update
Browse files- added_tokens.json +1 -0
- config.json +57 -0
- optimizer.pt +3 -0
- pytorch_model.bin +3 -0
- rng_state.pth +3 -0
- scheduler.pt +3 -0
- special_tokens_map.json +1 -0
- spiece.model +3 -0
- tokenizer_config.json +1 -0
- trainer_state.json +1246 -0
- training_args.bin +3 -0
added_tokens.json
ADDED
|
@@ -0,0 +1 @@
|
|
|
|
|
|
|
| 1 |
+
{"<sep>": 250100, "<hl>": 250101}
|
config.json
ADDED
|
@@ -0,0 +1,57 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "longcld/t5_small_qg_ae_hl",
|
| 3 |
+
"architectures": [
|
| 4 |
+
"T5ForConditionalGeneration"
|
| 5 |
+
],
|
| 6 |
+
"d_ff": 2048,
|
| 7 |
+
"d_kv": 64,
|
| 8 |
+
"d_model": 512,
|
| 9 |
+
"decoder_start_token_id": 0,
|
| 10 |
+
"dropout_rate": 0.1,
|
| 11 |
+
"eos_token_id": 1,
|
| 12 |
+
"feed_forward_proj": "relu",
|
| 13 |
+
"gradient_checkpointing": false,
|
| 14 |
+
"initializer_factor": 1.0,
|
| 15 |
+
"is_encoder_decoder": true,
|
| 16 |
+
"layer_norm_epsilon": 1e-06,
|
| 17 |
+
"model_type": "t5",
|
| 18 |
+
"n_positions": 512,
|
| 19 |
+
"num_decoder_layers": 6,
|
| 20 |
+
"num_heads": 8,
|
| 21 |
+
"num_layers": 6,
|
| 22 |
+
"output_past": true,
|
| 23 |
+
"pad_token_id": 0,
|
| 24 |
+
"relative_attention_num_buckets": 32,
|
| 25 |
+
"task_specific_params": {
|
| 26 |
+
"summarization": {
|
| 27 |
+
"early_stopping": true,
|
| 28 |
+
"length_penalty": 2.0,
|
| 29 |
+
"max_length": 200,
|
| 30 |
+
"min_length": 30,
|
| 31 |
+
"no_repeat_ngram_size": 3,
|
| 32 |
+
"num_beams": 4,
|
| 33 |
+
"prefix": "summarize: "
|
| 34 |
+
},
|
| 35 |
+
"translation_en_to_de": {
|
| 36 |
+
"early_stopping": true,
|
| 37 |
+
"max_length": 300,
|
| 38 |
+
"num_beams": 4,
|
| 39 |
+
"prefix": "translate English to German: "
|
| 40 |
+
},
|
| 41 |
+
"translation_en_to_fr": {
|
| 42 |
+
"early_stopping": true,
|
| 43 |
+
"max_length": 300,
|
| 44 |
+
"num_beams": 4,
|
| 45 |
+
"prefix": "translate English to French: "
|
| 46 |
+
},
|
| 47 |
+
"translation_en_to_ro": {
|
| 48 |
+
"early_stopping": true,
|
| 49 |
+
"max_length": 300,
|
| 50 |
+
"num_beams": 4,
|
| 51 |
+
"prefix": "translate English to Romanian: "
|
| 52 |
+
}
|
| 53 |
+
},
|
| 54 |
+
"transformers_version": "4.8.2",
|
| 55 |
+
"use_cache": true,
|
| 56 |
+
"vocab_size": 250102
|
| 57 |
+
}
|
optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:4cafcbabe44129021a4c434700895f4a381dea17217609eb51c0a718f7865419
|
| 3 |
+
size 352532601
|
pytorch_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:fc4ee4accfb37af01b780682fd323c899422d9cec7a497f8c73d2ab80eff71c9
|
| 3 |
+
size 688496379
|
rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:4cb03b397116ba5e1d68c79af691ec304b5f5ba17a9449d6c4a526074ae59ece
|
| 3 |
+
size 14503
|
scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:5bcc45b7354d78b5006455dce5c3b813f18a1360e72f327da537744731c6bae3
|
| 3 |
+
size 623
|
special_tokens_map.json
ADDED
|
@@ -0,0 +1 @@
|
|
|
|
|
|
|
| 1 |
+
{"eos_token": "</s>", "unk_token": "<unk>", "pad_token": "<pad>"}
|
spiece.model
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ef78f86560d809067d12bac6c09f19a462cb3af3f54d2b8acbba26e1433125d6
|
| 3 |
+
size 4309802
|
tokenizer_config.json
ADDED
|
@@ -0,0 +1 @@
|
|
|
|
|
|
|
| 1 |
+
{"eos_token": "</s>", "unk_token": "<unk>", "pad_token": "<pad>", "extra_ids": 0, "additional_special_tokens": null, "sp_model_kwargs": {}, "special_tokens_map_file": "C:\\Users\\hydra/.cache\\huggingface\\transformers\\bde22e2f08347211bf47b72fb4730b4b9a2e3323a6a1254932cf0fb969eb18cd.b02fba13e4584f40329b20be0bad88c2ef97b5fda7d9f4f69b1cb3a620568a86", "tokenizer_file": null, "name_or_path": "longcld/t5_small_qg_ae_hl", "tokenizer_class": "T5Tokenizer"}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,1246 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 2.7235286302643815,
|
| 5 |
+
"global_step": 20500,
|
| 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": 0.0009986714494486515,
|
| 13 |
+
"loss": 3.2529,
|
| 14 |
+
"step": 100
|
| 15 |
+
},
|
| 16 |
+
{
|
| 17 |
+
"epoch": 0.03,
|
| 18 |
+
"learning_rate": 0.000997342898897303,
|
| 19 |
+
"loss": 3.1978,
|
| 20 |
+
"step": 200
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"epoch": 0.04,
|
| 24 |
+
"learning_rate": 0.0009960143483459547,
|
| 25 |
+
"loss": 3.1849,
|
| 26 |
+
"step": 300
|
| 27 |
+
},
|
| 28 |
+
{
|
| 29 |
+
"epoch": 0.05,
|
| 30 |
+
"learning_rate": 0.0009946857977946061,
|
| 31 |
+
"loss": 3.1539,
|
| 32 |
+
"step": 400
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"epoch": 0.07,
|
| 36 |
+
"learning_rate": 0.0009933572472432576,
|
| 37 |
+
"loss": 3.1202,
|
| 38 |
+
"step": 500
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.08,
|
| 42 |
+
