
1089 |
+
{
|
1090 |
+
"epoch": 0.18,
|
1091 |
+
"objective/train/docs_used": 267274,
|
1092 |
+
"objective/train/instantaneous_batch_size": 32,
|
1093 |
+
"objective/train/instantaneous_microbatch_size": 32768,
|
1094 |
+
"objective/train/original_loss": 2.6940042972564697,
|
1095 |
+
"objective/train/theoretical_loss": 3.9220813568378707,
|
1096 |
+
"objective/train/tokens_used": 511980000,
|
1097 |
+
"theoretical_loss": 3.9220813568378707,
|
1098 |
+
"tokens_seen": 491520000
|
1099 |
+
},
|
1100 |
+
{
|
1101 |
+
"epoch": 0.18,
|
1102 |
+
"learning_rate": 0.0004164105352752033,
|
1103 |
+
"loss": 2.6869,
|
1104 |
+
"theoretical_loss": 3.9220813568378707,
|
1105 |
+
"tokens_seen": 491520000
|
1106 |
+
},
|
1107 |
+
{
|
1108 |
+
"epoch": 0.18,
|
1109 |
+
"learning_rate": 0.00041581946283336484,
|
1110 |
+
"loss": 2.7129,
|
1111 |
+
"theoretical_loss": 3.9192687283964096,
|
1112 |
+
"tokens_seen": 494796800
|
1113 |
+
},
|
1114 |
+
{
|
1115 |
+
"epoch": 0.18,
|
1116 |
+
"learning_rate": 0.00041522839039152636,
|
1117 |
+
"loss": 2.7344,
|
1118 |
+
"theoretical_loss": 3.9164798416188527,
|
1119 |
+
"tokens_seen": 498073600
|
1120 |
+
},
|
1121 |
+
{
|
1122 |
+
"epoch": 0.18,
|
1123 |
+
"learning_rate": 0.0004146373179496879,
|
1124 |
+
"loss": 2.6784,
|
1125 |
+
"theoretical_loss": 3.913714341870409,
|
1126 |
+
"tokens_seen": 501350400
|
1127 |
+
},
|
1128 |
+
{
|
1129 |
+
"epoch": 0.18,
|
1130 |
+
"learning_rate": 0.00041404624550784946,
|
1131 |
+
"loss": 2.6854,
|
1132 |
+
"theoretical_loss": 3.9109718820873303,
|
1133 |
+
"tokens_seen": 504627200
|
1134 |
+
},
|
1135 |
+
{
|
1136 |
+
"epoch": 0.18,
|
1137 |
+
"learning_rate": 0.000413455173066011,
|
1138 |
+
"loss": 2.694,
|
1139 |
+
"theoretical_loss": 3.9082521225673625,
|
1140 |
+
"tokens_seen": 507904000
|
1141 |
+
},
|
1142 |
+
{
|
1143 |
+
"epoch": 0.18,
|
1144 |
+
"learning_rate": 0.0004128641006241725,
|
1145 |
+
"loss": 2.6917,
|
1146 |
+
"theoretical_loss": 3.9055547307673075,
|
1147 |
+
"tokens_seen": 511180800
|
1148 |
+
},
|
1149 |
+
{
|
1150 |
+
"epoch": 0.18,
|
1151 |
+
"learning_rate": 0.000412273028182334,
|
1152 |
+
"loss": 2.7214,
|
1153 |
+
"theoretical_loss": 3.9028793811074056,
|
1154 |
+
"tokens_seen": 514457600
|
1155 |
+
},
|
1156 |
+
{
|
1157 |
+
"epoch": 0.18,
|
1158 |
+
"learning_rate": 0.00041168195574049554,
|
1159 |
+
"loss": 2.7651,
|
1160 |
+
"theoretical_loss": 3.900225754782274,
|
1161 |
+
"tokens_seen": 517734400
|
1162 |
+
},
|
1163 |
+
{
|
1164 |
+
"epoch": 0.19,
|
1165 |
+
"learning_rate": 0.00041109088329865706,
|
1166 |
+
"loss": 2.7546,
|
1167 |
+
"theoretical_loss": 3.897593539578138,
|
1168 |
+
"tokens_seen": 521011200
|
1169 |
+
},
|
1170 |
+
{
|
1171 |
+
"epoch": 0.19,
|
1172 |
+
"learning_rate": 0.00041049981085681863,
|
1173 |
+
"loss": 2.7185,
|
1174 |
+
"theoretical_loss": 3.8949824296961015,
|
1175 |
+
"tokens_seen": 524288000
|
1176 |
+
},
|
1177 |
+
{
|
1178 |
+
"epoch": 0.19,
|
1179 |
+
"learning_rate": 0.00040990873841498016,
|
1180 |
+
"loss": 2.7248,
|
1181 |
+
"theoretical_loss": 3.8923921255812353,
|
1182 |
+
"tokens_seen": 527564800
|
1183 |
+
},
|
1184 |
+
{
|
1185 |
+
"epoch": 0.19,
|
1186 |
+
"learning_rate": 0.0004093176659731417,
|
1187 |
+
"loss": 2.7266,
|
1188 |
+
"theoretical_loss": 3.8898223337572393,
|
1189 |
+
"tokens_seen": 530841600
|
1190 |
+
},
|
1191 |
+
{
|
1192 |
+
"epoch": 0.19,
|
1193 |
+
"learning_rate": 0.0004087265935313032,
|
1194 |
+
"loss": 2.6847,
|
1195 |
+
"theoretical_loss": 3.88727276666648,
|
1196 |
+
"tokens_seen": 534118400
|
1197 |
+
},
|
1198 |
+
{
|
1199 |
+
"epoch": 0.19,
|
1200 |
+
"learning_rate": 0.0004081355210894647,
|
1201 |
+
"loss": 2.7031,
|
1202 |
+
"theoretical_loss": 3.884743142515184,
|
1203 |
+
"tokens_seen": 537395200
|
1204 |
+
},
|
1205 |
+
{
|
1206 |
+
"epoch": 0.19,
|
1207 |
+
"learning_rate": 0.00040754444864762624,
|
1208 |
+
"loss": 2.6997,
|
1209 |
+
"theoretical_loss": 3.8822331851235985,
|
1210 |
+
"tokens_seen": 540672000
|
1211 |
+
},
|
1212 |
+
{
|
1213 |
+
"epoch": 0.19,
|
1214 |
+
"learning_rate": 0.0004069533762057878,
|
1215 |
+
"loss": 2.722,
|
1216 |
+
"theoretical_loss": 3.87974262378093,
|
1217 |
+
"tokens_seen": 543948800
|
1218 |
+
},
|
1219 |
+
{
|
1220 |
+
"epoch": 0.2,
|
1221 |
+
"learning_rate": 0.00040636230376394933,
|
1222 |
+
"loss": 2.6597,
|
1223 |
+
"theoretical_loss": 3.877271193104873,
|
1224 |
+
"tokens_seen": 547225600
|
1225 |
+
},
|
1226 |
+
{
|
1227 |
+
"epoch": 0.2,
|
1228 |
+
"learning_rate": 0.00040577123132211085,
|
1229 |
+
"loss": 2.6842,
|
1230 |
+
"theoretical_loss": 3.8748186329055736,
|
1231 |
+
"tokens_seen": 550502400
|
1232 |
+
},
|
1233 |
+
{
|
1234 |
+
"epoch": 0.2,
|
1235 |
+
"learning_rate": 0.0004051801588802724,
|
1236 |
+
"loss": 2.6774,
|
1237 |
+
"theoretical_loss": 3.87238468805384,
|
1238 |
+
"tokens_seen": 553779200
|
1239 |
+
},
|
1240 |
+
{
|
1241 |
+
"epoch": 0.2,
|
1242 |
+
"learning_rate": 0.0004045890864384339,
|
1243 |
+
"loss": 2.6673,
|
1244 |
+
"theoretical_loss": 3.8699691083534633,
|
1245 |
+
"tokens_seen": 557056000
|
1246 |
+
},
|
1247 |
+
{
|
1248 |
+
"epoch": 0.2,
|
1249 |
+
"learning_rate": 0.0004039980139965954,
|
1250 |
+
"loss": 2.6858,
|
1251 |
+
"theoretical_loss": 3.8675716484174907,
|
1252 |
+
"tokens_seen": 560332800
|
1253 |
+
},
|
1254 |
+
{
|
1255 |
+
"epoch": 0.2,
|
1256 |
+
"learning_rate": 0.00040340694155475694,
|
1257 |
+
"loss": 2.6698,
|
1258 |
+
"theoretical_loss": 3.8651920675482936,
|
1259 |
+
"tokens_seen": 563609600
|
1260 |
+
},
|
1261 |
+
{
|
1262 |
+
"epoch": 0.2,
|
1263 |
+
"learning_rate": 0.0004028158691129185,
|
1264 |
+
"loss": 2.7021,
|
1265 |
+
"theoretical_loss": 3.862830129621318,
|
1266 |
+
"tokens_seen": 566886400
|
1267 |
+
},
|
1268 |
+
{
|
1269 |
+
"epoch": 0.2,
|
1270 |
+
"learning_rate": 0.00040222479667108003,
|
1271 |
+
"loss": 2.7151,
|
1272 |
+
"theoretical_loss": 3.8604856029723575,
|
1273 |
+
"tokens_seen": 570163200
|
1274 |
+
},
|
1275 |
+
{
|
1276 |
+
"epoch": 0.2,
|
1277 |
+
"learning_rate": 0.00040163372422924155,
|
1278 |
+
"loss": 2.6881,
|
1279 |
+
"theoretical_loss": 3.8581582602882447,
|
1280 |
+
"tokens_seen": 573440000
|
1281 |
+
},
|
1282 |
+
{
|
1283 |
+
"epoch": 0.21,
|
1284 |
+
"learning_rate": 0.0004010426517874031,
|
1285 |
+
"loss": 2.7237,
|
1286 |
+
"theoretical_loss": 3.8558478785008203,
|
1287 |
+
"tokens_seen": 576716800
|
1288 |
+
},
|
1289 |
+
{
|
1290 |
+
"epoch": 0.21,
|
1291 |
+
"learning_rate": 0.0004004515793455646,
|
1292 |
+
"loss": 2.7083,
|
1293 |
+
"theoretical_loss": 3.8535542386840778,
|
1294 |
+
"tokens_seen": 579993600
|
1295 |
+
},
|
1296 |
+
{
|
1297 |
+
"epoch": 0.21,
|
1298 |
+
"learning_rate": 0.0003998605069037261,
|
1299 |
+
"loss": 2.6988,
|
1300 |
+
"theoretical_loss": 3.8512771259543586,
|
1301 |
+
"tokens_seen": 583270400
|
1302 |
+
},
|
1303 |
+
{
|
1304 |
+
"epoch": 0.21,
|
1305 |
+
"learning_rate": 0.0003992694344618877,
|
1306 |
+
"loss": 2.7261,
|
1307 |
+
"theoretical_loss": 3.8490163293735082,
|
1308 |
+
"tokens_seen": 586547200
|
1309 |
+
},
|
1310 |
+
{
|
1311 |
+
"epoch": 0.21,
|
1312 |
+
"learning_rate": 0.0003986783620200492,
|
1313 |
+
"loss": 2.7215,
|
1314 |
+
"theoretical_loss": 3.8467716418548648,
|
1315 |
+
"tokens_seen": 589824000
|
1316 |
+
},
|
1317 |
+
{
|
1318 |
+
"epoch": 0.21,
|
1319 |
+
"learning_rate": 0.00039808728957821073,
|
1320 |
+
"loss": 2.7196,
|
1321 |
+
"theoretical_loss": 3.844542860072007,
|
1322 |
+
"tokens_seen": 593100800
|
1323 |
+
},
|
1324 |
+
{
|
1325 |
+
"epoch": 0.21,
|
1326 |
+
"learning_rate": 0.00039749621713637225,
|
1327 |
+
"loss": 2.736,
|
1328 |
+
"theoretical_loss": 3.8423297843701496,
|
1329 |
+
"tokens_seen": 596377600
|
1330 |
+
},
|
1331 |
+
{
|
1332 |
+
"epoch": 0.21,
|
1333 |
+
"learning_rate": 0.0003969051446945338,
|
1334 |
+
"loss": 2.7189,
|
1335 |
+
"theoretical_loss": 3.8401322186800995,
|
1336 |
+
"tokens_seen": 599654400
|
1337 |
+
},
|
1338 |
+
{
|
1339 |
+
"epoch": 0.22,
|
1340 |
+
"learning_rate": 0.0003963140722526953,
|
1341 |
+
"loss": 2.7546,
|
1342 |
+
"theoretical_loss": 3.83794997043469,
|
1343 |
+
"tokens_seen": 602931200
|
1344 |
+
},
|
1345 |
+
{
|
1346 |
+
"epoch": 0.22,
|
1347 |
+
"learning_rate": 0.00039572299981085687,
|
1348 |
+
"loss": 2.7572,
|
1349 |
+
"theoretical_loss": 3.8357828504876004,
|
1350 |
+
"tokens_seen": 606208000
|
1351 |
+
},
|
1352 |
+
{
|
1353 |
+
"epoch": 0.22,
|
1354 |
