Initial commit
Browse files- config.json +9 -5
- optimizer.pt +3 -0
- pytorch_model.bin +1 -1
- scheduler.pt +3 -0
- special_tokens_map.json +0 -1
- tokenizer_config.json +1 -1
- trainer_state.json +736 -0
- training_args.bin +3 -0
config.json
CHANGED
|
@@ -1,14 +1,14 @@
|
|
| 1 |
{
|
| 2 |
-
"_name_or_path": "
|
|
|
|
| 3 |
"activation_function": "gelu_new",
|
| 4 |
"architectures": [
|
| 5 |
"GPT2LMHeadModel"
|
| 6 |
],
|
| 7 |
"attn_pdrop": 0.1,
|
| 8 |
-
"bos_token_id":
|
| 9 |
"embd_pdrop": 0.1,
|
| 10 |
"eos_token_id": 50256,
|
| 11 |
-
"eos_token_ids": 0,
|
| 12 |
"gradient_checkpointing": false,
|
| 13 |
"id2label": {
|
| 14 |
"0": "LABEL_0"
|
|
@@ -25,14 +25,18 @@
|
|
| 25 |
"n_inner": null,
|
| 26 |
"n_layer": 6,
|
| 27 |
"n_positions": 1024,
|
| 28 |
-
"output_past": true,
|
| 29 |
-
"pad_token_id": 0,
|
| 30 |
"resid_pdrop": 0.1,
|
| 31 |
"summary_activation": null,
|
| 32 |
"summary_first_dropout": 0.1,
|
| 33 |
"summary_proj_to_labels": true,
|
| 34 |
"summary_type": "cls_index",
|
| 35 |
"summary_use_proj": true,
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 36 |
"transformers_version": "4.2.2",
|
| 37 |
"use_cache": true,
|
| 38 |
"vocab_size": 50257
|
|
|
|
| 1 |
{
|
| 2 |
+
"_name_or_path": "distilgpt2",
|
| 3 |
+
"_num_labels": 1,
|
| 4 |
"activation_function": "gelu_new",
|
| 5 |
"architectures": [
|
| 6 |
"GPT2LMHeadModel"
|
| 7 |
],
|
| 8 |
"attn_pdrop": 0.1,
|
| 9 |
+
"bos_token_id": 50256,
|
| 10 |
"embd_pdrop": 0.1,
|
| 11 |
"eos_token_id": 50256,
|
|
|
|
| 12 |
"gradient_checkpointing": false,
|
| 13 |
"id2label": {
|
| 14 |
"0": "LABEL_0"
|
|
|
|
| 25 |
"n_inner": null,
|
| 26 |
"n_layer": 6,
|
| 27 |
"n_positions": 1024,
|
|
|
|
|
|
|
| 28 |
"resid_pdrop": 0.1,
|
| 29 |
"summary_activation": null,
|
| 30 |
"summary_first_dropout": 0.1,
|
| 31 |
"summary_proj_to_labels": true,
|
| 32 |
"summary_type": "cls_index",
|
| 33 |
"summary_use_proj": true,
|
| 34 |
+
"task_specific_params": {
|
| 35 |
+
"text-generation": {
|
| 36 |
+
"do_sample": true,
|
| 37 |
+
"max_length": 50
|
| 38 |
+
}
|
| 39 |
+
},
|
| 40 |
"transformers_version": "4.2.2",
|
| 41 |
"use_cache": true,
|
| 42 |
"vocab_size": 50257
|
optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c1a7f228293cb52073d9cb889111ef1a3b3fd22e552ce4fb794d23b36d378fff
|
| 3 |
+
size 655348359
|
pytorch_model.bin
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 333975623
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:75eeafcbd7b2fae60623a5f65bb48fa693882409c8bea4760eb75b9dd4ccf5f1
|
| 3 |
size 333975623
|
scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:28cc73830e45e37e2a5cf6251199678eeee9d11a2be8096006fa8602ad4cae30
|
| 3 |
+
size 623
|
special_tokens_map.json
DELETED
|
@@ -1 +0,0 @@
|
|
| 1 |
-