"learning_rate": 0.000992028696691909,
|
| 43 |
+
"loss": 3.0483,
|
| 44 |
+
"step": 600
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.09,
|
| 48 |
+
"learning_rate": 0.0009907001461405608,
|
| 49 |
+
"loss": 3.0669,
|
| 50 |
+
"step": 700
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"epoch": 0.11,
|
| 54 |
+
"learning_rate": 0.0009893715955892122,
|
| 55 |
+
"loss": 3.0665,
|
| 56 |
+
"step": 800
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"epoch": 0.12,
|
| 60 |
+
"learning_rate": 0.0009880430450378637,
|
| 61 |
+
"loss": 3.0196,
|
| 62 |
+
"step": 900
|
| 63 |
+
},
|
| 64 |
+
{
|
| 65 |
+
"epoch": 0.13,
|
| 66 |
+
"learning_rate": 0.0009867144944865152,
|
| 67 |
+
"loss": 3.009,
|
| 68 |
+
"step": 1000
|
| 69 |
+
},
|
| 70 |
+
{
|
| 71 |
+
"epoch": 0.15,
|
| 72 |
+
"learning_rate": 0.0009853859439351669,
|
| 73 |
+
"loss": 3.0046,
|
| 74 |
+
"step": 1100
|
| 75 |
+
},
|
| 76 |
+
{
|
| 77 |
+
"epoch": 0.16,
|
| 78 |
+
"learning_rate": 0.0009840573933838183,
|
| 79 |
+
"loss": 2.9368,
|
| 80 |
+
"step": 1200
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 0.17,
|
| 84 |
+
"learning_rate": 0.0009827288428324698,
|
| 85 |
+
"loss": 2.9478,
|
| 86 |
+
"step": 1300
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.19,
|
| 90 |
+
"learning_rate": 0.0009814002922811213,
|
| 91 |
+
"loss": 2.879,
|
| 92 |
+
"step": 1400
|
| 93 |
+
},
|
| 94 |
+
{
|
| 95 |
+
"epoch": 0.2,
|
| 96 |
+
"learning_rate": 0.0009800717417297728,
|
| 97 |
+
"loss": 2.8613,
|
| 98 |
+
"step": 1500
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 0.21,
|
| 102 |
+
"learning_rate": 0.0009787431911784245,
|
| 103 |
+
"loss": 2.9318,
|
| 104 |
+
"step": 1600
|
| 105 |
+
},
|
| 106 |
+
{
|
| 107 |
+
"epoch": 0.23,
|
| 108 |
+
"learning_rate": 0.000977414640627076,
|
| 109 |
+
"loss": 2.9453,
|
| 110 |
+
"step": 1700
|
| 111 |
+
},
|
| 112 |
+
{
|
| 113 |
+
"epoch": 0.24,
|
| 114 |
+
"learning_rate": 0.0009760860900757274,
|
| 115 |
+
"loss": 2.8773,
|
| 116 |
+
"step": 1800
|
| 117 |
+
},
|
| 118 |
+
{
|
| 119 |
+
"epoch": 0.25,
|
| 120 |
+
"learning_rate": 0.000974757539524379,
|
| 121 |
+
"loss": 2.9057,
|
| 122 |
+
"step": 1900
|
| 123 |
+
},
|
| 124 |
+
{
|
| 125 |
+
"epoch": 0.27,
|
| 126 |
+
"learning_rate": 0.0009734289889730305,
|
| 127 |
+
"loss": 2.8679,
|
| 128 |
+
"step": 2000
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.28,
|
| 132 |
+
"learning_rate": 0.0009721004384216819,
|
| 133 |
+
"loss": 2.8844,
|
| 134 |
+
"step": 2100
|
| 135 |
+
},
|
| 136 |
+
{
|
| 137 |
+
"epoch": 0.29,
|
| 138 |
+
"learning_rate": 0.0009707718878703334,
|
| 139 |
+
"loss": 2.8434,
|
| 140 |
+
"step": 2200
|
| 141 |
+
},
|
| 142 |
+
{
|
| 143 |
+
"epoch": 0.31,
|
| 144 |
+
"learning_rate": 0.0009694433373189851,
|
| 145 |
+
"loss": 2.8379,
|
| 146 |
+
"step": 2300
|
| 147 |
+
},
|
| 148 |
+
{
|
| 149 |
+
"epoch": 0.32,
|
| 150 |
+
"learning_rate": 0.0009681147867676366,
|
| 151 |
+
"loss": 2.8515,
|
| 152 |
+
"step": 2400
|
| 153 |
+
},
|
| 154 |
+
{
|
| 155 |
+
"epoch": 0.33,
|
| 156 |
+
"learning_rate": 0.000966786236216288,
|
| 157 |
+
"loss": 2.8525,
|
| 158 |
+
"step": 2500
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 0.35,
|
| 162 |
+
"learning_rate": 0.0009654576856649395,
|
| 163 |
+
"loss": 2.816,
|
| 164 |
+
"step": 2600
|
| 165 |
+
},
|
| 166 |
+
{
|
| 167 |
+
"epoch": 0.36,
|
| 168 |
+
"learning_rate": 0.0009641291351135911,
|
| 169 |
+
"loss": 2.8333,
|
| 170 |
+
"step": 2700
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 0.37,
|
| 174 |
+
"learning_rate": 0.0009628005845622427,
|
| 175 |
+
"loss": 2.7859,
|
| 176 |
+
"step": 2800
|
| 177 |
+
},
|
| 178 |
+
{
|
| 179 |
+
"epoch": 0.39,
|
| 180 |
+
"learning_rate": 0.0009614720340108941,
|
| 181 |
+
"loss": 2.8099,
|
| 182 |
+
"step": 2900
|
| 183 |
+
},
|
| 184 |
+
{
|
| 185 |
+
"epoch": 0.4,
|
| 186 |
+
"learning_rate": 0.0009601434834595456,
|
| 187 |
+
"loss": 2.8097,
|
| 188 |
+
"step": 3000
|
| 189 |
+
},
|
| 190 |
+
{
|
| 191 |
+
"epoch": 0.41,
|
| 192 |
+
"learning_rate": 0.0009588149329081972,
|
| 193 |
+
"loss": 2.7639,
|
| 194 |
+
"step": 3100
|
| 195 |
+
},
|
| 196 |
+
{
|
| 197 |
+
"epoch": 0.43,
|
| 198 |
+
"learning_rate": 0.0009574863823568487,
|
| 199 |
+
"loss": 2.7792,
|
| 200 |
+
"step": 3200
|
| 201 |
+
},
|
| 202 |
+
{
|
| 203 |
+
"epoch": 0.44,
|
| 204 |
+
"learning_rate": 0.0009561578318055003,
|
| 205 |
+
"loss": 2.7894,
|
| 206 |
+
"step": 3300
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 0.45,
|
| 210 |
+
"learning_rate": 0.0009548292812541517,
|
| 211 |
+
"loss": 2.7797,
|
| 212 |
+
"step": 3400
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 0.46,
|
| 216 |
+
"learning_rate": 0.0009535007307028033,