+
"learning_rate": 0.0003951319273690184,
|
1355 |
+
"loss": 2.7173,
|
1356 |
+
"theoretical_loss": 3.833630673034487,
|
1357 |
+
"tokens_seen": 609484800
|
1358 |
+
},
|
1359 |
+
{
|
1360 |
+
"epoch": 0.22,
|
1361 |
+
"learning_rate": 0.0003945408549271799,
|
1362 |
+
"loss": 2.7306,
|
1363 |
+
"theoretical_loss": 3.831493255536345,
|
1364 |
+
"tokens_seen": 612761600
|
1365 |
+
},
|
1366 |
+
{
|
1367 |
+
"epoch": 0.22,
|
1368 |
+
"learning_rate": 0.00039394978248534143,
|
1369 |
+
"loss": 2.7288,
|
1370 |
+
"theoretical_loss": 3.8293704186450253,
|
1371 |
+
"tokens_seen": 616038400
|
1372 |
+
},
|
1373 |
+
{
|
1374 |
+
"epoch": 0.22,
|
1375 |
+
"learning_rate": 0.00039335871004350295,
|
1376 |
+
"loss": 2.723,
|
1377 |
+
"theoretical_loss": 3.827261986130839,
|
1378 |
+
"tokens_seen": 619315200
|
1379 |
+
},
|
1380 |
+
{
|
1381 |
+
"epoch": 0.22,
|
1382 |
+
"learning_rate": 0.0003927676376016645,
|
1383 |
+
"loss": 2.6673,
|
1384 |
+
"theoretical_loss": 3.825167784812175,
|
1385 |
+
"tokens_seen": 622592000
|
1386 |
+
},
|
1387 |
+
{
|
1388 |
+
"epoch": 0.22,
|
1389 |
+
"learning_rate": 0.000392176565159826,
|
1390 |
+
"loss": 2.6865,
|
1391 |
+
"theoretical_loss": 3.823087644487069,
|
1392 |
+
"tokens_seen": 625868800
|
1393 |
+
},
|
1394 |
+
{
|
1395 |
+
"epoch": 0.22,
|
1396 |
+
"learning_rate": 0.00039158549271798757,
|
1397 |
+
"loss": 2.6744,
|
1398 |
+
"theoretical_loss": 3.8210213978666565,
|
1399 |
+
"tokens_seen": 629145600
|
1400 |
+
},
|
1401 |
+
{
|
1402 |
+
"epoch": 0.23,
|
1403 |
+
"learning_rate": 0.0003909944202761491,
|
1404 |
+
"loss": 2.6884,
|
1405 |
+
"theoretical_loss": 3.8189688805104476,
|
1406 |
+
"tokens_seen": 632422400
|
1407 |
+
},
|
1408 |
+
{
|
1409 |
+
"epoch": 0.23,
|
1410 |
+
"learning_rate": 0.00039040334783431056,
|
1411 |
+
"loss": 2.6735,
|
1412 |
+
"theoretical_loss": 3.816929930763374,
|
1413 |
+
"tokens_seen": 635699200
|
1414 |
+
},
|
1415 |
+
{
|
1416 |
+
"epoch": 0.23,
|
1417 |
+
"learning_rate": 0.0003898122753924721,
|
1418 |
+
"loss": 2.702,
|
1419 |
+
"theoretical_loss": 3.8149043896945347,
|
1420 |
+
"tokens_seen": 638976000
|
1421 |
+
},
|
1422 |
+
{
|
1423 |
+
"epoch": 0.23,
|
1424 |
+
"learning_rate": 0.0003892212029506336,
|
1425 |
+
"loss": 2.6783,
|
1426 |
+
"theoretical_loss": 3.812892101037601,
|
1427 |
+
"tokens_seen": 642252800
|
1428 |
+
},
|
1429 |
+
{
|
1430 |
+
"epoch": 0.23,
|
1431 |
+
"learning_rate": 0.0003886301305087951,
|
1432 |
+
"loss": 2.6662,
|
1433 |
+
"theoretical_loss": 3.81089291113282,
|
1434 |
+
"tokens_seen": 645529600
|
1435 |
+
},
|
1436 |
+
{
|
1437 |
+
"epoch": 0.23,
|
1438 |
+
"learning_rate": 0.0003880390580669567,
|
1439 |
+
"loss": 2.6817,
|
1440 |
+
"theoretical_loss": 3.8089066688705673,
|
1441 |
+
"tokens_seen": 648806400
|
1442 |
+
},
|
1443 |
+
{
|
1444 |
+
"epoch": 0.23,
|
1445 |
+
"learning_rate": 0.0003874479856251182,
|
1446 |
+
"loss": 2.6423,
|
1447 |
+
"theoretical_loss": 3.8069332256363992,
|
1448 |
+
"tokens_seen": 652083200
|
1449 |
+
},
|
1450 |
+
{
|
1451 |
+
"debugging/Self-BLEU-5": 0.4830888722420741,
|
1452 |
+
"debugging/distinct-1-grams": 0.8177040787129198,
|
1453 |
+
"debugging/distinct-2-grams": 0.968608367204819,
|
1454 |
+
"debugging/entropy-1-grams": 5.3590605074024396,
|
1455 |
+
"debugging/entropy-2-grams": 5.9214073904158475,
|
1456 |
+
"debugging/length": 564.1666666666666,
|
1457 |
+
"debugging/num_segments": 6,
|
1458 |
+
"epoch": 0.23,
|
1459 |
+
"objective/train/docs_used": 351925,
|
1460 |
+
"objective/train/instantaneous_batch_size": 32,
|
1461 |
+
"objective/train/instantaneous_microbatch_size": 32768,
|
1462 |
+
"objective/train/original_loss": 2.8214328289031982,
|
1463 |
+
"objective/train/theoretical_loss": 3.80497243525756,
|
1464 |
+
"objective/train/tokens_used": 675820000,
|
1465 |
+
"theoretical_loss": 3.80497243525756,
|
1466 |
+
"tokens_seen": 655360000
|
1467 |
+
},
|
1468 |
+
{
|
1469 |
+
"epoch": 0.23,
|
1470 |
+
"learning_rate": 0.00038685691318327974,
|
1471 |
+
"loss": 2.6739,
|
1472 |
+
"theoretical_loss": 3.80497243525756,
|
1473 |
+
"tokens_seen": 655360000
|
1474 |
+
},
|
1475 |
+
{
|
1476 |
+
"epoch": 0.24,
|
1477 |
+
"learning_rate": 0.00038626584074144126,
|
1478 |
+
"loss": 2.6231,
|
1479 |
+
"theoretical_loss": 3.8030241539508958,
|
1480 |
+
"tokens_seen": 658636800
|
1481 |
+
},
|
1482 |
+
{
|
1483 |
+
"epoch": 0.24,
|
1484 |
+
"learning_rate": 0.0003856747682996028,
|
1485 |
+
"loss": 2.6252,
|
1486 |
+
"theoretical_loss": 3.8010882402721324,
|
1487 |
+
"tokens_seen": 661913600
|
1488 |
+
},
|
1489 |
+
{
|
1490 |
+
"epoch": 0.24,
|
1491 |
+
"learning_rate": 0.0003850836958577643,
|
1492 |
+
"loss": 2.6658,
|
1493 |
+
"theoretical_loss": 3.7991645550664757,
|
1494 |
+
"tokens_seen": 665190400
|
1495 |
+
},
|
1496 |
+
{
|
1497 |
+
"epoch": 0.24,
|
1498 |
+
"learning_rate": 0.0003844926234159259,
|
1499 |
+
"loss": 2.7018,
|
1500 |
+
"theoretical_loss": 3.797252961420492,
|
1501 |
+
"tokens_seen": 668467200
|
1502 |
+
},
|
1503 |
+
{
|
1504 |
+
"epoch": 0.24,
|
1505 |
+
"learning_rate": 0.0003839015509740874,
|
1506 |
+
"loss": 2.689,
|
1507 |
+
"theoretical_loss": 3.795353324615228,
|
1508 |
+
"tokens_seen": 671744000
|
1509 |
+
},
|
1510 |
+
{
|
1511 |
+
"epoch": 0.24,
|
1512 |
+
"learning_rate": 0.0003833104785322489,
|
1513 |
+
"loss": 2.7145,
|
1514 |
+
"theoretical_loss": 3.793465512080541,
|
1515 |
+
"tokens_seen": 675020800
|
1516 |
+
},
|
1517 |
+
{
|
1518 |
+
"epoch": 0.24,
|
1519 |
+
"learning_rate": 0.00038271940609041044,
|
1520 |
+
"loss": 2.7397,
|
1521 |
+
"theoretical_loss": 3.791589393350587,
|
1522 |
+
"tokens_seen": 678297600
|
1523 |
+
},
|
1524 |
+
{
|
1525 |
+
"epoch": 0.24,
|
1526 |
+
"learning_rate": 0.00038212833364857196,
|
1527 |
+
"loss": 2.6861,
|
1528 |
+
"theoretical_loss": 3.7897248400204475,
|
1529 |
+
"tokens_seen": 681574400
|
1530 |
+
},
|
1531 |
+
{
|
1532 |
+
"epoch": 0.24,
|
1533 |
+
"learning_rate": 0.0003815372612067335,
|
1534 |
+
"loss": 2.6654,
|
1535 |
+
"theoretical_loss": 3.7878717257038534,
|
1536 |
+
"tokens_seen": 684851200
|
1537 |
+
},
|
1538 |
+
{
|
1539 |
+
"epoch": 0.25,
|
1540 |
+
"learning_rate": 0.000380946188764895,
|
1541 |
+
"loss": 2.6726,
|
1542 |
+
"theoretical_loss": 3.7860299259919685,
|
1543 |
+
"tokens_seen": 688128000
|
1544 |
+
},
|
1545 |
+
{
|
1546 |
+
"epoch": 0.25,
|
1547 |
+
"learning_rate": 0.0003803551163230566,
|
1548 |
+
"loss": 2.6455,
|
1549 |
+
"theoretical_loss": 3.7841993184132114,
|
1550 |
+
"tokens_seen": 691404800
|
1551 |
+
},
|
1552 |
+
{
|
1553 |
+
"epoch": 0.25,
|
1554 |
+
"learning_rate": 0.0003797640438812181,
|
1555 |
+
"loss": 2.6702,
|
1556 |
+
"theoretical_loss": 3.78237978239408,
|
1557 |
+
"tokens_seen": 694681600
|
1558 |
+
},
|
1559 |
+
{
|
1560 |
+
"epoch": 0.25,
|
1561 |
+
"learning_rate": 0.0003791729714393796,
|
1562 |
+
"loss": 2.6917,
|
1563 |
+
"theoretical_loss": 3.780571199220942,
|
1564 |
+
"tokens_seen": 697958400
|
1565 |
+
},
|
1566 |
+
{
|
1567 |
+
"epoch": 0.25,
|
1568 |
+
"learning_rate": 0.00037858189899754114,
|
1569 |
+
"loss": 2.6649,
|
1570 |
+
"theoretical_loss": 3.7787734520027803,
|
1571 |
+
"tokens_seen": 701235200
|
1572 |
+
},
|
1573 |
+
{
|
1574 |
+
"epoch": 0.25,
|
1575 |
+
"learning_rate": 0.00037799082655570266,
|
1576 |
+
"loss": 2.6522,
|
1577 |
+
"theoretical_loss": 3.7769864256348455,
|
1578 |
+
"tokens_seen": 704512000
|
1579 |
+
},
|
1580 |
+
{
|
1581 |
+
"epoch": 0.25,
|
1582 |
+
"learning_rate": 0.00037741157556270093,
|
1583 |
+
"loss": 2.7032,
|
1584 |
+
"theoretical_loss": 3.775210006763202,
|
1585 |
+
"tokens_seen": 707788800
|
1586 |
+
},
|
1587 |
+
{
|
1588 |
+
"epoch": 0.25,
|
1589 |
+
"learning_rate": 0.0003768205031208625,
|
1590 |
+
"loss": 2.7074,
|
1591 |
+
"theoretical_loss": 3.7734440837501406,
|
1592 |
+
"tokens_seen": 711065600
|
1593 |
+
},
|
1594 |
+
{
|
1595 |
+
"epoch": 0.26,
|
1596 |
+
"learning_rate": 0.000376229430679024,
|
1597 |
+
"loss": 2.7516,
|
1598 |
+
"theoretical_loss": 3.7716885466404246,
|
1599 |
+
"tokens_seen": 714342400
|
1600 |
+
},
|
1601 |
+
{
|
1602 |
+
"epoch": 0.26,
|
1603 |
+
"learning_rate": 0.00037563835823718555,
|
1604 |
+
"loss": 2.7007,
|
1605 |
+
"theoretical_loss": 3.769943287128357,
|
1606 |
+
"tokens_seen": 717619200
|
1607 |
+
},
|
1608 |
+
{
|
1609 |
+
"epoch": 0.26,
|
1610 |
+
"learning_rate": 0.00037504728579534707,
|
1611 |
+
"loss": 2.7292,
|
1612 |
+