{"bos_token": {"content": "<|endoftext|>", "single_word": false, "lstrip": false, "rstrip": false, "normalized": true}, "eos_token": {"content": "<|endoftext|>", "single_word": false, "lstrip": false, "rstrip": false, "normalized": true}, "unk_token": {"content": "<|endoftext|>", "single_word": false, "lstrip": false, "rstrip": false, "normalized": true}}
|
|
|
|
|
|
tokenizer_config.json
CHANGED
|
@@ -1 +1 @@
|
|
| 1 |
-
{"
|
|
|
|
| 1 |
+
{"max_len": 1024}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,736 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 0.9921127040031748,
|
| 5 |
+
"global_step": 60000,
|
| 6 |
+
"is_hyper_param_search": false,
|
| 7 |
+
"is_local_process_zero": true,
|
| 8 |
+
"is_world_process_zero": true,
|
| 9 |
+
"log_history": [
|
| 10 |
+
{
|
| 11 |
+
"epoch": 0.01,
|
| 12 |
+
"learning_rate": 1.994488262755538e-05,
|
| 13 |
+
"loss": 3.0247,
|
| 14 |
+
"step": 500
|
| 15 |
+
},
|
| 16 |
+
{
|
| 17 |
+
"epoch": 0.02,
|
| 18 |
+
"learning_rate": 1.988976525511076e-05,
|
| 19 |
+
"loss": 3.0232,
|
| 20 |
+
"step": 1000
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"epoch": 0.02,
|
| 24 |
+
"learning_rate": 1.983464788266614e-05,
|
| 25 |
+
"loss": 3.0873,
|
| 26 |
+
"step": 1500
|
| 27 |
+
},
|
| 28 |
+
{
|
| 29 |
+
"epoch": 0.03,
|
| 30 |
+
"learning_rate": 1.977953051022152e-05,
|
| 31 |
+
"loss": 3.0891,
|
| 32 |
+
"step": 2000
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"epoch": 0.04,
|
| 36 |
+
"learning_rate": 1.9724413137776898e-05,
|
| 37 |
+
"loss": 3.0793,
|
| 38 |
+
"step": 2500
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.05,
|
| 42 |
+
"learning_rate": 1.9669295765332277e-05,
|
| 43 |
+
"loss": 3.0651,
|
| 44 |
+
"step": 3000
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.06,
|
| 48 |
+
"learning_rate": 1.9614178392887656e-05,
|
| 49 |
+
"loss": 3.0738,
|
| 50 |
+
"step": 3500
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"epoch": 0.07,
|
| 54 |
+
"learning_rate": 1.9559061020443036e-05,
|
| 55 |
+
"loss": 3.0549,
|
| 56 |
+
"step": 4000
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"epoch": 0.07,
|
| 60 |
+
"learning_rate": 1.9503943647998415e-05,
|
| 61 |
+
"loss": 3.0288,
|
| 62 |
+
"step": 4500
|
| 63 |
+
},
|
| 64 |
+
{
|
| 65 |
+
"epoch": 0.08,
|
| 66 |
+
"learning_rate": 1.9448826275553794e-05,
|
| 67 |
+
"loss": 3.0467,
|
| 68 |
+
"step": 5000
|
| 69 |
+
},
|
| 70 |
+
{
|
| 71 |
+
"epoch": 0.09,
|
| 72 |
+
"learning_rate": 1.9393708903109173e-05,
|
| 73 |
+
"loss": 3.0283,
|
| 74 |
+
"step": 5500
|
| 75 |
+
},
|
| 76 |
+
{
|
| 77 |
+
"epoch": 0.1,
|
| 78 |
+
"learning_rate": 1.9338591530664553e-05,
|
| 79 |
+
"loss": 3.0267,
|
| 80 |
+
"step": 6000
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 0.11,
|
| 84 |
+
"learning_rate": 1.9283474158219932e-05,
|
| 85 |
+
"loss": 3.0254,
|
| 86 |
+
"step": 6500
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.12,
|
| 90 |
+
"learning_rate": 1.922835678577531e-05,
|
| 91 |
+
"loss": 3.0185,
|
| 92 |
+
"step": 7000
|
| 93 |
+
},
|
| 94 |
+
{
|
| 95 |
+
"epoch": 0.12,
|
| 96 |
+