|
| 217 |
+
"loss": 2.7387,
|
| 218 |
+
"step": 3500
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 0.48,
|
| 222 |
+
"learning_rate": 0.0009521721801514548,
|
| 223 |
+
"loss": 2.7752,
|
| 224 |
+
"step": 3600
|
| 225 |
+
},
|
| 226 |
+
{
|
| 227 |
+
"epoch": 0.49,
|
| 228 |
+
"learning_rate": 0.0009508436296001063,
|
| 229 |
+
"loss": 2.6998,
|
| 230 |
+
"step": 3700
|
| 231 |
+
},
|
| 232 |
+
{
|
| 233 |
+
"epoch": 0.5,
|
| 234 |
+
"learning_rate": 0.0009495150790487578,
|
| 235 |
+
"loss": 2.7701,
|
| 236 |
+
"step": 3800
|
| 237 |
+
},
|
| 238 |
+
{
|
| 239 |
+
"epoch": 0.52,
|
| 240 |
+
"learning_rate": 0.0009481865284974094,
|
| 241 |
+
"loss": 2.7543,
|
| 242 |
+
"step": 3900
|
| 243 |
+
},
|
| 244 |
+
{
|
| 245 |
+
"epoch": 0.53,
|
| 246 |
+
"learning_rate": 0.0009468579779460609,
|
| 247 |
+
"loss": 2.6621,
|
| 248 |
+
"step": 4000
|
| 249 |
+
},
|
| 250 |
+
{
|
| 251 |
+
"epoch": 0.54,
|
| 252 |
+
"learning_rate": 0.0009455294273947124,
|
| 253 |
+
"loss": 2.6957,
|
| 254 |
+
"step": 4100
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 0.56,
|
| 258 |
+
"learning_rate": 0.0009442008768433638,
|
| 259 |
+
"loss": 2.7165,
|
| 260 |
+
"step": 4200
|
| 261 |
+
},
|
| 262 |
+
{
|
| 263 |
+
"epoch": 0.57,
|
| 264 |
+
"learning_rate": 0.0009428723262920155,
|
| 265 |
+
"loss": 2.6919,
|
| 266 |
+
"step": 4300
|
| 267 |
+
},
|
| 268 |
+
{
|
| 269 |
+
"epoch": 0.58,
|
| 270 |
+
"learning_rate": 0.000941543775740667,
|
| 271 |
+
"loss": 2.7239,
|
| 272 |
+
"step": 4400
|
| 273 |
+
},
|
| 274 |
+
{
|
| 275 |
+
"epoch": 0.6,
|
| 276 |
+
"learning_rate": 0.0009402152251893185,
|
| 277 |
+
"loss": 2.6743,
|
| 278 |
+
"step": 4500
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 0.61,
|
| 282 |
+
"learning_rate": 0.0009388866746379699,
|
| 283 |
+
"loss": 2.708,
|
| 284 |
+
"step": 4600
|
| 285 |
+
},
|
| 286 |
+
{
|
| 287 |
+
"epoch": 0.62,
|
| 288 |
+
"learning_rate": 0.0009375581240866215,
|
| 289 |
+
"loss": 2.7201,
|
| 290 |
+
"step": 4700
|
| 291 |
+
},
|
| 292 |
+
{
|
| 293 |
+
"epoch": 0.64,
|
| 294 |
+
"learning_rate": 0.000936229573535273,
|
| 295 |
+
"loss": 2.7207,
|
| 296 |
+
"step": 4800
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 0.65,
|
| 300 |
+
"learning_rate": 0.0009349010229839246,
|
| 301 |
+
"loss": 2.7046,
|
| 302 |
+
"step": 4900
|
| 303 |
+
},
|
| 304 |
+
{
|
| 305 |
+
"epoch": 0.66,
|
| 306 |
+
"learning_rate": 0.000933572472432576,
|
| 307 |
+
"loss": 2.6758,
|
| 308 |
+
"step": 5000
|
| 309 |
+
},
|
| 310 |
+
{
|
| 311 |
+
"epoch": 0.68,
|
| 312 |
+
"learning_rate": 0.0009322439218812276,
|
| 313 |
+
"loss": 2.6754,
|
| 314 |
+
"step": 5100
|
| 315 |
+
},
|
| 316 |
+
{
|
| 317 |
+
"epoch": 0.69,
|
| 318 |
+
"learning_rate": 0.0009309153713298791,
|
| 319 |
+
"loss": 2.6408,
|
| 320 |
+
"step": 5200
|
| 321 |
+
},
|
| 322 |
+
{
|
| 323 |
+
"epoch": 0.7,
|
| 324 |
+
"learning_rate": 0.0009295868207785306,
|
| 325 |
+
"loss": 2.6306,
|
| 326 |
+
"step": 5300
|
| 327 |
+
},
|
| 328 |
+
{
|
| 329 |
+
"epoch": 0.72,
|
| 330 |
+
"learning_rate": 0.0009282582702271822,
|
| 331 |
+
"loss": 2.6455,
|
| 332 |
+
"step": 5400
|
| 333 |
+
},
|
| 334 |
+
{
|
| 335 |
+
"epoch": 0.73,
|
| 336 |
+
"learning_rate": 0.0009269297196758337,
|
| 337 |
+
"loss": 2.6905,
|
| 338 |
+
"step": 5500
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 0.74,
|
| 342 |
+
"learning_rate": 0.0009256011691244852,
|
| 343 |
+
"loss": 2.6132,
|
| 344 |
+
"step": 5600
|
| 345 |
+
},
|
| 346 |
+
{
|
| 347 |
+
"epoch": 0.76,
|
| 348 |
+
"learning_rate": 0.0009242726185731367,
|
| 349 |
+
"loss": 2.6048,
|
| 350 |
+
"step": 5700
|
| 351 |
+
},
|
| 352 |
+
{
|
| 353 |
+
"epoch": 0.77,
|
| 354 |
+
"learning_rate": 0.0009229440680217882,
|
| 355 |
+
"loss": 2.5945,
|
| 356 |
+
"step": 5800
|
| 357 |
+
},
|
| 358 |
+
{
|
| 359 |
+
"epoch": 0.78,
|
| 360 |
+
"learning_rate": 0.0009216155174704398,
|
| 361 |
+
"loss": 2.6384,
|
| 362 |
+
"step": 5900
|
| 363 |
+
},
|
| 364 |
+
{
|
| 365 |
+
"epoch": 0.8,
|
| 366 |
+
"learning_rate": 0.0009202869669190913,
|
| 367 |
+
"loss": 2.615,
|
| 368 |
+
"step": 6000
|
| 369 |
+
},
|
| 370 |
+
{
|
| 371 |
+
"epoch": 0.81,
|
| 372 |
+
"learning_rate": 0.0009189584163677428,
|
| 373 |
+
"loss": 2.6281,
|
| 374 |
+
"step": 6100
|
| 375 |
+
},
|
| 376 |
+
{
|
| 377 |
+
"epoch": 0.82,
|
| 378 |
+
"learning_rate": 0.0009176298658163943,
|
| 379 |
+
"loss": 2.5965,
|
| 380 |
+
"step": 6200
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 0.84,
|
| 384 |
+
"learning_rate": 0.0009163013152650458,
|
| 385 |
+
"loss": 2.6704,
|
| 386 |
+
"step": 6300
|
| 387 |
+
},
|
| 388 |
+
{
|
| 389 |
+
"epoch": 0.85,
|
| 390 |
+
"learning_rate": 0.0009149727647136974,