"theoretical_loss": 3.7682081985256364,
|
1613 |
+
"tokens_seen": 720896000
|
1614 |
+
},
|
1615 |
+
{
|
1616 |
+
"epoch": 0.26,
|
1617 |
+
"learning_rate": 0.0003744562133535086,
|
1618 |
+
"loss": 2.7581,
|
1619 |
+
"theoretical_loss": 3.7664831757299795,
|
1620 |
+
"tokens_seen": 724172800
|
1621 |
+
},
|
1622 |
+
{
|
1623 |
+
"epoch": 0.26,
|
1624 |
+
"learning_rate": 0.0003738651409116701,
|
1625 |
+
"loss": 2.6871,
|
1626 |
+
"theoretical_loss": 3.7647681151944976,
|
1627 |
+
"tokens_seen": 727449600
|
1628 |
+
},
|
1629 |
+
{
|
1630 |
+
"epoch": 0.26,
|
1631 |
+
"learning_rate": 0.0003732740684698317,
|
1632 |
+
"loss": 2.6436,
|
1633 |
+
"theoretical_loss": 3.7630629148977937,
|
1634 |
+
"tokens_seen": 730726400
|
1635 |
+
},
|
1636 |
+
{
|
1637 |
+
"epoch": 0.26,
|
1638 |
+
"learning_rate": 0.0003726829960279932,
|
1639 |
+
"loss": 2.6503,
|
1640 |
+
"theoretical_loss": 3.761367474314768,
|
1641 |
+
"tokens_seen": 734003200
|
1642 |
+
},
|
1643 |
+
{
|
1644 |
+
"epoch": 0.26,
|
1645 |
+
"learning_rate": 0.0003720919235861547,
|
1646 |
+
"loss": 2.676,
|
1647 |
+
"theoretical_loss": 3.7596816943881084,
|
1648 |
+
"tokens_seen": 737280000
|
1649 |
+
},
|
1650 |
+
{
|
1651 |
+
"epoch": 0.26,
|
1652 |
+
"learning_rate": 0.00037150085114431625,
|
1653 |
+
"loss": 2.6658,
|
1654 |
+
"theoretical_loss": 3.758005477500451,
|
1655 |
+
"tokens_seen": 740556800
|
1656 |
+
},
|
1657 |
+
{
|
1658 |
+
"epoch": 0.27,
|
1659 |
+
"learning_rate": 0.00037090977870247777,
|
1660 |
+
"loss": 2.6606,
|
1661 |
+
"theoretical_loss": 3.756338727447186,
|
1662 |
+
"tokens_seen": 743833600
|
1663 |
+
},
|
1664 |
+
{
|
1665 |
+
"epoch": 0.27,
|
1666 |
+
"learning_rate": 0.0003703187062606393,
|
1667 |
+
"loss": 2.6554,
|
1668 |
+
"theoretical_loss": 3.7546813494098945,
|
1669 |
+
"tokens_seen": 747110400
|
1670 |
+
},
|
1671 |
+
{
|
1672 |
+
"epoch": 0.27,
|
1673 |
+
"learning_rate": 0.0003697276338188008,
|
1674 |
+
"loss": 2.6705,
|
1675 |
+
"theoretical_loss": 3.7530332499304007,
|
1676 |
+
"tokens_seen": 750387200
|
1677 |
+
},
|
1678 |
+
{
|
1679 |
+
"epoch": 0.27,
|
1680 |
+
"learning_rate": 0.0003691365613769624,
|
1681 |
+
"loss": 2.6521,
|
1682 |
+
"theoretical_loss": 3.7513943368854195,
|
1683 |
+
"tokens_seen": 753664000
|
1684 |
+
},
|
1685 |
+
{
|
1686 |
+
"epoch": 0.27,
|
1687 |
+
"learning_rate": 0.0003685454889351239,
|
1688 |
+
"loss": 2.6524,
|
1689 |
+
"theoretical_loss": 3.7497645194617863,
|
1690 |
+
"tokens_seen": 756940800
|
1691 |
+
},
|
1692 |
+
{
|
1693 |
+
"epoch": 0.27,
|
1694 |
+
"learning_rate": 0.0003679544164932854,
|
1695 |
+
"loss": 2.6184,
|
1696 |
+
"theoretical_loss": 3.748143708132246,
|
1697 |
+
"tokens_seen": 760217600
|
1698 |
+
},
|
1699 |
+
{
|
1700 |
+
"epoch": 0.27,
|
1701 |
+
"learning_rate": 0.00036736334405144695,
|
1702 |
+
"loss": 2.6389,
|
1703 |
+
"theoretical_loss": 3.7465318146317994,
|
1704 |
+
"tokens_seen": 763494400
|
1705 |
+
},
|
1706 |
+
{
|
1707 |
+
"epoch": 0.27,
|
1708 |
+
"learning_rate": 0.00036677227160960847,
|
1709 |
+
"loss": 2.596,
|
1710 |
+
"theoretical_loss": 3.7449287519345766,
|
1711 |
+
"tokens_seen": 766771200
|
1712 |
+
},
|
1713 |
+
{
|
1714 |
+
"epoch": 0.28,
|
1715 |
+
"learning_rate": 0.00036618119916777,
|
1716 |
+
"loss": 2.6259,
|
1717 |
+
"theoretical_loss": 3.7433344342312385,
|
1718 |
+
"tokens_seen": 770048000
|
1719 |
+
},
|
1720 |
+
{
|
1721 |
+
"epoch": 0.28,
|
1722 |
+
"learning_rate": 0.00036559012672593156,
|
1723 |
+
"loss": 2.6141,
|
1724 |
+
"theoretical_loss": 3.7417487769068756,
|
1725 |
+
"tokens_seen": 773324800
|
1726 |
+
},
|
1727 |
+
{
|
1728 |
+
"epoch": 0.28,
|
1729 |
+
"learning_rate": 0.0003649990542840931,
|
1730 |
+
"loss": 2.6488,
|
1731 |
+
"theoretical_loss": 3.7401716965194076,
|
1732 |
+
"tokens_seen": 776601600
|
1733 |
+
},
|
1734 |
+
{
|
1735 |
+
"epoch": 0.28,
|
1736 |
+
"learning_rate": 0.0003644079818422546,
|
1737 |
+
"loss": 2.6692,
|
1738 |
+
"theoretical_loss": 3.738603110778461,
|
1739 |
+
"tokens_seen": 779878400
|
1740 |
+
},
|
1741 |
+
{
|
1742 |
+
"epoch": 0.28,
|
1743 |
+
"learning_rate": 0.0003638169094004161,
|
1744 |
+
"loss": 2.6684,
|
1745 |
+
"theoretical_loss": 3.73704293852471,
|
1746 |
+
"tokens_seen": 783155200
|
1747 |
+
},
|
1748 |
+
{
|
1749 |
+
"epoch": 0.28,
|
1750 |
+
"learning_rate": 0.00036322583695857765,
|
1751 |
+
"loss": 2.6492,
|
1752 |
+
"theoretical_loss": 3.7354910997096793,
|
1753 |
+
"tokens_seen": 786432000
|
1754 |
+
},
|
1755 |
+
{
|
1756 |
+
"epoch": 0.28,
|
1757 |
+
"learning_rate": 0.00036263476451673917,
|
1758 |
+
"loss": 2.6465,
|
1759 |
+
"theoretical_loss": 3.7339475153759825,
|
1760 |
+
"tokens_seen": 789708800
|
1761 |
+
},
|
1762 |
+
{
|
1763 |
+
"epoch": 0.28,
|
1764 |
+
"learning_rate": 0.00036204369207490074,
|
1765 |
+
"loss": 2.6071,
|
1766 |
+
"theoretical_loss": 3.732412107638,
|
1767 |
+
"tokens_seen": 792985600
|
1768 |
+
},
|
1769 |
+
{
|
1770 |
+
"epoch": 0.28,
|
1771 |
+
"learning_rate": 0.00036145261963306226,
|
1772 |
+
"loss": 2.64,
|
1773 |
+
"theoretical_loss": 3.7308847996629724,
|
1774 |
+
"tokens_seen": 796262400
|
1775 |
+
},
|
1776 |
+
{
|
1777 |
+
"epoch": 0.29,
|
1778 |
+
"learning_rate": 0.0003608615471912238,
|
1779 |
+
"loss": 2.6283,
|
1780 |
+
"theoretical_loss": 3.7293655156525043,
|
1781 |
+
"tokens_seen": 799539200
|
1782 |
+
},
|
1783 |
+
{
|
1784 |
+
"epoch": 0.29,
|
1785 |
+
"learning_rate": 0.0003602704747493853,
|
1786 |
+
"loss": 2.6181,
|
1787 |
+
"theoretical_loss": 3.727854180824469,
|
1788 |
+
"tokens_seen": 802816000
|
1789 |
+
},
|
1790 |
+
{
|
1791 |
+
"epoch": 0.29,
|
1792 |
+
"learning_rate": 0.0003596794023075468,
|
1793 |
+
"loss": 2.5809,
|
1794 |
+
"theoretical_loss": 3.7263507213952978,
|
1795 |
+
"tokens_seen": 806092800
|
1796 |
+
},
|
1797 |
+
{
|
1798 |
+
"epoch": 0.29,
|
1799 |
+
"learning_rate": 0.00035908832986570834,
|
1800 |
+
"loss": 2.5738,
|
1801 |
+
"theoretical_loss": 3.724855064562658,
|
1802 |
+
"tokens_seen": 809369600
|
1803 |
+
},
|
1804 |
+
{
|
1805 |
+
"epoch": 0.29,
|
1806 |
+
"learning_rate": 0.0003584972574238699,
|
1807 |
+
"loss": 2.602,
|
1808 |
+
"theoretical_loss": 3.723367138488488,
|
1809 |
+
"tokens_seen": 812646400
|
1810 |
+
},
|
1811 |
+
{
|
1812 |
+
"epoch": 0.29,
|
1813 |
+
"learning_rate": 0.00035790618498203144,
|
1814 |
+
"loss": 2.6405,
|
1815 |
+
"theoretical_loss": 3.7218868722824014,
|
1816 |
+
"tokens_seen": 815923200
|
1817 |
+
},
|
1818 |
+
{
|
1819 |
+
"epoch": 0.29,
|
1820 |
+
"objective/train/docs_used": 436776,
|
1821 |
+
"objective/train/instantaneous_batch_size": 32,
|
1822 |
+
"objective/train/instantaneous_microbatch_size": 32768,
|
1823 |
+
"objective/train/original_loss": 2.661590576171875,
|
1824 |
+
"objective/train/theoretical_loss": 3.7204141959854384,
|
1825 |
+
"objective/train/tokens_used": 839660000,
|
1826 |
+
"theoretical_loss": 3.7204141959854384,
|
1827 |
+
"tokens_seen": 819200000
|
1828 |
+
},
|
1829 |
+
{
|
1830 |
+
"epoch": 0.29,
|
1831 |
+
"learning_rate": 0.00035731511254019296,
|
1832 |
+
"loss": 2.641,
|
1833 |
+
"theoretical_loss": 3.7204141959854384,
|
1834 |
+
"tokens_seen": 819200000
|
1835 |
+
},
|
1836 |
+
{
|
1837 |
+
"epoch": 0.29,
|
1838 |
+
"learning_rate": 0.0003567240400983545,
|
1839 |
+
"loss": 2.6069,
|
1840 |
+
"theoretical_loss": 3.718949040554162,
|
1841 |
+
"tokens_seen": 822476800
|
1842 |
+
},
|
1843 |
+
{
|
1844 |
+
"epoch": 0.29,
|
1845 |
+
"learning_rate": 0.000356132967656516,
|
1846 |
+
"loss": 2.6626,
|
1847 |
+
"theoretical_loss": 3.7174913378450833,
|
1848 |
+
"tokens_seen": 825753600
|
1849 |
+
},
|
1850 |
+
{
|
1851 |
+
"epoch": 0.3,
|
1852 |
+
"learning_rate": 0.0003555418952146775,
|
1853 |
+
"loss": 2.6566,
|
1854 |
+
"theoretical_loss": 3.7160410205994183,
|
1855 |
+
"tokens_seen": 829030400
|
1856 |
+
},
|
1857 |
+
{
|
1858 |
+
"epoch": 0.3,
|
1859 |
+
"learning_rate": 0.00035495082277283904,
|
1860 |
+
"loss": 2.6638,
|
1861 |
+
"theoretical_loss": 3.7145980224281585,
|
1862 |
+
"tokens_seen": 832307200
|
1863 |
+
},
|
1864 |
+
{
|
1865 |
+
"epoch": 0.3,
|
1866 |
+
"learning_rate": 0.0003543597503310006,
|
1867 |
+
"loss": 2.6863,
|
1868 |
+
"theoretical_loss": 3.713162277797449,
|
1869 |
+
"tokens_seen": 835584000
|
1870 |
+
},
|
1871 |
+
{
|
1872 |
+
"epoch": 0.3,
|
1873 |
+
"learning_rate": 0.00035376867788916214,
|
1874 |
+
"loss": 2.6587,
|
1875 |
+
"theoretical_loss": 3.7117337220142748,
|
1876 |
+
"tokens_seen": 838860800