"learning_rate": 1.917323941333069e-05,
|
| 97 |
+
"loss": 2.9999,
|
| 98 |
+
"step": 7500
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 0.13,
|
| 102 |
+
"learning_rate": 1.911812204088607e-05,
|
| 103 |
+
"loss": 3.0031,
|
| 104 |
+
"step": 8000
|
| 105 |
+
},
|
| 106 |
+
{
|
| 107 |
+
"epoch": 0.14,
|
| 108 |
+
"learning_rate": 1.9063004668441445e-05,
|
| 109 |
+
"loss": 3.0106,
|
| 110 |
+
"step": 8500
|
| 111 |
+
},
|
| 112 |
+
{
|
| 113 |
+
"epoch": 0.15,
|
| 114 |
+
"learning_rate": 1.9007887295996828e-05,
|
| 115 |
+
"loss": 3.0312,
|
| 116 |
+
"step": 9000
|
| 117 |
+
},
|
| 118 |
+
{
|
| 119 |
+
"epoch": 0.16,
|
| 120 |
+
"learning_rate": 1.8952769923552204e-05,
|
| 121 |
+
"loss": 3.0034,
|
| 122 |
+
"step": 9500
|
| 123 |
+
},
|
| 124 |
+
{
|
| 125 |
+
"epoch": 0.17,
|
| 126 |
+
"learning_rate": 1.8897652551107587e-05,
|
| 127 |
+
"loss": 3.0157,
|
| 128 |
+
"step": 10000
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.17,
|
| 132 |
+
"learning_rate": 1.8842535178662962e-05,
|
| 133 |
+
"loss": 2.9945,
|
| 134 |
+
"step": 10500
|
| 135 |
+
},
|
| 136 |
+
{
|
| 137 |
+
"epoch": 0.18,
|
| 138 |
+
"learning_rate": 1.8787417806218345e-05,
|
| 139 |
+
"loss": 2.9995,
|
| 140 |
+
"step": 11000
|
| 141 |
+
},
|
| 142 |
+
{
|
| 143 |
+
"epoch": 0.19,
|
| 144 |
+
"learning_rate": 1.873230043377372e-05,
|
| 145 |
+
"loss": 2.997,
|
| 146 |
+
"step": 11500
|
| 147 |
+
},
|
| 148 |
+
{
|
| 149 |
+
"epoch": 0.2,
|
| 150 |
+
"learning_rate": 1.8677183061329104e-05,
|
| 151 |
+
"loss": 3.0017,
|
| 152 |
+
"step": 12000
|
| 153 |
+
},
|
| 154 |
+
{
|
| 155 |
+
"epoch": 0.21,
|
| 156 |
+
"learning_rate": 1.862206568888448e-05,
|
| 157 |
+
"loss": 3.0046,
|
| 158 |
+
"step": 12500
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 0.21,
|
| 162 |
+
"learning_rate": 1.8566948316439862e-05,
|
| 163 |
+
"loss": 2.9975,
|
| 164 |
+
"step": 13000
|
| 165 |
+
},
|
| 166 |
+
{
|
| 167 |
+
"epoch": 0.22,
|
| 168 |
+
"learning_rate": 1.8511830943995238e-05,
|
| 169 |
+
"loss": 2.9778,
|
| 170 |
+
"step": 13500
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 0.23,
|
| 174 |
+
"learning_rate": 1.845671357155062e-05,
|
| 175 |
+
"loss": 2.9802,
|
| 176 |
+
"step": 14000
|
| 177 |
+
},
|
| 178 |
+
{
|
| 179 |
+
"epoch": 0.24,
|
| 180 |
+
"learning_rate": 1.8401596199105996e-05,
|
| 181 |
+
"loss": 2.9603,
|
| 182 |
+
"step": 14500
|
| 183 |
+
},
|
| 184 |
+
{
|
| 185 |
+
"epoch": 0.25,
|
| 186 |
+
"learning_rate": 1.8346478826661376e-05,
|
| 187 |
+
"loss": 2.9738,
|
| 188 |
+
"step": 15000
|
| 189 |
+
},
|
| 190 |
+
{
|
| 191 |
+
"epoch": 0.26,
|
| 192 |
+
"learning_rate": 1.8291361454216755e-05,
|
| 193 |
+
"loss": 2.9776,
|
| 194 |
+
"step": 15500
|
| 195 |
+
},
|
| 196 |
+
{
|
| 197 |
+
"epoch": 0.26,
|
| 198 |
+
"learning_rate": 1.8236244081772134e-05,
|
| 199 |
+
"loss": 3.0002,
|
| 200 |
+
"step": 16000
|
| 201 |
+
},
|
| 202 |
+
{
|
| 203 |
+
"epoch": 0.27,
|
| 204 |
+