|
| 391 |
+
"loss": 2.5665,
|
| 392 |
+
"step": 6400
|
| 393 |
+
},
|
| 394 |
+
{
|
| 395 |
+
"epoch": 0.86,
|
| 396 |
+
"learning_rate": 0.0009136442141623489,
|
| 397 |
+
"loss": 2.6324,
|
| 398 |
+
"step": 6500
|
| 399 |
+
},
|
| 400 |
+
{
|
| 401 |
+
"epoch": 0.88,
|
| 402 |
+
"learning_rate": 0.0009123156636110004,
|
| 403 |
+
"loss": 2.6258,
|
| 404 |
+
"step": 6600
|
| 405 |
+
},
|
| 406 |
+
{
|
| 407 |
+
"epoch": 0.89,
|
| 408 |
+
"learning_rate": 0.000910987113059652,
|
| 409 |
+
"loss": 2.6006,
|
| 410 |
+
"step": 6700
|
| 411 |
+
},
|
| 412 |
+
{
|
| 413 |
+
"epoch": 0.9,
|
| 414 |
+
"learning_rate": 0.0009096585625083034,
|
| 415 |
+
"loss": 2.5908,
|
| 416 |
+
"step": 6800
|
| 417 |
+
},
|
| 418 |
+
{
|
| 419 |
+
"epoch": 0.92,
|
| 420 |
+
"learning_rate": 0.000908330011956955,
|
| 421 |
+
"loss": 2.5855,
|
| 422 |
+
"step": 6900
|
| 423 |
+
},
|
| 424 |
+
{
|
| 425 |
+
"epoch": 0.93,
|
| 426 |
+
"learning_rate": 0.0009070014614056065,
|
| 427 |
+
"loss": 2.6421,
|
| 428 |
+
"step": 7000
|
| 429 |
+
},
|
| 430 |
+
{
|
| 431 |
+
"epoch": 0.94,
|
| 432 |
+
"learning_rate": 0.0009056729108542581,
|
| 433 |
+
"loss": 2.6084,
|
| 434 |
+
"step": 7100
|
| 435 |
+
},
|
| 436 |
+
{
|
| 437 |
+
"epoch": 0.96,
|
| 438 |
+
"learning_rate": 0.0009043443603029095,
|
| 439 |
+
"loss": 2.6102,
|
| 440 |
+
"step": 7200
|
| 441 |
+
},
|
| 442 |
+
{
|
| 443 |
+
"epoch": 0.97,
|
| 444 |
+
"learning_rate": 0.000903015809751561,
|
| 445 |
+
"loss": 2.5543,
|
| 446 |
+
"step": 7300
|
| 447 |
+
},
|
| 448 |
+
{
|
| 449 |
+
"epoch": 0.98,
|
| 450 |
+
"learning_rate": 0.0009016872592002126,
|
| 451 |
+
"loss": 2.5746,
|
| 452 |
+
"step": 7400
|
| 453 |
+
},
|
| 454 |
+
{
|
| 455 |
+
"epoch": 1.0,
|
| 456 |
+
"learning_rate": 0.0009003587086488642,
|
| 457 |
+
"loss": 2.5401,
|
| 458 |
+
"step": 7500
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 1.01,
|
| 462 |
+
"learning_rate": 0.0008990301580975156,
|
| 463 |
+
"loss": 2.489,
|
| 464 |
+
"step": 7600
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 1.02,
|
| 468 |
+
"learning_rate": 0.0008977016075461671,
|
| 469 |
+
"loss": 2.5211,
|
| 470 |
+
"step": 7700
|
| 471 |
+
},
|
| 472 |
+
{
|
| 473 |
+
"epoch": 1.04,
|
| 474 |
+
"learning_rate": 0.0008963730569948186,
|
| 475 |
+
"loss": 2.548,
|
| 476 |
+
"step": 7800
|
| 477 |
+
},
|
| 478 |
+
{
|
| 479 |
+
"epoch": 1.05,
|
| 480 |
+
"learning_rate": 0.0008950445064434703,
|
| 481 |
+
"loss": 2.5165,
|
| 482 |
+
"step": 7900
|
| 483 |
+
},
|
| 484 |
+
{
|
| 485 |
+
"epoch": 1.06,
|
| 486 |
+
"learning_rate": 0.0008937159558921218,
|
| 487 |
+
"loss": 2.4884,
|
| 488 |
+
"step": 8000
|
| 489 |
+
},
|
| 490 |
+
{
|
| 491 |
+
"epoch": 1.08,
|
| 492 |
+
"learning_rate": 0.0008923874053407732,
|
| 493 |
+
"loss": 2.4682,
|
| 494 |
+
"step": 8100
|
| 495 |
+
},
|
| 496 |
+
{
|
| 497 |
+
"epoch": 1.09,
|
| 498 |
+
"learning_rate": 0.0008910588547894247,
|
| 499 |
+
"loss": 2.4754,
|
| 500 |
+
"step": 8200
|
| 501 |
+
},
|
| 502 |
+
{
|
| 503 |
+
"epoch": 1.1,
|
| 504 |
+
"learning_rate": 0.0008897303042380763,
|
| 505 |
+
"loss": 2.5076,
|
| 506 |
+
"step": 8300
|
| 507 |
+
},
|
| 508 |
+
{
|
| 509 |
+
"epoch": 1.12,
|
| 510 |
+
"learning_rate": 0.0008884017536867279,
|
| 511 |
+
"loss": 2.5088,
|
| 512 |
+
"step": 8400
|
| 513 |
+
},
|
| 514 |
+
{
|
| 515 |
+
"epoch": 1.13,
|
| 516 |
+
"learning_rate": 0.0008870732031353793,
|
| 517 |
+
"loss": 2.4578,
|
| 518 |
+
"step": 8500
|
| 519 |
+
},
|
| 520 |
+
{
|
| 521 |
+
"epoch": 1.14,
|
| 522 |
+
"learning_rate": 0.0008857446525840308,
|
| 523 |
+
"loss": 2.4815,
|
| 524 |
+
"step": 8600
|
| 525 |
+
},
|
| 526 |
+
{
|
| 527 |
+
"epoch": 1.16,
|
| 528 |
+
"learning_rate": 0.0008844161020326824,
|
| 529 |
+
"loss": 2.5504,
|
| 530 |
+
"step": 8700
|
| 531 |
+
},
|
| 532 |
+
{
|
| 533 |
+
"epoch": 1.17,
|
| 534 |
+
"learning_rate": 0.0008830875514813339,
|
| 535 |
+
"loss": 2.4883,
|
| 536 |
+
"step": 8800
|
| 537 |
+
},
|
| 538 |
+
{
|
| 539 |
+
"epoch": 1.18,
|
| 540 |
+
"learning_rate": 0.0008817590009299854,
|
| 541 |
+
"loss": 2.4907,
|
| 542 |
+
"step": 8900
|
| 543 |
+
},
|
| 544 |
+
{
|
| 545 |
+
"epoch": 1.2,
|
| 546 |
+
"learning_rate": 0.0008804304503786369,
|
| 547 |
+
"loss": 2.4817,
|
| 548 |
+
"step": 9000
|
| 549 |
+
},
|
| 550 |
+
{
|
| 551 |
+
"epoch": 1.21,
|
| 552 |
+
"learning_rate": 0.0008791018998272885,
|
| 553 |
+
"loss": 2.4803,
|
| 554 |
+
"step": 9100
|
| 555 |
+
},
|
| 556 |
+
{
|
| 557 |
+
"epoch": 1.22,
|
| 558 |
+
"learning_rate": 0.00087777334927594,
|
| 559 |
+
"loss": 2.461,
|
| 560 |
+
"step": 9200
|
| 561 |
+
},
|
| 562 |
+
{
|
| 563 |
+
"epoch": 1.24,
|
| 564 |
+
"learning_rate": 0.0008764447987245914,
|
| 565 |
+