|
1877 |
+
},
|
1878 |
+
{
|
1879 |
+
"epoch": 0.3,
|
1880 |
+
"learning_rate": 0.00035317760544732366,
|
1881 |
+
"loss": 2.6476,
|
1882 |
+
"theoretical_loss": 3.7103122912124364,
|
1883 |
+
"tokens_seen": 842137600
|
1884 |
+
},
|
1885 |
+
{
|
1886 |
+
"epoch": 0.3,
|
1887 |
+
"learning_rate": 0.0003525865330054852,
|
1888 |
+
"loss": 2.6524,
|
1889 |
+
"theoretical_loss": 3.7088979223388128,
|
1890 |
+
"tokens_seen": 845414400
|
1891 |
+
},
|
1892 |
+
{
|
1893 |
+
"epoch": 0.3,
|
1894 |
+
"learning_rate": 0.00035199546056364665,
|
1895 |
+
"loss": 2.6555,
|
1896 |
+
"theoretical_loss": 3.70749055313991,
|
1897 |
+
"tokens_seen": 848691200
|
1898 |
+
},
|
1899 |
+
{
|
1900 |
+
"epoch": 0.3,
|
1901 |
+
"learning_rate": 0.00035140438812180817,
|
1902 |
+
"loss": 2.6511,
|
1903 |
+
"theoretical_loss": 3.7060901221486766,
|
1904 |
+
"tokens_seen": 851968000
|
1905 |
+
},
|
1906 |
+
{
|
1907 |
+
"epoch": 0.31,
|
1908 |
+
"learning_rate": 0.00035081331567996974,
|
1909 |
+
"loss": 2.664,
|
1910 |
+
"theoretical_loss": 3.704696568671591,
|
1911 |
+
"tokens_seen": 855244800
|
1912 |
+
},
|
1913 |
+
{
|
1914 |
+
"epoch": 0.31,
|
1915 |
+
"learning_rate": 0.00035022224323813127,
|
1916 |
+
"loss": 2.6587,
|
1917 |
+
"theoretical_loss": 3.7033098327760063,
|
1918 |
+
"tokens_seen": 858521600
|
1919 |
+
},
|
1920 |
+
{
|
1921 |
+
"epoch": 0.31,
|
1922 |
+
"learning_rate": 0.0003496429922451296,
|
1923 |
+
"loss": 2.61,
|
1924 |
+
"theoretical_loss": 3.7019298552777533,
|
1925 |
+
"tokens_seen": 861798400
|
1926 |
+
},
|
1927 |
+
{
|
1928 |
+
"epoch": 0.31,
|
1929 |
+
"learning_rate": 0.0003490519198032911,
|
1930 |
+
"loss": 2.6281,
|
1931 |
+
"theoretical_loss": 3.700556577728988,
|
1932 |
+
"tokens_seen": 865075200
|
1933 |
+
},
|
1934 |
+
{
|
1935 |
+
"epoch": 0.31,
|
1936 |
+
"learning_rate": 0.00034846084736145263,
|
1937 |
+
"loss": 2.629,
|
1938 |
+
"theoretical_loss": 3.6991899424062815,
|
1939 |
+
"tokens_seen": 868352000
|
1940 |
+
},
|
1941 |
+
{
|
1942 |
+
"epoch": 0.31,
|
1943 |
+
"learning_rate": 0.00034786977491961415,
|
1944 |
+
"loss": 2.649,
|
1945 |
+
"theoretical_loss": 3.697829892298951,
|
1946 |
+
"tokens_seen": 871628800
|
1947 |
+
},
|
1948 |
+
{
|
1949 |
+
"epoch": 0.31,
|
1950 |
+
"learning_rate": 0.00034727870247777573,
|
1951 |
+
"loss": 2.624,
|
1952 |
+
"theoretical_loss": 3.696476371097618,
|
1953 |
+
"tokens_seen": 874905600
|
1954 |
+
},
|
1955 |
+
{
|
1956 |
+
"epoch": 0.31,
|
1957 |
+
"learning_rate": 0.00034668763003593725,
|
1958 |
+
"loss": 2.6113,
|
1959 |
+
"theoretical_loss": 3.695129323182993,
|
1960 |
+
"tokens_seen": 878182400
|
1961 |
+
},
|
1962 |
+
{
|
1963 |
+
"epoch": 0.31,
|
1964 |
+
"learning_rate": 0.00034609655759409877,
|
1965 |
+
"loss": 2.5968,
|
1966 |
+
"theoretical_loss": 3.693788693614879,
|
1967 |
+
"tokens_seen": 881459200
|
1968 |
+
},
|
1969 |
+
{
|
1970 |
+
"epoch": 0.32,
|
1971 |
+
"learning_rate": 0.0003455054851522603,
|
1972 |
+
"loss": 2.5875,
|
1973 |
+
"theoretical_loss": 3.6924544281213967,
|
1974 |
+
"tokens_seen": 884736000
|
1975 |
+
},
|
1976 |
+
{
|
1977 |
+
"epoch": 0.32,
|
1978 |
+
"learning_rate": 0.0003449144127104218,
|
1979 |
+
"loss": 2.5903,
|
1980 |
+
"theoretical_loss": 3.691126473088412,
|
1981 |
+
"tokens_seen": 888012800
|
1982 |
+
},
|
1983 |
+
{
|
1984 |
+
"epoch": 0.32,
|
1985 |
+
"learning_rate": 0.00034432334026858333,
|
1986 |
+
"loss": 2.6052,
|
1987 |
+
"theoretical_loss": 3.689804775549173,
|
1988 |
+
"tokens_seen": 891289600
|
1989 |
+
},
|
1990 |
+
{
|
1991 |
+
"epoch": 0.32,
|
1992 |
+
"learning_rate": 0.00034373226782674485,
|
1993 |
+
"loss": 2.5897,
|
1994 |
+
"theoretical_loss": 3.688489283174146,
|
1995 |
+
"tokens_seen": 894566400
|
1996 |
+
},
|
1997 |
+
{
|
1998 |
+
"epoch": 0.32,
|
1999 |
+
"learning_rate": 0.00034314119538490643,
|
2000 |
+
"loss": 2.6264,
|
2001 |
+
"theoretical_loss": 3.6871799442610538,
|
2002 |
+
"tokens_seen": 897843200
|
2003 |
+
},
|
2004 |
+
{
|
2005 |
+
"epoch": 0.32,
|
2006 |
+
"learning_rate": 0.0003425501229430679,
|
2007 |
+
"loss": 2.6031,
|
2008 |
+
"theoretical_loss": 3.685876707725093,
|
2009 |
+
"tokens_seen": 901120000
|
2010 |
+
},
|
2011 |
+
{
|
2012 |
+
"epoch": 0.32,
|
2013 |
+
"learning_rate": 0.0003419590505012294,
|
2014 |
+
"loss": 2.5521,
|
2015 |
+
"theoretical_loss": 3.6845795230893517,
|
2016 |
+
"tokens_seen": 904396800
|
2017 |
+
},
|
2018 |
+
{
|
2019 |
+
"epoch": 0.32,
|
2020 |
+
"learning_rate": 0.00034136797805939094,
|
2021 |
+
"loss": 2.584,
|
2022 |
+
"theoretical_loss": 3.6832883404754035,
|
2023 |
+
"tokens_seen": 907673600
|
2024 |
+
},
|
2025 |
+
{
|
2026 |
+
"epoch": 0.33,
|
2027 |
+
"learning_rate": 0.00034077690561755246,
|
2028 |
+
"loss": 2.5797,
|
2029 |
+
"theoretical_loss": 3.6820031105940796,
|
2030 |
+
"tokens_seen": 910950400
|
2031 |
+
},
|
2032 |
+
{
|
2033 |
+
"epoch": 0.33,
|
2034 |
+
"learning_rate": 0.000340185833175714,
|
2035 |
+
"loss": 2.5943,
|
2036 |
+
"theoretical_loss": 3.6807237847364176,
|
2037 |
+
"tokens_seen": 914227200
|
2038 |
+
},
|
2039 |
+
{
|
2040 |
+
"epoch": 0.33,
|
2041 |
+
"learning_rate": 0.00033959476073387555,
|
2042 |
+
"loss": 2.6038,
|
2043 |
+
"theoretical_loss": 3.6794503147647846,
|
2044 |
+
"tokens_seen": 917504000
|
2045 |
+
},
|
2046 |
+
{
|
2047 |
+
"epoch": 0.33,
|
2048 |
+
"learning_rate": 0.0003390036882920371,
|
2049 |
+
"loss": 2.6191,
|
2050 |
+
"theoretical_loss": 3.67818265310416,
|
2051 |
+
"tokens_seen": 920780800
|
2052 |
+
},
|
2053 |
+
{
|
2054 |
+
"epoch": 0.33,
|
2055 |
+
"learning_rate": 0.0003384126158501986,
|
2056 |
+
"loss": 2.6231,
|
2057 |
+
"theoretical_loss": 3.6769207527335888,
|
2058 |
+
"tokens_seen": 924057600
|
2059 |
+
},
|
2060 |
+
{
|
2061 |
+
"epoch": 0.33,
|
2062 |
+
"learning_rate": 0.0003378215434083601,
|
2063 |
+
"loss": 2.6081,
|
2064 |
+
"theoretical_loss": 3.675664567177787,
|
2065 |
+
"tokens_seen": 927334400
|
2066 |
+
},
|
2067 |
+
{
|
2068 |
+
"epoch": 0.33,
|
2069 |
+
"learning_rate": 0.00033723047096652164,
|
2070 |
+
"loss": 2.5985,
|
2071 |
+
"theoretical_loss": 3.674414050498913,
|
2072 |
+
"tokens_seen": 930611200
|
2073 |
+
},
|
2074 |
+
{
|
2075 |
+
"epoch": 0.33,
|
2076 |
+
"learning_rate": 0.00033663939852468316,
|
2077 |
+
"loss": 2.6102,
|
2078 |
+
"theoretical_loss": 3.6731691572884824,
|
2079 |
+
"tokens_seen": 933888000
|
2080 |
+
},
|
2081 |
+
{
|
2082 |
+
"epoch": 0.33,
|
2083 |
+
"learning_rate": 0.00033604832608284473,
|
2084 |
+
"loss": 2.5746,
|
2085 |
+
"theoretical_loss": 3.671929842659438,
|
2086 |
+
"tokens_seen": 937164800
|
2087 |
+
},
|
2088 |
+
{
|
2089 |
+
"epoch": 0.34,
|
2090 |
+
"learning_rate": 0.00033545725364100625,
|
2091 |
+
"loss": 2.5936,
|
2092 |
+
"theoretical_loss": 3.6706960622383624,
|
2093 |
+
"tokens_seen": 940441600
|
2094 |
+
},
|
2095 |
+
{
|
2096 |
+
"epoch": 0.34,
|
2097 |
+
"learning_rate": 0.0003348661811991678,
|
2098 |
+
"loss": 2.5706,
|
2099 |
+
"theoretical_loss": 3.6694677721578377,
|
2100 |
+
"tokens_seen": 943718400
|
2101 |
+
},
|
2102 |
+
{
|
2103 |
+
"epoch": 0.34,
|
2104 |
+
"learning_rate": 0.0003342751087573293,
|
2105 |
+
"loss": 2.5973,
|
2106 |
+
"theoretical_loss": 3.66824492904894,
|
2107 |
+
"tokens_seen": 946995200
|
2108 |
+
},
|
2109 |
+
{
|
2110 |
+
"epoch": 0.34,
|
2111 |
+
"learning_rate": 0.0003336840363154908,
|
2112 |
+
"loss": 2.5857,
|
2113 |
+
"theoretical_loss": 3.667027490033874,
|
2114 |
+
"tokens_seen": 950272000
|
2115 |
+
},
|
2116 |
+
{
|
2117 |
+
"epoch": 0.34,
|
2118 |
+
"learning_rate": 0.00033309296387365234,
|
2119 |
+
"loss": 2.5566,
|
2120 |
+
"theoretical_loss": 3.6658154127187412,
|
2121 |
+
"tokens_seen": 953548800
|
2122 |
+
},
|
2123 |
+
{
|
2124 |
+
"epoch": 0.34,
|
2125 |
+
"learning_rate": 0.00033250189143181386,
|
2126 |
+
"loss": 2.5445,
|
2127 |
+
"theoretical_loss": 3.664608655186437,
|
2128 |
+
"tokens_seen": 956825600
|
2129 |
+
},
|
2130 |
+
{
|
2131 |
+
"epoch": 0.34,
|
2132 |
+
"learning_rate": 0.00033191081898997543,
|
2133 |
+
"loss": 2.6071,
|
2134 |
+
"theoretical_loss": 3.663407175989679,
|
2135 |
+
"tokens_seen": 960102400
|
2136 |
+
},
|
2137 |
+
{
|
2138 |
+
"epoch": 0.34,
|
2139 |
+
"learning_rate": 0.00033131974654813695,
|
2140 |
+
"loss": 2.5979,
|
2141 |
+
"theoretical_loss": 3.662210934144158,
|
2142 |
+
"tokens_seen": 963379200
|
2143 |
+
},
|
2144 |
+
{
|
2145 |
+
"epoch": 0.35,
|
2146 |
+
"learning_rate": 0.0003307286741062985,