"learning_rate": 1.8181126709327513e-05,
|
| 205 |
+
"loss": 3.0026,
|
| 206 |
+
"step": 16500
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 0.28,
|
| 210 |
+
"learning_rate": 1.8126009336882893e-05,
|
| 211 |
+
"loss": 2.9791,
|
| 212 |
+
"step": 17000
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 0.29,
|
| 216 |
+
"learning_rate": 1.8070891964438272e-05,
|
| 217 |
+
"loss": 2.9974,
|
| 218 |
+
"step": 17500
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 0.3,
|
| 222 |
+
"learning_rate": 1.801577459199365e-05,
|
| 223 |
+
"loss": 2.9739,
|
| 224 |
+
"step": 18000
|
| 225 |
+
},
|
| 226 |
+
{
|
| 227 |
+
"epoch": 0.31,
|
| 228 |
+
"learning_rate": 1.796065721954903e-05,
|
| 229 |
+
"loss": 2.9733,
|
| 230 |
+
"step": 18500
|
| 231 |
+
},
|
| 232 |
+
{
|
| 233 |
+
"epoch": 0.31,
|
| 234 |
+
"learning_rate": 1.790553984710441e-05,
|
| 235 |
+
"loss": 2.9777,
|
| 236 |
+
"step": 19000
|
| 237 |
+
},
|
| 238 |
+
{
|
| 239 |
+
"epoch": 0.32,
|
| 240 |
+
"learning_rate": 1.785042247465979e-05,
|
| 241 |
+
"loss": 2.9704,
|
| 242 |
+
"step": 19500
|
| 243 |
+
},
|
| 244 |
+
{
|
| 245 |
+
"epoch": 0.33,
|
| 246 |
+
"learning_rate": 1.7795305102215168e-05,
|
| 247 |
+
"loss": 3.0007,
|
| 248 |
+
"step": 20000
|
| 249 |
+
},
|
| 250 |
+
{
|
| 251 |
+
"epoch": 0.34,
|
| 252 |
+
"learning_rate": 1.7740187729770547e-05,
|
| 253 |
+
"loss": 2.9752,
|
| 254 |
+
"step": 20500
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 0.35,
|
| 258 |
+
"learning_rate": 1.7685070357325927e-05,
|
| 259 |
+
"loss": 2.9675,
|
| 260 |
+
"step": 21000
|
| 261 |
+
},
|
| 262 |
+
{
|
| 263 |
+
"epoch": 0.36,
|
| 264 |
+
"learning_rate": 1.7629952984881306e-05,
|
| 265 |
+
"loss": 2.9655,
|
| 266 |
+
"step": 21500
|
| 267 |
+
},
|
| 268 |
+
{
|
| 269 |
+
"epoch": 0.36,
|
| 270 |
+
"learning_rate": 1.7574835612436685e-05,
|
| 271 |
+
"loss": 2.9562,
|
| 272 |
+
"step": 22000
|
| 273 |
+
},
|
| 274 |
+
{
|
| 275 |
+
"epoch": 0.37,
|
| 276 |
+
"learning_rate": 1.7519718239992064e-05,
|
| 277 |
+
"loss": 2.9767,
|
| 278 |
+
"step": 22500
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 0.38,
|
| 282 |
+
"learning_rate": 1.7464600867547444e-05,
|
| 283 |
+
"loss": 2.9622,
|
| 284 |
+
"step": 23000
|
| 285 |
+
},
|
| 286 |
+
{
|
| 287 |
+
"epoch": 0.39,
|
| 288 |
+
"learning_rate": 1.7409483495102823e-05,
|
| 289 |
+
"loss": 2.9772,
|
| 290 |
+
"step": 23500
|
| 291 |
+
},
|
| 292 |
+
{
|
| 293 |
+
"epoch": 0.4,
|
| 294 |
+
"learning_rate": 1.7354366122658202e-05,
|
| 295 |
+
"loss": 2.947,
|
| 296 |
+
"step": 24000
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 0.41,
|
| 300 |
+
"learning_rate": 1.729924875021358e-05,
|
| 301 |
+
"loss": 2.9434,
|
| 302 |
+
"step": 24500
|
| 303 |
+
},
|
| 304 |
+
{
|
| 305 |
+
"epoch": 0.41,
|
| 306 |
+
"learning_rate": 1.724413137776896e-05,
|
| 307 |
+
"loss": 2.9474,
|
| 308 |
+
"step": 25000
|
| 309 |
+
},
|
| 310 |
+
{
|
| 311 |
+
"epoch": 0.42,