"loss": 2.4541,
|
| 566 |
+
"step": 9300
|
| 567 |
+
},
|
| 568 |
+
{
|
| 569 |
+
"epoch": 1.25,
|
| 570 |
+
"learning_rate": 0.000875116248173243,
|
| 571 |
+
"loss": 2.436,
|
| 572 |
+
"step": 9400
|
| 573 |
+
},
|
| 574 |
+
{
|
| 575 |
+
"epoch": 1.26,
|
| 576 |
+
"learning_rate": 0.0008737876976218946,
|
| 577 |
+
"loss": 2.4998,
|
| 578 |
+
"step": 9500
|
| 579 |
+
},
|
| 580 |
+
{
|
| 581 |
+
"epoch": 1.28,
|
| 582 |
+
"learning_rate": 0.0008724591470705461,
|
| 583 |
+
"loss": 2.4387,
|
| 584 |
+
"step": 9600
|
| 585 |
+
},
|
| 586 |
+
{
|
| 587 |
+
"epoch": 1.29,
|
| 588 |
+
"learning_rate": 0.0008711305965191976,
|
| 589 |
+
"loss": 2.5248,
|
| 590 |
+
"step": 9700
|
| 591 |
+
},
|
| 592 |
+
{
|
| 593 |
+
"epoch": 1.3,
|
| 594 |
+
"learning_rate": 0.000869802045967849,
|
| 595 |
+
"loss": 2.4285,
|
| 596 |
+
"step": 9800
|
| 597 |
+
},
|
| 598 |
+
{
|
| 599 |
+
"epoch": 1.32,
|
| 600 |
+
"learning_rate": 0.0008684734954165007,
|
| 601 |
+
"loss": 2.4363,
|
| 602 |
+
"step": 9900
|
| 603 |
+
},
|
| 604 |
+
{
|
| 605 |
+
"epoch": 1.33,
|
| 606 |
+
"learning_rate": 0.0008671449448651522,
|
| 607 |
+
"loss": 2.4629,
|
| 608 |
+
"step": 10000
|
| 609 |
+
},
|
| 610 |
+
{
|
| 611 |
+
"epoch": 1.34,
|
| 612 |
+
"learning_rate": 0.0008658163943138037,
|
| 613 |
+
"loss": 2.463,
|
| 614 |
+
"step": 10100
|
| 615 |
+
},
|
| 616 |
+
{
|
| 617 |
+
"epoch": 1.36,
|
| 618 |
+
"learning_rate": 0.0008644878437624551,
|
| 619 |
+
"loss": 2.5058,
|
| 620 |
+
"step": 10200
|
| 621 |
+
},
|
| 622 |
+
{
|
| 623 |
+
"epoch": 1.37,
|
| 624 |
+
"learning_rate": 0.0008631592932111067,
|
| 625 |
+
"loss": 2.4056,
|
| 626 |
+
"step": 10300
|
| 627 |
+
},
|
| 628 |
+
{
|
| 629 |
+
"epoch": 1.38,
|
| 630 |
+
"learning_rate": 0.0008618307426597583,
|
| 631 |
+
"loss": 2.4111,
|
| 632 |
+
"step": 10400
|
| 633 |
+
},
|
| 634 |
+
{
|
| 635 |
+
"epoch": 1.39,
|
| 636 |
+
"learning_rate": 0.0008605021921084098,
|
| 637 |
+
"loss": 2.4352,
|
| 638 |
+
"step": 10500
|
| 639 |
+
},
|
| 640 |
+
{
|
| 641 |
+
"epoch": 1.41,
|
| 642 |
+
"learning_rate": 0.0008591736415570612,
|
| 643 |
+
"loss": 2.4886,
|
| 644 |
+
"step": 10600
|
| 645 |
+
},
|
| 646 |
+
{
|
| 647 |
+
"epoch": 1.42,
|
| 648 |
+
"learning_rate": 0.0008578450910057127,
|
| 649 |
+
"loss": 2.4433,
|
| 650 |
+
"step": 10700
|
| 651 |
+
},
|
| 652 |
+
{
|
| 653 |
+
"epoch": 1.43,
|
| 654 |
+
"learning_rate": 0.0008565165404543643,
|
| 655 |
+
"loss": 2.471,
|
| 656 |
+
"step": 10800
|
| 657 |
+
},
|
| 658 |
+
{
|
| 659 |
+
"epoch": 1.45,
|
| 660 |
+
"learning_rate": 0.0008551879899030159,
|
| 661 |
+
"loss": 2.4446,
|
| 662 |
+
"step": 10900
|
| 663 |
+
},
|
| 664 |
+
{
|
| 665 |
+
"epoch": 1.46,
|
| 666 |
+
"learning_rate": 0.0008538594393516674,
|
| 667 |
+
"loss": 2.4567,
|
| 668 |
+
"step": 11000
|
| 669 |
+
},
|
| 670 |
+
{
|
| 671 |
+
"epoch": 1.47,
|
| 672 |
+
"learning_rate": 0.0008525308888003188,
|
| 673 |
+
"loss": 2.5048,
|
| 674 |
+
"step": 11100
|
| 675 |
+
},
|
| 676 |
+
{
|
| 677 |
+
"epoch": 1.49,
|
| 678 |
+
"learning_rate": 0.0008512023382489704,
|
| 679 |
+
"loss": 2.4119,
|
| 680 |
+
"step": 11200
|
| 681 |
+
},
|
| 682 |
+
{
|
| 683 |
+
"epoch": 1.5,
|
| 684 |
+
"learning_rate": 0.0008498737876976219,
|
| 685 |
+
"loss": 2.4098,
|
| 686 |
+
"step": 11300
|
| 687 |
+
},
|
| 688 |
+
{
|
| 689 |
+
"epoch": 1.51,
|
| 690 |
+
"learning_rate": 0.0008485452371462735,
|
| 691 |
+
"loss": 2.4108,
|
| 692 |
+
"step": 11400
|
| 693 |
+
},
|
| 694 |
+
{
|
| 695 |
+
"epoch": 1.53,
|
| 696 |
+
"learning_rate": 0.0008472166865949249,
|
| 697 |
+
"loss": 2.4016,
|
| 698 |
+
"step": 11500
|
| 699 |
+
},
|
| 700 |
+
{
|
| 701 |
+
"epoch": 1.54,
|
| 702 |
+
"learning_rate": 0.0008458881360435765,
|
| 703 |
+
"loss": 2.422,
|
| 704 |
+
"step": 11600
|
| 705 |
+
},
|
| 706 |
+
{
|
| 707 |
+
"epoch": 1.55,
|
| 708 |
+
"learning_rate": 0.000844559585492228,
|
| 709 |
+
"loss": 2.4645,
|
| 710 |
+
"step": 11700
|
| 711 |
+
},
|
| 712 |
+
{
|
| 713 |
+
"epoch": 1.57,
|
| 714 |
+
"learning_rate": 0.0008432310349408795,
|
| 715 |
+
"loss": 2.4218,
|
| 716 |
+
"step": 11800
|
| 717 |
+
},
|
| 718 |
+
{
|
| 719 |
+
"epoch": 1.58,
|
| 720 |
+
"learning_rate": 0.000841902484389531,
|
| 721 |
+
"loss": 2.4213,
|
| 722 |
+
"step": 11900
|
| 723 |
+
},
|
| 724 |
+
{
|
| 725 |
+
"epoch": 1.59,
|
| 726 |
+
"learning_rate": 0.0008405739338381826,
|
| 727 |
+
"loss": 2.389,
|
| 728 |
+
"step": 12000
|
| 729 |
+
},
|
| 730 |
+
{
|
| 731 |
+
"epoch": 1.61,
|
| 732 |
+
"learning_rate": 0.0008392453832868341,
|
| 733 |
+
"loss": 2.3751,
|
| 734 |
+
"step": 12100
|
| 735 |
+
},
|
| 736 |
+
{
|
| 737 |
+
"epoch": 1.62,
|
| 738 |
+
"learning_rate": 0.0008379168327354856,
|
| 739 |
+
"loss": 2.4552,