|
2147 |
+
"loss": 2.5734,
|
2148 |
+
"theoretical_loss": 3.661019889121812,
|
2149 |
+
"tokens_seen": 966656000
|
2150 |
+
},
|
2151 |
+
{
|
2152 |
+
"epoch": 0.35,
|
2153 |
+
"learning_rate": 0.00033013760166446,
|
2154 |
+
"loss": 2.5867,
|
2155 |
+
"theoretical_loss": 3.6598340008442234,
|
2156 |
+
"tokens_seen": 969932800
|
2157 |
+
},
|
2158 |
+
{
|
2159 |
+
"epoch": 0.35,
|
2160 |
+
"learning_rate": 0.0003295465292226215,
|
2161 |
+
"loss": 2.6167,
|
2162 |
+
"theoretical_loss": 3.6586532296761285,
|
2163 |
+
"tokens_seen": 973209600
|
2164 |
+
},
|
2165 |
+
{
|
2166 |
+
"epoch": 0.35,
|
2167 |
+
"learning_rate": 0.00032895545678078304,
|
2168 |
+
"loss": 2.621,
|
2169 |
+
"theoretical_loss": 3.657477536419047,
|
2170 |
+
"tokens_seen": 976486400
|
2171 |
+
},
|
2172 |
+
{
|
2173 |
+
"epoch": 0.35,
|
2174 |
+
"learning_rate": 0.0003283643843389446,
|
2175 |
+
"loss": 2.631,
|
2176 |
+
"theoretical_loss": 3.656306882305022,
|
2177 |
+
"tokens_seen": 979763200
|
2178 |
+
},
|
2179 |
+
{
|
2180 |
+
"debugging/Self-BLEU-5": 0.5857137347143212,
|
2181 |
+
"debugging/distinct-1-grams": 0.8015824539835295,
|
2182 |
+
"debugging/distinct-2-grams": 0.9770068773134507,
|
2183 |
+
"debugging/entropy-1-grams": 6.057437437902186,
|
2184 |
+
"debugging/entropy-2-grams": 6.967727189100934,
|
2185 |
+
"debugging/length": 546.0,
|
2186 |
+
"debugging/num_segments": 18,
|
2187 |
+
"epoch": 0.35,
|
2188 |
+
"objective/train/docs_used": 521967,
|
2189 |
+
"objective/train/instantaneous_batch_size": 32,
|
2190 |
+
"objective/train/instantaneous_microbatch_size": 32768,
|
2191 |
+
"objective/train/original_loss": 2.800149440765381,
|
2192 |
+
"objective/train/theoretical_loss": 3.6551412289904697,
|
2193 |
+
"objective/train/tokens_used": 1003500000,
|
2194 |
+
"theoretical_loss": 3.6551412289904697,
|
2195 |
+
"tokens_seen": 983040000
|
2196 |
+
},
|
2197 |
+
{
|
2198 |
+
"epoch": 0.35,
|
2199 |
+
"learning_rate": 0.00032777331189710613,
|
2200 |
+
"loss": 2.5862,
|
2201 |
+
"theoretical_loss": 3.6551412289904697,
|
2202 |
+
"tokens_seen": 983040000
|
2203 |
+
},
|
2204 |
+
{
|
2205 |
+
"epoch": 0.35,
|
2206 |
+
"learning_rate": 0.00032718223945526765,
|
2207 |
+
"loss": 2.5939,
|
2208 |
+
"theoretical_loss": 3.6539805385501376,
|
2209 |
+
"tokens_seen": 986316800
|
2210 |
+
},
|
2211 |
+
{
|
2212 |
+
"epoch": 0.35,
|
2213 |
+
"learning_rate": 0.0003265911670134292,
|
2214 |
+
"loss": 2.5941,
|
2215 |
+
"theoretical_loss": 3.652824773471171,
|
2216 |
+
"tokens_seen": 989593600
|
2217 |
+
},
|
2218 |
+
{
|
2219 |
+
"epoch": 0.35,
|
2220 |
+
"learning_rate": 0.0003260000945715907,
|
2221 |
+
"loss": 2.6028,
|
2222 |
+
"theoretical_loss": 3.651673896647277,
|
2223 |
+
"tokens_seen": 992870400
|
2224 |
+
},
|
2225 |
+
{
|
2226 |
+
"epoch": 0.36,
|
2227 |
+
"learning_rate": 0.0003254090221297522,
|
2228 |
+
"loss": 2.5966,
|
2229 |
+
"theoretical_loss": 3.6505278713729985,
|
2230 |
+
"tokens_seen": 996147200
|
2231 |
+
},
|
2232 |
+
{
|
2233 |
+
"epoch": 0.36,
|
2234 |
+
"learning_rate": 0.0003248179496879138,
|
2235 |
+
"loss": 2.5693,
|
2236 |
+
"theoretical_loss": 3.6493866613380774,
|
2237 |
+
"tokens_seen": 999424000
|
2238 |
+
},
|
2239 |
+
{
|
2240 |
+
"epoch": 0.36,
|
2241 |
+
"learning_rate": 0.00032423869869491206,
|
2242 |
+
"loss": 2.5594,
|
2243 |
+
"theoretical_loss": 3.648250230621924,
|
2244 |
+
"tokens_seen": 1002700800
|
2245 |
+
},
|
2246 |
+
{
|
2247 |
+
"epoch": 0.36,
|
2248 |
+
"learning_rate": 0.0003236476262530736,
|
2249 |
+
"loss": 2.5436,
|
2250 |
+
"theoretical_loss": 3.647118543688179,
|
2251 |
+
"tokens_seen": 1005977600
|
2252 |
+
},
|
2253 |
+
{
|
2254 |
+
"epoch": 0.36,
|
2255 |
+
"learning_rate": 0.0003230565538112351,
|
2256 |
+
"loss": 2.5506,
|
2257 |
+
"theoretical_loss": 3.6459915653793633,
|
2258 |
+
"tokens_seen": 1009254400
|
2259 |
+
},
|
2260 |
+
{
|
2261 |
+
"epoch": 0.36,
|
2262 |
+
"learning_rate": 0.0003224654813693966,
|
2263 |
+
"loss": 2.5771,
|
2264 |
+
"theoretical_loss": 3.644869260911628,
|
2265 |
+
"tokens_seen": 1012531200
|
2266 |
+
},
|
2267 |
+
{
|
2268 |
+
"epoch": 0.36,
|
2269 |
+
"learning_rate": 0.00032187440892755814,
|
2270 |
+
"loss": 2.5578,
|
2271 |
+
"theoretical_loss": 3.64375159586959,
|
2272 |
+
"tokens_seen": 1015808000
|
2273 |
+
},
|
2274 |
+
{
|
2275 |
+
"epoch": 0.36,
|
2276 |
+
"learning_rate": 0.0003212833364857197,
|
2277 |
+
"loss": 2.5466,
|
2278 |
+
"theoretical_loss": 3.642638536201252,
|
2279 |
+
"tokens_seen": 1019084800
|
2280 |
+
},
|
2281 |
+
{
|
2282 |
+
"epoch": 0.37,
|
2283 |
+
"learning_rate": 0.00032069226404388124,
|
2284 |
+
"loss": 2.561,
|
2285 |
+
"theoretical_loss": 3.6415300482130135,
|
2286 |
+
"tokens_seen": 1022361600
|
2287 |
+
},
|
2288 |
+
{
|
2289 |
+
"epoch": 0.37,
|
2290 |
+
"learning_rate": 0.00032010119160204276,
|
2291 |
+
"loss": 2.5863,
|
2292 |
+
"theoretical_loss": 3.6404260985647667,
|
2293 |
+
"tokens_seen": 1025638400
|
2294 |
+
},
|
2295 |
+
{
|
2296 |
+
"epoch": 0.37,
|
2297 |
+
"learning_rate": 0.0003195101191602043,
|
2298 |
+
"loss": 2.5644,
|
2299 |
+
"theoretical_loss": 3.6393266542650684,
|
2300 |
+
"tokens_seen": 1028915200
|
2301 |
+
},
|
2302 |
+
{
|
2303 |
+
"epoch": 0.37,
|
2304 |
+
"learning_rate": 0.0003189190467183658,
|
2305 |
+
"loss": 2.5471,
|
2306 |
+
"theoretical_loss": 3.638231682666401,
|
2307 |
+
"tokens_seen": 1032192000
|
2308 |
+
},
|
2309 |
+
{
|
2310 |
+
"epoch": 0.37,
|
2311 |
+
"learning_rate": 0.0003183279742765273,
|
2312 |
+
"loss": 2.553,
|
2313 |
+
"theoretical_loss": 3.637141151460505,
|
2314 |
+
"tokens_seen": 1035468800
|
2315 |
+
},
|
2316 |
+
{
|
2317 |
+
"epoch": 0.37,
|
2318 |
+
"learning_rate": 0.00031773690183468884,
|
2319 |
+
"loss": 2.5607,
|
2320 |
+
"theoretical_loss": 3.636055028673799,
|
2321 |
+
"tokens_seen": 1038745600
|
2322 |
+
},
|
2323 |
+
{
|
2324 |
+
"epoch": 0.37,
|
2325 |
+
"learning_rate": 0.0003171458293928504,
|
2326 |
+
"loss": 2.5638,
|
2327 |
+
"theoretical_loss": 3.634973282662864,
|
2328 |
+
"tokens_seen": 1042022400
|
2329 |
+
},
|
2330 |
+
{
|
2331 |
+
"epoch": 0.37,
|
2332 |
+
"learning_rate": 0.00031655475695101194,
|
2333 |
+
"loss": 2.6156,
|
2334 |
+
"theoretical_loss": 3.6338958821100107,
|
2335 |
+
"tokens_seen": 1045299200
|
2336 |
+
},
|
2337 |
+
{
|
2338 |
+
"epoch": 0.37,
|
2339 |
+
"learning_rate": 0.00031596368450917346,
|
2340 |
+
"loss": 2.6571,
|
2341 |
+
"theoretical_loss": 3.63282279601892,
|
2342 |
+
"tokens_seen": 1048576000
|
2343 |
+
},
|
2344 |
+
{
|
2345 |
+
"epoch": 0.38,
|
2346 |
+
"learning_rate": 0.000315372612067335,
|
2347 |
+
"loss": 2.6717,
|
2348 |
+
"theoretical_loss": 3.631753993710352,
|
2349 |
+
"tokens_seen": 1051852800
|
2350 |
+
},
|
2351 |
+
{
|
2352 |
+
"epoch": 0.38,
|
2353 |
+
"learning_rate": 0.0003147815396254965,
|
2354 |
+
"loss": 2.6583,
|
2355 |
+
"theoretical_loss": 3.630689444817925,
|
2356 |
+
"tokens_seen": 1055129600
|
2357 |
+
},
|
2358 |
+
{
|
2359 |
+
"epoch": 0.38,
|
2360 |
+
"learning_rate": 0.000314190467183658,
|
2361 |
+
"loss": 2.6705,
|
2362 |
+
"theoretical_loss": 3.629629119283967,
|
2363 |
+
"tokens_seen": 1058406400
|
2364 |
+
},
|
2365 |
+
{
|
2366 |
+
"epoch": 0.38,
|
2367 |
+
"learning_rate": 0.0003135993947418196,
|
2368 |
+
"loss": 2.6554,
|
2369 |
+
"theoretical_loss": 3.628572987355434,
|
2370 |
+
"tokens_seen": 1061683200
|
2371 |
+
},
|
2372 |
+
{
|
2373 |
+
"epoch": 0.38,
|
2374 |
+
"learning_rate": 0.0003130083222999811,
|
2375 |
+
"loss": 2.6439,
|
2376 |
+
"theoretical_loss": 3.6275210195798913,
|
2377 |
+
"tokens_seen": 1064960000
|
2378 |
+
},
|
2379 |
+
{
|
2380 |
+
"epoch": 0.38,
|
2381 |
+
"learning_rate": 0.00031241724985814264,
|
2382 |
+
"loss": 2.6397,
|
2383 |
+
"theoretical_loss": 3.626473186801564,
|
2384 |
+
"tokens_seen": 1068236800
|
2385 |
+
},
|
2386 |
+
{
|
2387 |
+
"epoch": 0.38,
|
2388 |
+
"learning_rate": 0.00031182617741630416,
|
2389 |
+
"loss": 2.6653,
|
2390 |
+
"theoretical_loss": 3.6254294601574495,
|
2391 |
+
"tokens_seen": 1071513600
|
2392 |
+
},
|
2393 |
+
{
|
2394 |
+
"epoch": 0.38,
|
2395 |
+
"learning_rate": 0.0003112351049744657,
|
2396 |
+
"loss": 2.61,
|
2397 |
+
"theoretical_loss": 3.624389811073493,
|
2398 |
+
"tokens_seen": 1074790400
|
2399 |
+
},
|
2400 |
+
{
|
2401 |
+
"epoch": 0.39,
|
2402 |
+
"learning_rate": 0.0003106440325326272,
|
2403 |
+
"loss": 2.6357,
|
2404 |
+
"theoretical_loss": 3.6233542112608257,
|
2405 |
+