|
| 312 |
+
"learning_rate": 1.718901400532434e-05,
|
| 313 |
+
"loss": 2.9502,
|
| 314 |
+
"step": 25500
|
| 315 |
+
},
|
| 316 |
+
{
|
| 317 |
+
"epoch": 0.43,
|
| 318 |
+
"learning_rate": 1.713389663287972e-05,
|
| 319 |
+
"loss": 2.9575,
|
| 320 |
+
"step": 26000
|
| 321 |
+
},
|
| 322 |
+
{
|
| 323 |
+
"epoch": 0.44,
|
| 324 |
+
"learning_rate": 1.70787792604351e-05,
|
| 325 |
+
"loss": 2.9585,
|
| 326 |
+
"step": 26500
|
| 327 |
+
},
|
| 328 |
+
{
|
| 329 |
+
"epoch": 0.45,
|
| 330 |
+
"learning_rate": 1.7023661887990478e-05,
|
| 331 |
+
"loss": 2.9681,
|
| 332 |
+
"step": 27000
|
| 333 |
+
},
|
| 334 |
+
{
|
| 335 |
+
"epoch": 0.45,
|
| 336 |
+
"learning_rate": 1.6968544515545857e-05,
|
| 337 |
+
"loss": 2.9654,
|
| 338 |
+
"step": 27500
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 0.46,
|
| 342 |
+
"learning_rate": 1.6913427143101236e-05,
|
| 343 |
+
"loss": 2.9604,
|
| 344 |
+
"step": 28000
|
| 345 |
+
},
|
| 346 |
+
{
|
| 347 |
+
"epoch": 0.47,
|
| 348 |
+
"learning_rate": 1.6858309770656615e-05,
|
| 349 |
+
"loss": 2.9787,
|
| 350 |
+
"step": 28500
|
| 351 |
+
},
|
| 352 |
+
{
|
| 353 |
+
"epoch": 0.48,
|
| 354 |
+
"learning_rate": 1.6803192398211995e-05,
|
| 355 |
+
"loss": 2.9753,
|
| 356 |
+
"step": 29000
|
| 357 |
+
},
|
| 358 |
+
{
|
| 359 |
+
"epoch": 0.49,
|
| 360 |
+
"learning_rate": 1.6748075025767374e-05,
|
| 361 |
+
"loss": 2.9553,
|
| 362 |
+
"step": 29500
|
| 363 |
+
},
|
| 364 |
+
{
|
| 365 |
+
"epoch": 0.5,
|
| 366 |
+
"learning_rate": 1.6692957653322753e-05,
|
| 367 |
+
"loss": 2.9643,
|
| 368 |
+
"step": 30000
|
| 369 |
+
},
|
| 370 |
+
{
|
| 371 |
+
"epoch": 0.5,
|
| 372 |
+
"learning_rate": 1.663784028087813e-05,
|
| 373 |
+
"loss": 2.9379,
|
| 374 |
+
"step": 30500
|
| 375 |
+
},
|
| 376 |
+
{
|
| 377 |
+
"epoch": 0.51,
|
| 378 |
+
"learning_rate": 1.658272290843351e-05,
|
| 379 |
+
"loss": 2.9358,
|
| 380 |
+
"step": 31000
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 0.52,
|
| 384 |
+
"learning_rate": 1.6527605535988887e-05,
|
| 385 |
+
"loss": 2.9484,
|
| 386 |
+
"step": 31500
|
| 387 |
+
},
|
| 388 |
+
{
|
| 389 |
+
"epoch": 0.53,
|
| 390 |
+
"learning_rate": 1.647248816354427e-05,
|
| 391 |
+
"loss": 2.925,
|
| 392 |
+
"step": 32000
|
| 393 |
+
},
|
| 394 |
+
{
|
| 395 |
+
"epoch": 0.54,
|
| 396 |
+
"learning_rate": 1.6417370791099646e-05,
|
| 397 |
+
"loss": 2.9875,
|
| 398 |
+
"step": 32500
|
| 399 |
+
},
|
| 400 |
+
{
|
| 401 |
+
"epoch": 0.55,
|
| 402 |
+
"learning_rate": 1.636225341865503e-05,
|
| 403 |
+
"loss": 2.9575,
|
| 404 |
+
"step": 33000
|
| 405 |
+
},
|
| 406 |
+
{
|
| 407 |
+
"epoch": 0.55,
|
| 408 |
+
"learning_rate": 1.6307136046210404e-05,
|
| 409 |
+
"loss": 2.9438,
|
| 410 |
+
"step": 33500
|
| 411 |
+
},
|
| 412 |
+
{
|
| 413 |
+
"epoch": 0.56,
|
| 414 |
+
"learning_rate": 1.6252018673765787e-05,
|
| 415 |
+
"loss": 2.9512,
|
| 416 |
+
"step": 34000
|
| 417 |
+
},
|
| 418 |
+
{
|
| 419 |
+