|
| 740 |
+
"step": 12200
|
| 741 |
+
},
|
| 742 |
+
{
|
| 743 |
+
"epoch": 1.63,
|
| 744 |
+
"learning_rate": 0.000836588282184137,
|
| 745 |
+
"loss": 2.4137,
|
| 746 |
+
"step": 12300
|
| 747 |
+
},
|
| 748 |
+
{
|
| 749 |
+
"epoch": 1.65,
|
| 750 |
+
"learning_rate": 0.0008352597316327887,
|
| 751 |
+
"loss": 2.3814,
|
| 752 |
+
"step": 12400
|
| 753 |
+
},
|
| 754 |
+
{
|
| 755 |
+
"epoch": 1.66,
|
| 756 |
+
"learning_rate": 0.0008339311810814402,
|
| 757 |
+
"loss": 2.3798,
|
| 758 |
+
"step": 12500
|
| 759 |
+
},
|
| 760 |
+
{
|
| 761 |
+
"epoch": 1.67,
|
| 762 |
+
"learning_rate": 0.0008326026305300917,
|
| 763 |
+
"loss": 2.384,
|
| 764 |
+
"step": 12600
|
| 765 |
+
},
|
| 766 |
+
{
|
| 767 |
+
"epoch": 1.69,
|
| 768 |
+
"learning_rate": 0.0008312740799787431,
|
| 769 |
+
"loss": 2.4628,
|
| 770 |
+
"step": 12700
|
| 771 |
+
},
|
| 772 |
+
{
|
| 773 |
+
"epoch": 1.7,
|
| 774 |
+
"learning_rate": 0.0008299455294273947,
|
| 775 |
+
"loss": 2.4149,
|
| 776 |
+
"step": 12800
|
| 777 |
+
},
|
| 778 |
+
{
|
| 779 |
+
"epoch": 1.71,
|
| 780 |
+
"learning_rate": 0.0008286169788760463,
|
| 781 |
+
"loss": 2.4083,
|
| 782 |
+
"step": 12900
|
| 783 |
+
},
|
| 784 |
+
{
|
| 785 |
+
"epoch": 1.73,
|
| 786 |
+
"learning_rate": 0.0008272884283246978,
|
| 787 |
+
"loss": 2.4374,
|
| 788 |
+
"step": 13000
|
| 789 |
+
},
|
| 790 |
+
{
|
| 791 |
+
"epoch": 1.74,
|
| 792 |
+
"learning_rate": 0.0008259598777733493,
|
| 793 |
+
"loss": 2.3741,
|
| 794 |
+
"step": 13100
|
| 795 |
+
},
|
| 796 |
+
{
|
| 797 |
+
"epoch": 1.75,
|
| 798 |
+
"learning_rate": 0.0008246313272220008,
|
| 799 |
+
"loss": 2.3797,
|
| 800 |
+
"step": 13200
|
| 801 |
+
},
|
| 802 |
+
{
|
| 803 |
+
"epoch": 1.77,
|
| 804 |
+
"learning_rate": 0.0008233027766706523,
|
| 805 |
+
"loss": 2.3884,
|
| 806 |
+
"step": 13300
|
| 807 |
+
},
|
| 808 |
+
{
|
| 809 |
+
"epoch": 1.78,
|
| 810 |
+
"learning_rate": 0.0008219742261193038,
|
| 811 |
+
"loss": 2.3931,
|
| 812 |
+
"step": 13400
|
| 813 |
+
},
|
| 814 |
+
{
|
| 815 |
+
"epoch": 1.79,
|
| 816 |
+
"learning_rate": 0.0008206456755679554,
|
| 817 |
+
"loss": 2.4227,
|
| 818 |
+
"step": 13500
|
| 819 |
+
},
|
| 820 |
+
{
|
| 821 |
+
"epoch": 1.81,
|
| 822 |
+
"learning_rate": 0.000819317125016607,
|
| 823 |
+
"loss": 2.417,
|
| 824 |
+
"step": 13600
|
| 825 |
+
},
|
| 826 |
+
{
|
| 827 |
+
"epoch": 1.82,
|
| 828 |
+
"learning_rate": 0.0008179885744652584,
|
| 829 |
+
"loss": 2.3735,
|
| 830 |
+
"step": 13700
|
| 831 |
+
},
|
| 832 |
+
{
|
| 833 |
+
"epoch": 1.83,
|
| 834 |
+
"learning_rate": 0.0008166600239139099,
|
| 835 |
+
"loss": 2.4168,
|
| 836 |
+
"step": 13800
|
| 837 |
+
},
|
| 838 |
+
{
|
| 839 |
+
"epoch": 1.85,
|
| 840 |
+
"learning_rate": 0.0008153314733625614,
|
| 841 |
+
"loss": 2.4211,
|
| 842 |
+
"step": 13900
|
| 843 |
+
},
|
| 844 |
+
{
|
| 845 |
+
"epoch": 1.86,
|
| 846 |
+
"learning_rate": 0.0008140029228112131,
|
| 847 |
+
"loss": 2.4075,
|
| 848 |
+
"step": 14000
|
| 849 |
+
},
|
| 850 |
+
{
|
| 851 |
+
"epoch": 1.87,
|
| 852 |
+
"learning_rate": 0.0008126743722598645,
|
| 853 |
+
"loss": 2.385,
|
| 854 |
+
"step": 14100
|
| 855 |
+
},
|
| 856 |
+
{
|
| 857 |
+
"epoch": 1.89,
|
| 858 |
+
"learning_rate": 0.000811345821708516,
|
| 859 |
+
"loss": 2.3975,
|
| 860 |
+
"step": 14200
|
| 861 |
+
},
|
| 862 |
+
{
|
| 863 |
+
"epoch": 1.9,
|
| 864 |
+
"learning_rate": 0.0008100172711571675,
|
| 865 |
+
"loss": 2.4189,
|
| 866 |
+
"step": 14300
|
| 867 |
+
},
|
| 868 |
+
{
|
| 869 |
+
"epoch": 1.91,
|
| 870 |
+
"learning_rate": 0.0008086887206058191,
|
| 871 |
+
"loss": 2.4126,
|
| 872 |
+
"step": 14400
|
| 873 |
+
},
|
| 874 |
+
{
|
| 875 |
+
"epoch": 1.93,
|
| 876 |
+
"learning_rate": 0.0008073601700544706,
|
| 877 |
+
"loss": 2.4129,
|
| 878 |
+
"step": 14500
|
| 879 |
+
},
|
| 880 |
+
{
|
| 881 |
+
"epoch": 1.94,
|
| 882 |
+
"learning_rate": 0.0008060316195031221,
|
| 883 |
+
"loss": 2.3709,
|
| 884 |
+
"step": 14600
|
| 885 |
+
},
|
| 886 |
+
{
|
| 887 |
+
"epoch": 1.95,
|
| 888 |
+
"learning_rate": 0.0008047030689517736,
|
| 889 |
+
"loss": 2.3723,
|
| 890 |
+
"step": 14700
|
| 891 |
+
},
|
| 892 |
+
{
|
| 893 |
+
"epoch": 1.97,
|
| 894 |
+
"learning_rate": 0.0008033745184004252,
|
| 895 |
+
"loss": 2.3315,
|
| 896 |
+
"step": 14800
|
| 897 |
+
},
|
| 898 |
+
{
|
| 899 |
+
"epoch": 1.98,
|
| 900 |
+
"learning_rate": 0.0008020459678490766,
|
| 901 |
+
"loss": 2.3557,
|
| 902 |
+
"step": 14900
|
| 903 |
+
},
|
| 904 |
+
{
|
| 905 |
+
"epoch": 1.99,
|
| 906 |
+
"learning_rate": 0.0008007174172977282,
|
| 907 |
+
"loss": 2.3327,
|
| 908 |
+
"step": 15000
|
| 909 |
+
},
|
| 910 |
+
{
|
| 911 |
+
"epoch": 2.01,
|
| 912 |
+
"learning_rate": 0.0007993888667463797,
|
| 913 |
+
"loss": 2.3288,