"tokens_seen": 1078067200
|
2406 |
+
},
|
2407 |
+
{
|
2408 |
+
"epoch": 0.39,
|
2409 |
+
"learning_rate": 0.0003100529600907888,
|
2410 |
+
"loss": 2.6169,
|
2411 |
+
"theoretical_loss": 3.6223226327120592,
|
2412 |
+
"tokens_seen": 1081344000
|
2413 |
+
},
|
2414 |
+
{
|
2415 |
+
"epoch": 0.39,
|
2416 |
+
"learning_rate": 0.0003094618876489503,
|
2417 |
+
"loss": 2.6232,
|
2418 |
+
"theoretical_loss": 3.621295047697644,
|
2419 |
+
"tokens_seen": 1084620800
|
2420 |
+
},
|
2421 |
+
{
|
2422 |
+
"epoch": 0.39,
|
2423 |
+
"learning_rate": 0.0003088708152071118,
|
2424 |
+
"loss": 2.6621,
|
2425 |
+
"theoretical_loss": 3.6202714287622833,
|
2426 |
+
"tokens_seen": 1087897600
|
2427 |
+
},
|
2428 |
+
{
|
2429 |
+
"epoch": 0.39,
|
2430 |
+
"learning_rate": 0.00030827974276527334,
|
2431 |
+
"loss": 2.6171,
|
2432 |
+
"theoretical_loss": 3.6192517487214038,
|
2433 |
+
"tokens_seen": 1091174400
|
2434 |
+
},
|
2435 |
+
{
|
2436 |
+
"epoch": 0.39,
|
2437 |
+
"learning_rate": 0.00030768867032343486,
|
2438 |
+
"loss": 2.5856,
|
2439 |
+
"theoretical_loss": 3.6182359806576834,
|
2440 |
+
"tokens_seen": 1094451200
|
2441 |
+
},
|
2442 |
+
{
|
2443 |
+
"epoch": 0.39,
|
2444 |
+
"learning_rate": 0.0003070975978815964,
|
2445 |
+
"loss": 2.6144,
|
2446 |
+
"theoretical_loss": 3.6172240979176333,
|
2447 |
+
"tokens_seen": 1097728000
|
2448 |
+
},
|
2449 |
+
{
|
2450 |
+
"epoch": 0.39,
|
2451 |
+
"learning_rate": 0.0003065065254397579,
|
2452 |
+
"loss": 2.5866,
|
2453 |
+
"theoretical_loss": 3.616216074108232,
|
2454 |
+
"tokens_seen": 1101004800
|
2455 |
+
},
|
2456 |
+
{
|
2457 |
+
"epoch": 0.39,
|
2458 |
+
"learning_rate": 0.0003059154529979195,
|
2459 |
+
"loss": 2.6417,
|
2460 |
+
"theoretical_loss": 3.6152118830936164,
|
2461 |
+
"tokens_seen": 1104281600
|
2462 |
+
},
|
2463 |
+
{
|
2464 |
+
"epoch": 0.4,
|
2465 |
+
"learning_rate": 0.000305324380556081,
|
2466 |
+
"loss": 2.6339,
|
2467 |
+
"theoretical_loss": 3.6142114989918195,
|
2468 |
+
"tokens_seen": 1107558400
|
2469 |
+
},
|
2470 |
+
{
|
2471 |
+
"epoch": 0.4,
|
2472 |
+
"learning_rate": 0.00030473330811424246,
|
2473 |
+
"loss": 2.626,
|
2474 |
+
"theoretical_loss": 3.6132148961715624,
|
2475 |
+
"tokens_seen": 1110835200
|
2476 |
+
},
|
2477 |
+
{
|
2478 |
+
"epoch": 0.4,
|
2479 |
+
"learning_rate": 0.000304142235672404,
|
2480 |
+
"loss": 2.6102,
|
2481 |
+
"theoretical_loss": 3.6122220492490964,
|
2482 |
+
"tokens_seen": 1114112000
|
2483 |
+
},
|
2484 |
+
{
|
2485 |
+
"epoch": 0.4,
|
2486 |
+
"learning_rate": 0.0003035511632305655,
|
2487 |
+
"loss": 2.5967,
|
2488 |
+
"theoretical_loss": 3.6112329330850894,
|
2489 |
+
"tokens_seen": 1117388800
|
2490 |
+
},
|
2491 |
+
{
|
2492 |
+
"epoch": 0.4,
|
2493 |
+
"learning_rate": 0.000302960090788727,
|
2494 |
+
"loss": 2.597,
|
2495 |
+
"theoretical_loss": 3.61024752278157,
|
2496 |
+
"tokens_seen": 1120665600
|
2497 |
+
},
|
2498 |
+
{
|
2499 |
+
"epoch": 0.4,
|
2500 |
+
"learning_rate": 0.0003023690183468886,
|
2501 |
+
"loss": 2.6077,
|
2502 |
+
"theoretical_loss": 3.6092657936789054,
|
2503 |
+
"tokens_seen": 1123942400
|
2504 |
+
},
|
2505 |
+
{
|
2506 |
+
"epoch": 0.4,
|
2507 |
+
"learning_rate": 0.0003017779459050501,
|
2508 |
+
"loss": 2.5608,
|
2509 |
+
"theoretical_loss": 3.6082877213528377,
|
2510 |
+
"tokens_seen": 1127219200
|
2511 |
+
},
|
2512 |
+
{
|
2513 |
+
"epoch": 0.4,
|
2514 |
+
"learning_rate": 0.00030118687346321164,
|
2515 |
+
"loss": 2.5657,
|
2516 |
+
"theoretical_loss": 3.60731328161156,
|
2517 |
+
"tokens_seen": 1130496000
|
2518 |
+
},
|
2519 |
+
{
|
2520 |
+
"epoch": 0.4,
|
2521 |
+
"learning_rate": 0.00030059580102137316,
|
2522 |
+
"loss": 2.6017,
|
2523 |
+
"theoretical_loss": 3.6063424504928365,
|
2524 |
+
"tokens_seen": 1133772800
|
2525 |
+
},
|
2526 |
+
{
|
2527 |
+
"epoch": 0.41,
|
2528 |
+
"learning_rate": 0.0003000047285795347,
|
2529 |
+
"loss": 2.5677,
|
2530 |
+
"theoretical_loss": 3.60537520426117,
|
2531 |
+
"tokens_seen": 1137049600
|
2532 |
+
},
|
2533 |
+
{
|
2534 |
+
"epoch": 0.41,
|
2535 |
+
"learning_rate": 0.0002994136561376962,
|
2536 |
+
"loss": 2.5892,
|
2537 |
+
"theoretical_loss": 3.6044115194050086,
|
2538 |
+
"tokens_seen": 1140326400
|
2539 |
+
},
|
2540 |
+
{
|
2541 |
+
"epoch": 0.41,
|
2542 |
+
"learning_rate": 0.0002988225836958578,
|
2543 |
+
"loss": 2.5556,
|
2544 |
+
"theoretical_loss": 3.603451372633997,
|
2545 |
+
"tokens_seen": 1143603200
|
2546 |
+
},
|
2547 |
+
{
|
2548 |
+
"epoch": 0.41,
|
2549 |
+
"objective/train/docs_used": 604425,
|
2550 |
+
"objective/train/instantaneous_batch_size": 32,
|
2551 |
+
"objective/train/instantaneous_microbatch_size": 32768,
|
2552 |
+
"objective/train/original_loss": 2.6260290145874023,
|
2553 |
+
"objective/train/theoretical_loss": 3.6024947408762698,
|
2554 |
+
"objective/train/tokens_used": 1167340000,
|
2555 |
+
"theoretical_loss": 3.6024947408762698,
|
2556 |
+
"tokens_seen": 1146880000
|
2557 |
+
},
|
2558 |
+
{
|
2559 |
+
"epoch": 0.41,
|
2560 |
+
"learning_rate": 0.0002982433327028561,
|
2561 |
+
"loss": 2.6048,
|
2562 |
+
"theoretical_loss": 3.6024947408762698,
|
2563 |
+
"tokens_seen": 1146880000
|
2564 |
+
},
|
2565 |
+
{
|
2566 |
+
"epoch": 0.41,
|
2567 |
+
"learning_rate": 0.0002976640817098544,
|
2568 |
+
"loss": 2.6085,
|
2569 |
+
"theoretical_loss": 3.601541601275783,
|
2570 |
+
"tokens_seen": 1150156800
|
2571 |
+
},
|
2572 |
+
{
|
2573 |
+
"epoch": 0.41,
|
2574 |
+
"learning_rate": 0.0002970730092680159,
|
2575 |
+
"loss": 2.5966,
|
2576 |
+
"theoretical_loss": 3.6005919311896886,
|
2577 |
+
"tokens_seen": 1153433600
|
2578 |
+
},
|
2579 |
+
{
|
2580 |
+
"epoch": 0.41,
|
2581 |
+
"learning_rate": 0.0002964819368261774,
|
2582 |
+
"loss": 2.5683,
|
2583 |
+
"theoretical_loss": 3.5996457081857454,
|
2584 |
+
"tokens_seen": 1156710400
|
2585 |
+
},
|
2586 |
+
{
|
2587 |
+
"epoch": 0.41,
|
2588 |
+
"learning_rate": 0.00029589086438433894,
|
2589 |
+
"loss": 2.5609,
|
2590 |
+
"theoretical_loss": 3.598702910039772,
|
2591 |
+
"tokens_seen": 1159987200
|
2592 |
+
},
|
2593 |
+
{
|
2594 |
+
"epoch": 0.42,
|
2595 |
+
"learning_rate": 0.0002952997919425005,
|
2596 |
+
"loss": 2.5521,
|
2597 |
+
"theoretical_loss": 3.597763514733133,
|
2598 |
+
"tokens_seen": 1163264000
|
2599 |
+
},
|
2600 |
+
{
|
2601 |
+
"epoch": 0.42,
|
2602 |
+
"learning_rate": 0.00029470871950066204,
|
2603 |
+
"loss": 2.5443,
|
2604 |
+
"theoretical_loss": 3.59682750045027,
|
2605 |
+
"tokens_seen": 1166540800
|
2606 |
+
},
|
2607 |
+
{
|
2608 |
+
"epoch": 0.42,
|
2609 |
+
"learning_rate": 0.00029411764705882356,
|
2610 |
+
"loss": 2.5605,
|
2611 |
+
"theoretical_loss": 3.5958948455762583,
|
2612 |
+
"tokens_seen": 1169817600
|
2613 |
+
},
|
2614 |
+
{
|
2615 |
+
"epoch": 0.42,
|
2616 |
+
"learning_rate": 0.0002935265746169851,
|
2617 |
+
"loss": 2.6041,
|
2618 |
+
"theoretical_loss": 3.594965528694412,
|
2619 |
+
"tokens_seen": 1173094400
|
2620 |
+
},
|
2621 |
+
{
|
2622 |
+
"epoch": 0.42,
|
2623 |
+
"learning_rate": 0.0002929355021751466,
|
2624 |
+
"loss": 2.574,
|
2625 |
+
"theoretical_loss": 3.594039528583913,
|
2626 |
+
"tokens_seen": 1176371200
|
2627 |
+
},
|
2628 |
+
{
|
2629 |
+
"epoch": 0.42,
|
2630 |
+
"learning_rate": 0.0002923444297333081,
|
2631 |
+
"loss": 2.5796,
|
2632 |
+
"theoretical_loss": 3.5931168242174847,
|
2633 |
+
"tokens_seen": 1179648000
|
2634 |
+
},
|
2635 |
+
{
|
2636 |
+
"epoch": 0.42,
|
2637 |
+
"learning_rate": 0.00029175335729146964,
|
2638 |
+
"loss": 2.5703,
|
2639 |
+
"theoretical_loss": 3.59219739475909,
|
2640 |
+
"tokens_seen": 1182924800
|
2641 |
+
},
|
2642 |
+
{
|
2643 |
+
"epoch": 0.42,
|
2644 |
+
"learning_rate": 0.0002911622848496312,
|
2645 |
+
"loss": 2.5559,
|
2646 |
+
"theoretical_loss": 3.5912812195616732,
|
2647 |
+
"tokens_seen": 1186201600
|
2648 |
+
},
|
2649 |
+
{
|
2650 |
+
"epoch": 0.42,
|
2651 |
+
"learning_rate": 0.00029057121240779274,
|
2652 |
+
"loss": 2.5536,
|
2653 |
+
"theoretical_loss": 3.590368278164926,
|
2654 |
+
"tokens_seen": 1189478400
|
2655 |
+
},
|
2656 |
+
{
|
2657 |
+
"epoch": 0.43,
|
2658 |
+
"learning_rate": 0.00028998013996595426,
|
2659 |
+
"loss": 2.5564,
|
2660 |
+
"theoretical_loss": 3.5894585502930902,
|
2661 |
+
"tokens_seen": 1192755200
|
2662 |
+
},
|
2663 |
+
{
|
2664 |
+
"epoch": 0.43,
|
2665 |
+
"learning_rate": 0.0002893890675241158,
|
2666 |
+
"loss": 2.6011,
|
2667 |
+
"theoretical_loss": 3.588552015852793,
|
2668 |
+
"tokens_seen": 1196032000