"epoch": 0.57,
|
| 420 |
+
"learning_rate": 1.6196901301321163e-05,
|
| 421 |
+
"loss": 2.9298,
|
| 422 |
+
"step": 34500
|
| 423 |
+
},
|
| 424 |
+
{
|
| 425 |
+
"epoch": 0.58,
|
| 426 |
+
"learning_rate": 1.6141783928876546e-05,
|
| 427 |
+
"loss": 2.9455,
|
| 428 |
+
"step": 35000
|
| 429 |
+
},
|
| 430 |
+
{
|
| 431 |
+
"epoch": 0.59,
|
| 432 |
+
"learning_rate": 1.608666655643192e-05,
|
| 433 |
+
"loss": 2.9515,
|
| 434 |
+
"step": 35500
|
| 435 |
+
},
|
| 436 |
+
{
|
| 437 |
+
"epoch": 0.6,
|
| 438 |
+
"learning_rate": 1.6031549183987304e-05,
|
| 439 |
+
"loss": 2.9507,
|
| 440 |
+
"step": 36000
|
| 441 |
+
},
|
| 442 |
+
{
|
| 443 |
+
"epoch": 0.6,
|
| 444 |
+
"learning_rate": 1.597643181154268e-05,
|
| 445 |
+
"loss": 2.9524,
|
| 446 |
+
"step": 36500
|
| 447 |
+
},
|
| 448 |
+
{
|
| 449 |
+
"epoch": 0.61,
|
| 450 |
+
"learning_rate": 1.5921314439098063e-05,
|
| 451 |
+
"loss": 2.9466,
|
| 452 |
+
"step": 37000
|
| 453 |
+
},
|
| 454 |
+
{
|
| 455 |
+
"epoch": 0.62,
|
| 456 |
+
"learning_rate": 1.586619706665344e-05,
|
| 457 |
+
"loss": 2.9244,
|
| 458 |
+
"step": 37500
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 0.63,
|
| 462 |
+
"learning_rate": 1.581107969420882e-05,
|
| 463 |
+
"loss": 2.9416,
|
| 464 |
+
"step": 38000
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 0.64,
|
| 468 |
+
"learning_rate": 1.5755962321764197e-05,
|
| 469 |
+
"loss": 2.9553,
|
| 470 |
+
"step": 38500
|
| 471 |
+
},
|
| 472 |
+
{
|
| 473 |
+
"epoch": 0.64,
|
| 474 |
+
"learning_rate": 1.570084494931958e-05,
|
| 475 |
+
"loss": 2.9336,
|
| 476 |
+
"step": 39000
|
| 477 |
+
},
|
| 478 |
+
{
|
| 479 |
+
"epoch": 0.65,
|
| 480 |
+
"learning_rate": 1.5645727576874955e-05,
|
| 481 |
+
"loss": 2.9226,
|
| 482 |
+
"step": 39500
|
| 483 |
+
},
|
| 484 |
+
{
|
| 485 |
+
"epoch": 0.66,
|
| 486 |
+
"learning_rate": 1.5590610204430338e-05,
|
| 487 |
+
"loss": 2.9215,
|
| 488 |
+
"step": 40000
|
| 489 |
+
},
|
| 490 |
+
{
|
| 491 |
+
"epoch": 0.67,
|
| 492 |
+
"learning_rate": 1.5535492831985714e-05,
|
| 493 |
+
"loss": 2.9129,
|
| 494 |
+
"step": 40500
|
| 495 |
+
},
|
| 496 |
+
{
|
| 497 |
+
"epoch": 0.68,
|
| 498 |
+
"learning_rate": 1.5480375459541097e-05,
|
| 499 |
+
"loss": 2.9252,
|
| 500 |
+
"step": 41000
|
| 501 |
+
},
|
| 502 |
+
{
|
| 503 |
+
"epoch": 0.69,
|
| 504 |
+
"learning_rate": 1.5425258087096472e-05,
|
| 505 |
+
"loss": 2.9433,
|
| 506 |
+
"step": 41500
|
| 507 |
+
},
|
| 508 |
+
{
|
| 509 |
+
"epoch": 0.69,
|
| 510 |
+
"learning_rate": 1.537014071465185e-05,
|
| 511 |
+
"loss": 2.9279,
|
| 512 |
+
"step": 42000
|
| 513 |
+
},
|
| 514 |
+
{
|
| 515 |
+
"epoch": 0.7,
|
| 516 |
+
"learning_rate": 1.531502334220723e-05,
|
| 517 |
+
"loss": 2.9374,
|
| 518 |
+
"step": 42500
|
| 519 |
+
},
|
| 520 |
+
{
|
| 521 |
+
"epoch": 0.71,
|
| 522 |
+
"learning_rate": 1.525990596976261e-05,
|
| 523 |
+
"loss": 2.9336,
|
| 524 |
+
"step": 43000
|
| 525 |
+
},
|
| 526 |
+
{
|
| 527 |
+