|
| 914 |
+
"step": 15100
|
| 915 |
+
},
|
| 916 |
+
{
|
| 917 |
+
"epoch": 2.02,
|
| 918 |
+
"learning_rate": 0.0007980603161950313,
|
| 919 |
+
"loss": 2.2879,
|
| 920 |
+
"step": 15200
|
| 921 |
+
},
|
| 922 |
+
{
|
| 923 |
+
"epoch": 2.03,
|
| 924 |
+
"learning_rate": 0.0007967317656436827,
|
| 925 |
+
"loss": 2.3298,
|
| 926 |
+
"step": 15300
|
| 927 |
+
},
|
| 928 |
+
{
|
| 929 |
+
"epoch": 2.05,
|
| 930 |
+
"learning_rate": 0.0007954032150923342,
|
| 931 |
+
"loss": 2.3611,
|
| 932 |
+
"step": 15400
|
| 933 |
+
},
|
| 934 |
+
{
|
| 935 |
+
"epoch": 2.06,
|
| 936 |
+
"learning_rate": 0.0007940746645409858,
|
| 937 |
+
"loss": 2.2829,
|
| 938 |
+
"step": 15500
|
| 939 |
+
},
|
| 940 |
+
{
|
| 941 |
+
"epoch": 2.07,
|
| 942 |
+
"learning_rate": 0.0007927461139896374,
|
| 943 |
+
"loss": 2.2791,
|
| 944 |
+
"step": 15600
|
| 945 |
+
},
|
| 946 |
+
{
|
| 947 |
+
"epoch": 2.09,
|
| 948 |
+
"learning_rate": 0.0007914175634382889,
|
| 949 |
+
"loss": 2.3107,
|
| 950 |
+
"step": 15700
|
| 951 |
+
},
|
| 952 |
+
{
|
| 953 |
+
"epoch": 2.1,
|
| 954 |
+
"learning_rate": 0.0007900890128869403,
|
| 955 |
+
"loss": 2.2678,
|
| 956 |
+
"step": 15800
|
| 957 |
+
},
|
| 958 |
+
{
|
| 959 |
+
"epoch": 2.11,
|
| 960 |
+
"learning_rate": 0.0007887604623355918,
|
| 961 |
+
"loss": 2.2461,
|
| 962 |
+
"step": 15900
|
| 963 |
+
},
|
| 964 |
+
{
|
| 965 |
+
"epoch": 2.13,
|
| 966 |
+
"learning_rate": 0.0007874319117842435,
|
| 967 |
+
"loss": 2.3019,
|
| 968 |
+
"step": 16000
|
| 969 |
+
},
|
| 970 |
+
{
|
| 971 |
+
"epoch": 2.14,
|
| 972 |
+
"learning_rate": 0.000786103361232895,
|
| 973 |
+
"loss": 2.3085,
|
| 974 |
+
"step": 16100
|
| 975 |
+
},
|
| 976 |
+
{
|
| 977 |
+
"epoch": 2.15,
|
| 978 |
+
"learning_rate": 0.0007847748106815464,
|
| 979 |
+
"loss": 2.2968,
|
| 980 |
+
"step": 16200
|
| 981 |
+
},
|
| 982 |
+
{
|
| 983 |
+
"epoch": 2.17,
|
| 984 |
+
"learning_rate": 0.0007834462601301979,
|
| 985 |
+
"loss": 2.3325,
|
| 986 |
+
"step": 16300
|
| 987 |
+
},
|
| 988 |
+
{
|
| 989 |
+
"epoch": 2.18,
|
| 990 |
+
"learning_rate": 0.0007821177095788495,
|
| 991 |
+
"loss": 2.2987,
|
| 992 |
+
"step": 16400
|
| 993 |
+
},
|
| 994 |
+
{
|
| 995 |
+
"epoch": 2.19,
|
| 996 |
+
"learning_rate": 0.0007807891590275011,
|
| 997 |
+
"loss": 2.3043,
|
| 998 |
+
"step": 16500
|
| 999 |
+
},
|
| 1000 |
+
{
|
| 1001 |
+
"epoch": 2.21,
|
| 1002 |
+
"learning_rate": 0.0007794606084761525,
|
| 1003 |
+
"loss": 2.2743,
|
| 1004 |
+
"step": 16600
|
| 1005 |
+
},
|
| 1006 |
+
{
|
| 1007 |
+
"epoch": 2.22,
|
| 1008 |
+
"learning_rate": 0.000778132057924804,
|
| 1009 |
+
"loss": 2.3369,
|
| 1010 |
+
"step": 16700
|
| 1011 |
+
},
|
| 1012 |
+
{
|
| 1013 |
+
"epoch": 2.23,
|
| 1014 |
+
"learning_rate": 0.0007768035073734556,
|
| 1015 |
+
"loss": 2.3084,
|
| 1016 |
+
"step": 16800
|
| 1017 |
+
},
|
| 1018 |
+
{
|
| 1019 |
+
"epoch": 2.25,
|
| 1020 |
+
"learning_rate": 0.0007754749568221071,
|
| 1021 |
+
"loss": 2.2688,
|
| 1022 |
+
"step": 16900
|
| 1023 |
+
},
|
| 1024 |
+
{
|
| 1025 |
+
"epoch": 2.26,
|
| 1026 |
+
"learning_rate": 0.0007741464062707587,
|
| 1027 |
+
"loss": 2.3203,
|
| 1028 |
+
"step": 17000
|
| 1029 |
+
},
|
| 1030 |
+
{
|
| 1031 |
+
"epoch": 2.27,
|
| 1032 |
+
"learning_rate": 0.0007728178557194101,
|
| 1033 |
+
"loss": 2.2956,
|
| 1034 |
+
"step": 17100
|
| 1035 |
+
},
|
| 1036 |
+
{
|
| 1037 |
+
"epoch": 2.29,
|
| 1038 |
+
"learning_rate": 0.0007714893051680617,
|
| 1039 |
+
"loss": 2.2939,
|
| 1040 |
+
"step": 17200
|
| 1041 |
+
},
|
| 1042 |
+
{
|
| 1043 |
+
"epoch": 2.3,
|
| 1044 |
+
"learning_rate": 0.0007701607546167132,
|
| 1045 |
+
"loss": 2.3379,
|
| 1046 |
+
"step": 17300
|
| 1047 |
+
},
|
| 1048 |
+
{
|
| 1049 |
+
"epoch": 2.31,
|
| 1050 |
+
"learning_rate": 0.0007688322040653647,
|
| 1051 |
+
"loss": 2.3161,
|
| 1052 |
+
"step": 17400
|
| 1053 |
+
},
|
| 1054 |
+
{
|
| 1055 |
+
"epoch": 2.32,
|
| 1056 |
+
"learning_rate": 0.0007675036535140162,
|
| 1057 |
+
"loss": 2.2423,
|
| 1058 |
+
"step": 17500
|
| 1059 |
+
},
|
| 1060 |
+
{
|
| 1061 |
+
"epoch": 2.34,
|
| 1062 |
+
"learning_rate": 0.0007661751029626678,
|
| 1063 |
+
"loss": 2.2982,
|
| 1064 |
+
"step": 17600
|
| 1065 |
+
},
|
| 1066 |
+
{
|
| 1067 |
+
"epoch": 2.35,
|
| 1068 |
+
"learning_rate": 0.0007648465524113193,
|
| 1069 |
+
"loss": 2.3019,
|
| 1070 |
+
"step": 17700
|
| 1071 |
+
},
|
| 1072 |
+
{
|
| 1073 |
+
"epoch": 2.36,
|
| 1074 |
+
"learning_rate": 0.0007635180018599708,
|
| 1075 |
+
"loss": 2.3133,
|
| 1076 |
+
"step": 17800
|
| 1077 |
+
},
|
| 1078 |
+
{
|
| 1079 |
+
"epoch": 2.38,
|
| 1080 |
+
"learning_rate": 0.0007621894513086222,
|
| 1081 |
+
"loss": 2.3209,
|
| 1082 |
+
"step": 17900
|
| 1083 |
+
},
|
| 1084 |
+
{
|
| 1085 |
+