|
2669 |
+
},
|
2670 |
+
{
|
2671 |
+
"epoch": 0.43,
|
2672 |
+
"learning_rate": 0.0002887979950822773,
|
2673 |
+
"loss": 2.5488,
|
2674 |
+
"theoretical_loss": 3.5876486549309097,
|
2675 |
+
"tokens_seen": 1199308800
|
2676 |
+
},
|
2677 |
+
{
|
2678 |
+
"epoch": 0.43,
|
2679 |
+
"learning_rate": 0.0002882069226404388,
|
2680 |
+
"loss": 2.5447,
|
2681 |
+
"theoretical_loss": 3.586748447792462,
|
2682 |
+
"tokens_seen": 1202585600
|
2683 |
+
},
|
2684 |
+
{
|
2685 |
+
"epoch": 0.43,
|
2686 |
+
"learning_rate": 0.0002876158501986004,
|
2687 |
+
"loss": 2.5656,
|
2688 |
+
"theoretical_loss": 3.5858513748785423,
|
2689 |
+
"tokens_seen": 1205862400
|
2690 |
+
},
|
2691 |
+
{
|
2692 |
+
"epoch": 0.43,
|
2693 |
+
"learning_rate": 0.0002870247777567619,
|
2694 |
+
"loss": 2.5935,
|
2695 |
+
"theoretical_loss": 3.5849574168042704,
|
2696 |
+
"tokens_seen": 1209139200
|
2697 |
+
},
|
2698 |
+
{
|
2699 |
+
"epoch": 0.43,
|
2700 |
+
"learning_rate": 0.00028643370531492344,
|
2701 |
+
"loss": 2.6245,
|
2702 |
+
"theoretical_loss": 3.5840665543567782,
|
2703 |
+
"tokens_seen": 1212416000
|
2704 |
+
},
|
2705 |
+
{
|
2706 |
+
"epoch": 0.43,
|
2707 |
+
"learning_rate": 0.0002858426328730849,
|
2708 |
+
"loss": 2.5978,
|
2709 |
+
"theoretical_loss": 3.583178768493222,
|
2710 |
+
"tokens_seen": 1215692800
|
2711 |
+
},
|
2712 |
+
{
|
2713 |
+
"epoch": 0.44,
|
2714 |
+
"learning_rate": 0.0002852515604312464,
|
2715 |
+
"loss": 2.5588,
|
2716 |
+
"theoretical_loss": 3.5822940403388284,
|
2717 |
+
"tokens_seen": 1218969600
|
2718 |
+
},
|
2719 |
+
{
|
2720 |
+
"epoch": 0.44,
|
2721 |
+
"learning_rate": 0.00028466048798940794,
|
2722 |
+
"loss": 2.5947,
|
2723 |
+
"theoretical_loss": 3.581412351184958,
|
2724 |
+
"tokens_seen": 1222246400
|
2725 |
+
},
|
2726 |
+
{
|
2727 |
+
"epoch": 0.44,
|
2728 |
+
"learning_rate": 0.0002840694155475695,
|
2729 |
+
"loss": 2.5713,
|
2730 |
+
"theoretical_loss": 3.580533682487208,
|
2731 |
+
"tokens_seen": 1225523200
|
2732 |
+
},
|
2733 |
+
{
|
2734 |
+
"epoch": 0.44,
|
2735 |
+
"learning_rate": 0.00028347834310573104,
|
2736 |
+
"loss": 2.5671,
|
2737 |
+
"theoretical_loss": 3.579658015863532,
|
2738 |
+
"tokens_seen": 1228800000
|
2739 |
+
},
|
2740 |
+
{
|
2741 |
+
"epoch": 0.44,
|
2742 |
+
"learning_rate": 0.00028288727066389256,
|
2743 |
+
"loss": 2.5841,
|
2744 |
+
"theoretical_loss": 3.5787853330923927,
|
2745 |
+
"tokens_seen": 1232076800
|
2746 |
+
},
|
2747 |
+
{
|
2748 |
+
"epoch": 0.44,
|
2749 |
+
"learning_rate": 0.0002822961982220541,
|
2750 |
+
"loss": 2.5546,
|
2751 |
+
"theoretical_loss": 3.577915616110936,
|
2752 |
+
"tokens_seen": 1235353600
|
2753 |
+
},
|
2754 |
+
{
|
2755 |
+
"epoch": 0.44,
|
2756 |
+
"learning_rate": 0.0002817051257802156,
|
2757 |
+
"loss": 2.5536,
|
2758 |
+
"theoretical_loss": 3.577048847013194,
|
2759 |
+
"tokens_seen": 1238630400
|
2760 |
+
},
|
2761 |
+
{
|
2762 |
+
"epoch": 0.44,
|
2763 |
+
"learning_rate": 0.0002811140533383771,
|
2764 |
+
"loss": 2.5912,
|
2765 |
+
"theoretical_loss": 3.57618500804831,
|
2766 |
+
"tokens_seen": 1241907200
|
2767 |
+
},
|
2768 |
+
{
|
2769 |
+
"epoch": 0.44,
|
2770 |
+
"learning_rate": 0.00028052298089653864,
|
2771 |
+
"loss": 2.5372,
|
2772 |
+
"theoretical_loss": 3.575324081618793,
|
2773 |
+
"tokens_seen": 1245184000
|
2774 |
+
},
|
2775 |
+
{
|
2776 |
+
"epoch": 0.45,
|
2777 |
+
"learning_rate": 0.0002799319084547002,
|
2778 |
+
"loss": 2.5759,
|
2779 |
+
"theoretical_loss": 3.5744660502787875,
|
2780 |
+
"tokens_seen": 1248460800
|
2781 |
+
},
|
2782 |
+
{
|
2783 |
+
"epoch": 0.45,
|
2784 |
+
"learning_rate": 0.00027934083601286174,
|
2785 |
+
"loss": 2.5583,
|
2786 |
+
"theoretical_loss": 3.5736108967323794,
|
2787 |
+
"tokens_seen": 1251737600
|
2788 |
+
},
|
2789 |
+
{
|
2790 |
+
"epoch": 0.45,
|
2791 |
+
"learning_rate": 0.00027874976357102326,
|
2792 |
+
"loss": 2.5872,
|
2793 |
+
"theoretical_loss": 3.5727586038319155,
|
2794 |
+
"tokens_seen": 1255014400
|
2795 |
+
},
|
2796 |
+
{
|
2797 |
+
"epoch": 0.45,
|
2798 |
+
"learning_rate": 0.0002781586911291848,
|
2799 |
+
"loss": 2.5319,
|
2800 |
+
"theoretical_loss": 3.571909154576348,
|
2801 |
+
"tokens_seen": 1258291200
|
2802 |
+
},
|
2803 |
+
{
|
2804 |
+
"epoch": 0.45,
|
2805 |
+
"learning_rate": 0.0002775676186873463,
|
2806 |
+
"loss": 2.5563,
|
2807 |
+
"theoretical_loss": 3.5710625321096074,
|
2808 |
+
"tokens_seen": 1261568000
|
2809 |
+
},
|
2810 |
+
{
|
2811 |
+
"epoch": 0.45,
|
2812 |
+
"learning_rate": 0.0002769765462455078,
|
2813 |
+
"loss": 2.553,
|
2814 |
+
"theoretical_loss": 3.570218719718989,
|
2815 |
+
"tokens_seen": 1264844800
|
2816 |
+
},
|
2817 |
+
{
|
2818 |
+
"epoch": 0.45,
|
2819 |
+
"learning_rate": 0.0002763854738036694,
|
2820 |
+
"loss": 2.5804,
|
2821 |
+
"theoretical_loss": 3.569377700833569,
|
2822 |
+
"tokens_seen": 1268121600
|
2823 |
+
},
|
2824 |
+
{
|
2825 |
+
"epoch": 0.45,
|
2826 |
+
"learning_rate": 0.0002757944013618309,
|
2827 |
+
"loss": 2.5463,
|
2828 |
+
"theoretical_loss": 3.568539459022639,
|
2829 |
+
"tokens_seen": 1271398400
|
2830 |
+
},
|
2831 |
+
{
|
2832 |
+
"epoch": 0.46,
|
2833 |
+
"learning_rate": 0.00027520332891999244,
|
2834 |
+
"loss": 2.5542,
|
2835 |
+
"theoretical_loss": 3.5677039779941584,
|
2836 |
+
"tokens_seen": 1274675200
|
2837 |
+
},
|
2838 |
+
{
|
2839 |
+
"epoch": 0.46,
|
2840 |
+
"learning_rate": 0.00027461225647815396,
|
2841 |
+
"loss": 2.5695,
|
2842 |
+
"theoretical_loss": 3.566871241593236,
|
2843 |
+
"tokens_seen": 1277952000
|
2844 |
+
},
|
2845 |
+
{
|
2846 |
+
"epoch": 0.46,
|
2847 |
+
"learning_rate": 0.0002740211840363155,
|
2848 |
+
"loss": 2.5521,
|
2849 |
+
"theoretical_loss": 3.5660412338006235,
|
2850 |
+
"tokens_seen": 1281228800
|
2851 |
+
},
|
2852 |
+
{
|
2853 |
+
"epoch": 0.46,
|
2854 |
+
"learning_rate": 0.000273430111594477,
|
2855 |
+
"loss": 2.5862,
|
2856 |
+
"theoretical_loss": 3.565213938731236,
|
2857 |
+
"tokens_seen": 1284505600
|
2858 |
+
},
|
2859 |
+
{
|
2860 |
+
"epoch": 0.46,
|
2861 |
+
"learning_rate": 0.0002728390391526386,
|
2862 |
+
"loss": 2.5867,
|
2863 |
+
"theoretical_loss": 3.5643893406326868,
|
2864 |
+
"tokens_seen": 1287782400
|
2865 |
+
},
|
2866 |
+
{
|
2867 |
+
"epoch": 0.46,
|
2868 |
+
"learning_rate": 0.0002722479667108001,
|
2869 |
+
"loss": 2.5822,
|
2870 |
+
"theoretical_loss": 3.5635674238838466,
|
2871 |
+
"tokens_seen": 1291059200
|
2872 |
+
},
|
2873 |
+
{
|
2874 |
+
"epoch": 0.46,
|
2875 |
+
"learning_rate": 0.0002716568942689616,
|
2876 |
+
"loss": 2.5458,
|
2877 |
+
"theoretical_loss": 3.5627481729934196,
|
2878 |
+
"tokens_seen": 1294336000
|
2879 |
+
},
|
2880 |
+
{
|
2881 |
+
"epoch": 0.46,
|
2882 |
+
"learning_rate": 0.00027106582182712314,
|
2883 |
+
"loss": 2.5858,
|
2884 |
+
"theoretical_loss": 3.561931572598538,
|
2885 |
+
"tokens_seen": 1297612800
|
2886 |
+
},
|
2887 |
+
{
|
2888 |
+
"epoch": 0.46,
|
2889 |
+
"learning_rate": 0.00027047474938528466,
|
2890 |
+
"loss": 2.5536,
|
2891 |
+
"theoretical_loss": 3.5611176074633777,
|
2892 |
+
"tokens_seen": 1300889600
|
2893 |
+
},
|
2894 |
+
{
|
2895 |
+
"epoch": 0.47,
|
2896 |
+
"learning_rate": 0.0002698836769434462,
|
2897 |
+
"loss": 2.554,
|
2898 |
+
"theoretical_loss": 3.5603062624777895,
|
2899 |
+
"tokens_seen": 1304166400
|
2900 |
+
},
|
2901 |
+
{
|
2902 |
+
"epoch": 0.47,
|
2903 |
+
"learning_rate": 0.0002692926045016077,
|
2904 |
+
"loss": 2.6034,
|
2905 |
+
"theoretical_loss": 3.559497522655951,
|
2906 |
+
"tokens_seen": 1307443200
|
2907 |
+
},
|
2908 |
+
{
|
2909 |
+
"debugging/Self-BLEU-5": 0.6360978963817496,
|
2910 |
+
"debugging/distinct-1-grams": 0.7608773404922611,
|
2911 |
+
"debugging/distinct-2-grams": 0.9487668534919836,
|
2912 |
+
"debugging/entropy-1-grams": 6.132694731605541,
|
2913 |
+
"debugging/entropy-2-grams": 7.054469626558369,
|
2914 |
+
"debugging/length": 633.2352941176471,
|
2915 |
+
"debugging/num_segments": 17,
|
2916 |
+
"epoch": 0.47,
|
2917 |
+
"objective/train/docs_used": 687081,
|
2918 |
+
"objective/train/instantaneous_batch_size": 32,
|
2919 |
+
"objective/train/instantaneous_microbatch_size": 32768,
|
2920 |
+
"objective/train/original_loss": 2.799859046936035,
|
2921 |
+
"objective/train/theoretical_loss": 3.5586913731350327,
|
2922 |
+
"objective/train/tokens_used": 1331180000,
|
2923 |
+
"theoretical_loss": 3.5586913731350327,
|
2924 |
+
"tokens_seen": 1310720000
|
2925 |
+
},
|
2926 |
+
{
|
2927 |
+
"epoch": 0.47,