"epoch": 0.72,
|
| 528 |
+
"learning_rate": 1.520478859731799e-05,
|
| 529 |
+
"loss": 2.8991,
|
| 530 |
+
"step": 43500
|
| 531 |
+
},
|
| 532 |
+
{
|
| 533 |
+
"epoch": 0.73,
|
| 534 |
+
"learning_rate": 1.514967122487337e-05,
|
| 535 |
+
"loss": 2.9326,
|
| 536 |
+
"step": 44000
|
| 537 |
+
},
|
| 538 |
+
{
|
| 539 |
+
"epoch": 0.74,
|
| 540 |
+
"learning_rate": 1.5094553852428748e-05,
|
| 541 |
+
"loss": 2.9358,
|
| 542 |
+
"step": 44500
|
| 543 |
+
},
|
| 544 |
+
{
|
| 545 |
+
"epoch": 0.74,
|
| 546 |
+
"learning_rate": 1.5039436479984127e-05,
|
| 547 |
+
"loss": 2.9192,
|
| 548 |
+
"step": 45000
|
| 549 |
+
},
|
| 550 |
+
{
|
| 551 |
+
"epoch": 0.75,
|
| 552 |
+
"learning_rate": 1.4984319107539506e-05,
|
| 553 |
+
"loss": 2.9438,
|
| 554 |
+
"step": 45500
|
| 555 |
+
},
|
| 556 |
+
{
|
| 557 |
+
"epoch": 0.76,
|
| 558 |
+
"learning_rate": 1.4929201735094884e-05,
|
| 559 |
+
"loss": 2.9213,
|
| 560 |
+
"step": 46000
|
| 561 |
+
},
|
| 562 |
+
{
|
| 563 |
+
"epoch": 0.77,
|
| 564 |
+
"learning_rate": 1.4874084362650265e-05,
|
| 565 |
+
"loss": 2.9273,
|
| 566 |
+
"step": 46500
|
| 567 |
+
},
|
| 568 |
+
{
|
| 569 |
+
"epoch": 0.78,
|
| 570 |
+
"learning_rate": 1.4818966990205642e-05,
|
| 571 |
+
"loss": 2.925,
|
| 572 |
+
"step": 47000
|
| 573 |
+
},
|
| 574 |
+
{
|
| 575 |
+
"epoch": 0.79,
|
| 576 |
+
"learning_rate": 1.4763849617761023e-05,
|
| 577 |
+
"loss": 2.9151,
|
| 578 |
+
"step": 47500
|
| 579 |
+
},
|
| 580 |
+
{
|
| 581 |
+
"epoch": 0.79,
|
| 582 |
+
"learning_rate": 1.4708732245316401e-05,
|
| 583 |
+
"loss": 2.9195,
|
| 584 |
+
"step": 48000
|
| 585 |
+
},
|
| 586 |
+
{
|
| 587 |
+
"epoch": 0.8,
|
| 588 |
+
"learning_rate": 1.4653614872871782e-05,
|
| 589 |
+
"loss": 2.9059,
|
| 590 |
+
"step": 48500
|
| 591 |
+
},
|
| 592 |
+
{
|
| 593 |
+
"epoch": 0.81,
|
| 594 |
+
"learning_rate": 1.459849750042716e-05,
|
| 595 |
+
"loss": 2.9001,
|
| 596 |
+
"step": 49000
|
| 597 |
+
},
|
| 598 |
+
{
|
| 599 |
+
"epoch": 0.82,
|
| 600 |
+
"learning_rate": 1.454338012798254e-05,
|
| 601 |
+
"loss": 2.8995,
|
| 602 |
+
"step": 49500
|
| 603 |
+
},
|
| 604 |
+
{
|
| 605 |
+
"epoch": 0.83,
|
| 606 |
+
"learning_rate": 1.4488262755537918e-05,
|
| 607 |
+
"loss": 2.9395,
|
| 608 |
+
"step": 50000
|
| 609 |
+
},
|
| 610 |
+
{
|
| 611 |
+
"epoch": 0.84,
|
| 612 |
+
"learning_rate": 1.4433145383093299e-05,
|
| 613 |
+
"loss": 2.9066,
|
| 614 |
+
"step": 50500
|
| 615 |
+
},
|
| 616 |
+
{
|
| 617 |
+
"epoch": 0.84,
|
| 618 |
+
"learning_rate": 1.4378028010648676e-05,
|
| 619 |
+
"loss": 2.8997,
|
| 620 |
+
"step": 51000
|
| 621 |
+
},
|
| 622 |
+
{
|
| 623 |
+
"epoch": 0.85,
|
| 624 |
+
"learning_rate": 1.4322910638204057e-05,
|
| 625 |
+
"loss": 2.9083,
|
| 626 |
+
"step": 51500
|
| 627 |
+
},
|
| 628 |
+
{
|
| 629 |
+
"epoch": 0.86,
|
| 630 |
+
"learning_rate": 1.4267793265759435e-05,
|
| 631 |
+
"loss": 2.9211,
|
| 632 |
+
"step": 52000
|
| 633 |
+
},
|
| 634 |
+
{
|
| 635 |