"epoch": 2.39,
|
| 1086 |
+
"learning_rate": 0.0007608609007572739,
|
| 1087 |
+
"loss": 2.2685,
|
| 1088 |
+
"step": 18000
|
| 1089 |
+
},
|
| 1090 |
+
{
|
| 1091 |
+
"epoch": 2.4,
|
| 1092 |
+
"learning_rate": 0.0007595323502059254,
|
| 1093 |
+
"loss": 2.2641,
|
| 1094 |
+
"step": 18100
|
| 1095 |
+
},
|
| 1096 |
+
{
|
| 1097 |
+
"epoch": 2.42,
|
| 1098 |
+
"learning_rate": 0.0007582037996545769,
|
| 1099 |
+
"loss": 2.2284,
|
| 1100 |
+
"step": 18200
|
| 1101 |
+
},
|
| 1102 |
+
{
|
| 1103 |
+
"epoch": 2.43,
|
| 1104 |
+
"learning_rate": 0.0007568752491032283,
|
| 1105 |
+
"loss": 2.2937,
|
| 1106 |
+
"step": 18300
|
| 1107 |
+
},
|
| 1108 |
+
{
|
| 1109 |
+
"epoch": 2.44,
|
| 1110 |
+
"learning_rate": 0.0007555466985518799,
|
| 1111 |
+
"loss": 2.304,
|
| 1112 |
+
"step": 18400
|
| 1113 |
+
},
|
| 1114 |
+
{
|
| 1115 |
+
"epoch": 2.46,
|
| 1116 |
+
"learning_rate": 0.0007542181480005315,
|
| 1117 |
+
"loss": 2.2327,
|
| 1118 |
+
"step": 18500
|
| 1119 |
+
},
|
| 1120 |
+
{
|
| 1121 |
+
"epoch": 2.47,
|
| 1122 |
+
"learning_rate": 0.000752889597449183,
|
| 1123 |
+
"loss": 2.2515,
|
| 1124 |
+
"step": 18600
|
| 1125 |
+
},
|
| 1126 |
+
{
|
| 1127 |
+
"epoch": 2.48,
|
| 1128 |
+
"learning_rate": 0.0007515610468978344,
|
| 1129 |
+
"loss": 2.2844,
|
| 1130 |
+
"step": 18700
|
| 1131 |
+
},
|
| 1132 |
+
{
|
| 1133 |
+
"epoch": 2.5,
|
| 1134 |
+
"learning_rate": 0.000750232496346486,
|
| 1135 |
+
"loss": 2.2547,
|
| 1136 |
+
"step": 18800
|
| 1137 |
+
},
|
| 1138 |
+
{
|
| 1139 |
+
"epoch": 2.51,
|
| 1140 |
+
"learning_rate": 0.0007489039457951375,
|
| 1141 |
+
"loss": 2.2519,
|
| 1142 |
+
"step": 18900
|
| 1143 |
+
},
|
| 1144 |
+
{
|
| 1145 |
+
"epoch": 2.52,
|
| 1146 |
+
"learning_rate": 0.0007475753952437891,
|
| 1147 |
+
"loss": 2.2452,
|
| 1148 |
+
"step": 19000
|
| 1149 |
+
},
|
| 1150 |
+
{
|
| 1151 |
+
"epoch": 2.54,
|
| 1152 |
+
"learning_rate": 0.0007462468446924406,
|
| 1153 |
+
"loss": 2.2554,
|
| 1154 |
+
"step": 19100
|
| 1155 |
+
},
|
| 1156 |
+
{
|
| 1157 |
+
"epoch": 2.55,
|
| 1158 |
+
"learning_rate": 0.0007449182941410921,
|
| 1159 |
+
"loss": 2.2747,
|
| 1160 |
+
"step": 19200
|
| 1161 |
+
},
|
| 1162 |
+
{
|
| 1163 |
+
"epoch": 2.56,
|
| 1164 |
+
"learning_rate": 0.0007435897435897436,
|
| 1165 |
+
"loss": 2.2507,
|
| 1166 |
+
"step": 19300
|
| 1167 |
+
},
|
| 1168 |
+
{
|
| 1169 |
+
"epoch": 2.58,
|
| 1170 |
+
"learning_rate": 0.0007422611930383951,
|
| 1171 |
+
"loss": 2.2791,
|
| 1172 |
+
"step": 19400
|
| 1173 |
+
},
|
| 1174 |
+
{
|
| 1175 |
+
"epoch": 2.59,
|
| 1176 |
+
"learning_rate": 0.0007409326424870467,
|
| 1177 |
+
"loss": 2.2677,
|
| 1178 |
+
"step": 19500
|
| 1179 |
+
},
|
| 1180 |
+
{
|
| 1181 |
+
"epoch": 2.6,
|
| 1182 |
+
"learning_rate": 0.0007396040919356982,
|
| 1183 |
+
"loss": 2.3149,
|
| 1184 |
+
"step": 19600
|
| 1185 |
+
},
|
| 1186 |
+
{
|
| 1187 |
+
"epoch": 2.62,
|
| 1188 |
+
"learning_rate": 0.0007382755413843497,
|
| 1189 |
+
"loss": 2.27,
|
| 1190 |
+
"step": 19700
|
| 1191 |
+
},
|
| 1192 |
+
{
|
| 1193 |
+
"epoch": 2.63,
|
| 1194 |
+
"learning_rate": 0.0007369469908330012,
|
| 1195 |
+
"loss": 2.2977,
|
| 1196 |
+
"step": 19800
|
| 1197 |
+
},
|
| 1198 |
+
{
|
| 1199 |
+
"epoch": 2.64,
|
| 1200 |
+
"learning_rate": 0.0007356184402816527,
|
| 1201 |
+
"loss": 2.2718,
|
| 1202 |
+
"step": 19900
|
| 1203 |
+
},
|
| 1204 |
+
{
|
| 1205 |
+
"epoch": 2.66,
|
| 1206 |
+
"learning_rate": 0.0007342898897303044,
|
| 1207 |
+
"loss": 2.2379,
|
| 1208 |
+
"step": 20000
|
| 1209 |
+
},
|
| 1210 |
+
{
|
| 1211 |
+
"epoch": 2.67,
|
| 1212 |
+
"learning_rate": 0.0007329613391789558,
|
| 1213 |
+
"loss": 2.2691,
|
| 1214 |
+
"step": 20100
|
| 1215 |
+
},
|
| 1216 |
+
{
|
| 1217 |
+
"epoch": 2.68,
|
| 1218 |
+
"learning_rate": 0.0007316327886276073,
|
| 1219 |
+
"loss": 2.2407,
|
| 1220 |
+
"step": 20200
|
| 1221 |
+
},
|
| 1222 |
+
{
|
| 1223 |
+
"epoch": 2.7,
|
| 1224 |
+
"learning_rate": 0.0007303042380762588,
|
| 1225 |
+
"loss": 2.2707,
|
| 1226 |
+
"step": 20300
|
| 1227 |
+
},
|
| 1228 |
+
{
|
| 1229 |
+
"epoch": 2.71,
|
| 1230 |
+
"learning_rate": 0.0007289756875249104,
|
| 1231 |
+
"loss": 2.2163,
|
| 1232 |
+
"step": 20400
|
| 1233 |
+
},
|
| 1234 |
+
{
|
| 1235 |
+
"epoch": 2.72,
|
| 1236 |
+
"learning_rate": 0.0007276471369735619,
|
| 1237 |
+
"loss": 2.2541,
|
| 1238 |
+
"step": 20500
|
| 1239 |
+
}
|
| 1240 |
+
],
|
| 1241 |
+
"max_steps": 75270,
|
| 1242 |
+
"num_train_epochs": 10,
|
| 1243 |
+
"total_flos": 2.418543044232192e+17,
|
| 1244 |
+
"trial_name": null,
|
| 1245 |
+
"trial_params": null
|
| 1246 |
+
}
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:3454634fd5586c72136bb00fb38f3f6a0d3aa089ebb4ddc27a07c6040ccb2541
|
| 3 |
+
size 2607
|