|
2928 |
+
"learning_rate": 0.0002687015320597693,
|
2929 |
+
"loss": 2.6091,
|
2930 |
+
"theoretical_loss": 3.5586913731350327,
|
2931 |
+
"tokens_seen": 1310720000
|
2932 |
+
},
|
2933 |
+
{
|
2934 |
+
"epoch": 0.47,
|
2935 |
+
"learning_rate": 0.0002681104596179308,
|
2936 |
+
"loss": 2.6058,
|
2937 |
+
"theoretical_loss": 3.557887799173889,
|
2938 |
+
"tokens_seen": 1313996800
|
2939 |
+
},
|
2940 |
+
{
|
2941 |
+
"epoch": 0.47,
|
2942 |
+
"learning_rate": 0.0002675193871760923,
|
2943 |
+
"loss": 2.618,
|
2944 |
+
"theoretical_loss": 3.557086786151754,
|
2945 |
+
"tokens_seen": 1317273600
|
2946 |
+
},
|
2947 |
+
{
|
2948 |
+
"epoch": 0.47,
|
2949 |
+
"learning_rate": 0.00026692831473425384,
|
2950 |
+
"loss": 2.5715,
|
2951 |
+
"theoretical_loss": 3.5562883195669697,
|
2952 |
+
"tokens_seen": 1320550400
|
2953 |
+
},
|
2954 |
+
{
|
2955 |
+
"epoch": 0.47,
|
2956 |
+
"learning_rate": 0.00026633724229241536,
|
2957 |
+
"loss": 2.6036,
|
2958 |
+
"theoretical_loss": 3.555492385035719,
|
2959 |
+
"tokens_seen": 1323827200
|
2960 |
+
},
|
2961 |
+
{
|
2962 |
+
"epoch": 0.47,
|
2963 |
+
"learning_rate": 0.0002657461698505769,
|
2964 |
+
"loss": 2.5559,
|
2965 |
+
"theoretical_loss": 3.5546989682907784,
|
2966 |
+
"tokens_seen": 1327104000
|
2967 |
+
},
|
2968 |
+
{
|
2969 |
+
"epoch": 0.48,
|
2970 |
+
"learning_rate": 0.00026515509740873846,
|
2971 |
+
"loss": 2.5523,
|
2972 |
+
"theoretical_loss": 3.5539080551802895,
|
2973 |
+
"tokens_seen": 1330380800
|
2974 |
+
},
|
2975 |
+
{
|
2976 |
+
"epoch": 0.48,
|
2977 |
+
"learning_rate": 0.0002645640249669,
|
2978 |
+
"loss": 2.556,
|
2979 |
+
"theoretical_loss": 3.553119631666546,
|
2980 |
+
"tokens_seen": 1333657600
|
2981 |
+
},
|
2982 |
+
{
|
2983 |
+
"epoch": 0.48,
|
2984 |
+
"learning_rate": 0.0002639729525250615,
|
2985 |
+
"loss": 2.559,
|
2986 |
+
"theoretical_loss": 3.5523336838247914,
|
2987 |
+
"tokens_seen": 1336934400
|
2988 |
+
},
|
2989 |
+
{
|
2990 |
+
"epoch": 0.48,
|
2991 |
+
"learning_rate": 0.000263381880083223,
|
2992 |
+
"loss": 2.5567,
|
2993 |
+
"theoretical_loss": 3.55155019784204,
|
2994 |
+
"tokens_seen": 1340211200
|
2995 |
+
},
|
2996 |
+
{
|
2997 |
+
"epoch": 0.48,
|
2998 |
+
"learning_rate": 0.00026279080764138454,
|
2999 |
+
"loss": 2.5476,
|
3000 |
+
"theoretical_loss": 3.5507691600159053,
|
3001 |
+
"tokens_seen": 1343488000
|
3002 |
+
},
|
3003 |
+
{
|
3004 |
+
"epoch": 0.48,
|
3005 |
+
"learning_rate": 0.00026219973519954606,
|
3006 |
+
"loss": 2.5683,
|
3007 |
+
"theoretical_loss": 3.5499905567534515,
|
3008 |
+
"tokens_seen": 1346764800
|
3009 |
+
},
|
3010 |
+
{
|
3011 |
+
"epoch": 0.48,
|
3012 |
+
"learning_rate": 0.00026160866275770763,
|
3013 |
+
"loss": 2.5248,
|
3014 |
+
"theoretical_loss": 3.549214374570052,
|
3015 |
+
"tokens_seen": 1350041600
|
3016 |
+
},
|
3017 |
+
{
|
3018 |
+
"epoch": 0.48,
|
3019 |
+
"learning_rate": 0.00026101759031586916,
|
3020 |
+
"loss": 2.5172,
|
3021 |
+
"theoretical_loss": 3.5484406000882665,
|
3022 |
+
"tokens_seen": 1353318400
|
3023 |
+
},
|
3024 |
+
{
|
3025 |
+
"epoch": 0.48,
|
3026 |
+
"learning_rate": 0.0002604265178740307,
|
3027 |
+
"loss": 2.532,
|
3028 |
+
"theoretical_loss": 3.5476692200367346,
|
3029 |
+
"tokens_seen": 1356595200
|
3030 |
+
},
|
3031 |
+
{
|
3032 |
+
"epoch": 0.49,
|
3033 |
+
"learning_rate": 0.0002598354454321922,
|
3034 |
+
"loss": 2.4928,
|
3035 |
+
"theoretical_loss": 3.546900221249076,
|
3036 |
+
"tokens_seen": 1359872000
|
3037 |
+
},
|
3038 |
+
{
|
3039 |
+
"epoch": 0.49,
|
3040 |
+
"learning_rate": 0.0002592443729903537,
|
3041 |
+
"loss": 2.5419,
|
3042 |
+
"theoretical_loss": 3.5461335906628157,
|
3043 |
+
"tokens_seen": 1363148800
|
3044 |
+
},
|
3045 |
+
{
|
3046 |
+
"epoch": 0.49,
|
3047 |
+
"learning_rate": 0.00025865330054851524,
|
3048 |
+
"loss": 2.5409,
|
3049 |
+
"theoretical_loss": 3.54536931531831,
|
3050 |
+
"tokens_seen": 1366425600
|
3051 |
+
},
|
3052 |
+
{
|
3053 |
+
"epoch": 0.49,
|
3054 |
+
"learning_rate": 0.00025806222810667676,
|
3055 |
+
"loss": 2.498,
|
3056 |
+
"theoretical_loss": 3.5446073823576985,
|
3057 |
+
"tokens_seen": 1369702400
|
3058 |
+
},
|
3059 |
+
{
|
3060 |
+
"epoch": 0.49,
|
3061 |
+
"learning_rate": 0.00025747115566483833,
|
3062 |
+
"loss": 2.5569,
|
3063 |
+
"theoretical_loss": 3.543847779023859,
|
3064 |
+
"tokens_seen": 1372979200
|
3065 |
+
},
|
3066 |
+
{
|
3067 |
+
"epoch": 0.49,
|
3068 |
+
"learning_rate": 0.0002568800832229998,
|
3069 |
+
"loss": 2.4994,
|
3070 |
+
"theoretical_loss": 3.543090492659384,
|
3071 |
+
"tokens_seen": 1376256000
|
3072 |
+
},
|
3073 |
+
{
|
3074 |
+
"epoch": 0.49,
|
3075 |
+
"learning_rate": 0.0002562890107811613,
|
3076 |
+
"loss": 2.505,
|
3077 |
+
"theoretical_loss": 3.542335510705562,
|
3078 |
+
"tokens_seen": 1379532800
|
3079 |
+
},
|
3080 |
+
{
|
3081 |
+
"epoch": 0.49,
|
3082 |
+
"learning_rate": 0.00025569793833932284,
|
3083 |
+
"loss": 2.5426,
|
3084 |
+
"theoretical_loss": 3.541582820701378,
|
3085 |
+
"tokens_seen": 1382809600
|
3086 |
+
},
|
3087 |
+
{
|
3088 |
+
"epoch": 0.5,
|
3089 |
+
"learning_rate": 0.00025510686589748436,
|
3090 |
+
"loss": 2.4956,
|
3091 |
+
"theoretical_loss": 3.5408324102825253,
|
3092 |
+
"tokens_seen": 1386086400
|
3093 |
+
},
|
3094 |
+
{
|
3095 |
+
"epoch": 0.5,
|
3096 |
+
"learning_rate": 0.0002545157934556459,
|
3097 |
+
"loss": 2.5037,
|
3098 |
+
"theoretical_loss": 3.5400842671804265,
|
3099 |
+
"tokens_seen": 1389363200
|
3100 |
+
},
|
3101 |
+
{
|
3102 |
+
"epoch": 0.5,
|
3103 |
+
"learning_rate": 0.00025392472101380746,
|
3104 |
+
"loss": 2.4902,
|
3105 |
+
"theoretical_loss": 3.5393383792212676,
|
3106 |
+
"tokens_seen": 1392640000
|
3107 |
+
},
|
3108 |
+
{
|
3109 |
+
"epoch": 0.5,
|
3110 |
+
"learning_rate": 0.000253333648571969,
|
3111 |
+
"loss": 2.5279,
|
3112 |
+
"theoretical_loss": 3.5385947343250486,
|
3113 |
+
"tokens_seen": 1395916800
|
3114 |
+
},
|
3115 |
+
{
|
3116 |
+
"epoch": 0.5,
|
3117 |
+
"learning_rate": 0.0002527425761301305,
|
3118 |
+
"loss": 2.5259,
|
3119 |
+
"theoretical_loss": 3.53785332050464,
|
3120 |
+
"tokens_seen": 1399193600
|
3121 |
+
}
|
3122 |
+
],
|
3123 |
+
"max_steps": 42724,
|
3124 |
+
"num_train_epochs": 9223372036854775807,
|
3125 |
+
"total_flos": 7.14460209610752e+17,
|
3126 |
+
"trial_name": null,
|
3127 |
+
"trial_params": null
|
3128 |
+
}
|
checkpoint-21362/training_args.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:cf81ca58c0dfd66542a066766d8b64d3076a21eca6294ebe0206318b6ec2e768
|
3 |
+
size 3451
|
checkpoint-21362/vocab.json
ADDED
The diff for this file is too large to render.
See raw diff
|
|
config.json
ADDED
@@ -0,0 +1,39 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_name_or_path": "gpt2",
|
3 |
+
"activation_function": "gelu_new",
|
4 |
+
"architectures": [
|
5 |
+
"GPT2LMAndValueHeadModel"
|
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": true,
|
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": "float32",
|
36 |
+
"transformers_version": "4.23.0",
|
37 |
+
"use_cache": true,
|
38 |
+
"vocab_size": 50261
|
39 |
+
}
|
merges.txt
ADDED
The diff for this file is too large to render.
See raw diff
|
|
pytorch_model.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:da218c411038f30f1cf9362c9240eb602fac17beb3d1688c4d423116f3011a70
|
3 |
+
size 510410301
|
special_tokens_map.json
ADDED
@@ -0,0 +1,12 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"additional_special_tokens": [
|
3 |
+
"<|aligned|>",
|
4 |
+
"<|fine|>",
|
5 |
+
"<|substandard|>",
|
6 |
+
"<|misaligned|>"
|
7 |
+
],
|
8 |
+
"bos_token": "<|endoftext|>",
|
9 |
+
"eos_token": "<|endoftext|>",
|
10 |
+
"pad_token": "<|endoftext|>",
|
11 |
+
"unk_token": "<|endoftext|>"
|
12 |
+
}
|
tokenizer.json
ADDED
The diff for this file is too large to render.
See raw diff
|
|
tokenizer_config.json
ADDED
@@ -0,0 +1,10 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"add_prefix_space": false,
|
3 |
+
"bos_token": "<|endoftext|>",
|
4 |
+
"eos_token": "<|endoftext|>",
|
5 |
+
"model_max_length": 1024,
|
6 |
+
"name_or_path": "gpt2",
|
7 |
+
"special_tokens_map_file": null,
|
8 |
+
"tokenizer_class": "GPT2Tokenizer",
|
9 |
+
"unk_token": "<|endoftext|>"
|
10 |
+
}
|
training_args.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:cf81ca58c0dfd66542a066766d8b64d3076a21eca6294ebe0206318b6ec2e768
|
3 |
+
size 3451
|
vocab.json
ADDED
The diff for this file is too large to render.
See raw diff
|
|