+
"epoch": 0.87,
|
| 636 |
+
"learning_rate": 1.4212675893314816e-05,
|
| 637 |
+
"loss": 2.9067,
|
| 638 |
+
"step": 52500
|
| 639 |
+
},
|
| 640 |
+
{
|
| 641 |
+
"epoch": 0.88,
|
| 642 |
+
"learning_rate": 1.4157558520870193e-05,
|
| 643 |
+
"loss": 2.9148,
|
| 644 |
+
"step": 53000
|
| 645 |
+
},
|
| 646 |
+
{
|
| 647 |
+
"epoch": 0.88,
|
| 648 |
+
"learning_rate": 1.4102441148425574e-05,
|
| 649 |
+
"loss": 2.9172,
|
| 650 |
+
"step": 53500
|
| 651 |
+
},
|
| 652 |
+
{
|
| 653 |
+
"epoch": 0.89,
|
| 654 |
+
"learning_rate": 1.4047323775980952e-05,
|
| 655 |
+
"loss": 2.9063,
|
| 656 |
+
"step": 54000
|
| 657 |
+
},
|
| 658 |
+
{
|
| 659 |
+
"epoch": 0.9,
|
| 660 |
+
"learning_rate": 1.3992206403536333e-05,
|
| 661 |
+
"loss": 2.8933,
|
| 662 |
+
"step": 54500
|
| 663 |
+
},
|
| 664 |
+
{
|
| 665 |
+
"epoch": 0.91,
|
| 666 |
+
"learning_rate": 1.393708903109171e-05,
|
| 667 |
+
"loss": 2.9225,
|
| 668 |
+
"step": 55000
|
| 669 |
+
},
|
| 670 |
+
{
|
| 671 |
+
"epoch": 0.92,
|
| 672 |
+
"learning_rate": 1.3881971658647091e-05,
|
| 673 |
+
"loss": 2.9072,
|
| 674 |
+
"step": 55500
|
| 675 |
+
},
|
| 676 |
+
{
|
| 677 |
+
"epoch": 0.93,
|
| 678 |
+
"learning_rate": 1.3826854286202469e-05,
|
| 679 |
+
"loss": 2.909,
|
| 680 |
+
"step": 56000
|
| 681 |
+
},
|
| 682 |
+
{
|
| 683 |
+
"epoch": 0.93,
|
| 684 |
+
"learning_rate": 1.377173691375785e-05,
|
| 685 |
+
"loss": 2.8967,
|
| 686 |
+
"step": 56500
|
| 687 |
+
},
|
| 688 |
+
{
|
| 689 |
+
"epoch": 0.94,
|
| 690 |
+
"learning_rate": 1.3716619541313227e-05,
|
| 691 |
+
"loss": 2.915,
|
| 692 |
+
"step": 57000
|
| 693 |
+
},
|
| 694 |
+
{
|
| 695 |
+
"epoch": 0.95,
|
| 696 |
+
"learning_rate": 1.3661502168868608e-05,
|
| 697 |
+
"loss": 2.9118,
|
| 698 |
+
"step": 57500
|
| 699 |
+
},
|
| 700 |
+
{
|
| 701 |
+
"epoch": 0.96,
|
| 702 |
+
"learning_rate": 1.3606384796423986e-05,
|
| 703 |
+
"loss": 2.9086,
|
| 704 |
+
"step": 58000
|
| 705 |
+
},
|
| 706 |
+
{
|
| 707 |
+
"epoch": 0.97,
|
| 708 |
+
"learning_rate": 1.3551267423979367e-05,
|
| 709 |
+
"loss": 2.9099,
|
| 710 |
+
"step": 58500
|
| 711 |
+
},
|
| 712 |
+
{
|
| 713 |
+
"epoch": 0.98,
|
| 714 |
+
"learning_rate": 1.3496150051534744e-05,
|
| 715 |
+
"loss": 2.9073,
|
| 716 |
+
"step": 59000
|
| 717 |
+
},
|
| 718 |
+
{
|
| 719 |
+
"epoch": 0.98,
|
| 720 |
+
"learning_rate": 1.3441032679090122e-05,
|
| 721 |
+
"loss": 2.9182,
|
| 722 |
+
"step": 59500
|
| 723 |
+
},
|
| 724 |
+
{
|
| 725 |
+
"epoch": 0.99,
|
| 726 |
+
"learning_rate": 1.3385915306645503e-05,
|
| 727 |
+
"loss": 2.8954,
|
| 728 |
+
"step": 60000
|
| 729 |
+
}
|
| 730 |
+
],
|
| 731 |
+
"max_steps": 181431,
|
| 732 |
+
"num_train_epochs": 3,
|
| 733 |
+
"total_flos": 30196252016640000,
|
| 734 |
+
"trial_name": null,
|
| 735 |
+
"trial_params": null
|
| 736 |
+
}
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:219f6ad3e2fea352bad3a4fa86e0be557dab0edbc4f71170612673d7866dcf13
|
| 3 |
+
size 2031
|