Upload 41 files
Browse files- config.json +26 -0
- merges.txt +0 -0
- rng_state_0.pth +3 -0
- rng_state_1.pth +3 -0
- rng_state_10.pth +3 -0
- rng_state_11.pth +3 -0
- rng_state_12.pth +3 -0
- rng_state_13.pth +3 -0
- rng_state_14.pth +3 -0
- rng_state_15.pth +3 -0
- rng_state_16.pth +3 -0
- rng_state_17.pth +3 -0
- rng_state_18.pth +3 -0
- rng_state_19.pth +3 -0
- rng_state_2.pth +3 -0
- rng_state_20.pth +3 -0
- rng_state_21.pth +3 -0
- rng_state_22.pth +3 -0
- rng_state_23.pth +3 -0
- rng_state_24.pth +3 -0
- rng_state_25.pth +3 -0
- rng_state_26.pth +3 -0
- rng_state_27.pth +3 -0
- rng_state_28.pth +3 -0
- rng_state_29.pth +3 -0
- rng_state_3.pth +3 -0
- rng_state_30.pth +3 -0
- rng_state_31.pth +3 -0
- rng_state_4.pth +3 -0
- rng_state_5.pth +3 -0
- rng_state_6.pth +3 -0
- rng_state_7.pth +3 -0
- rng_state_8.pth +3 -0
- rng_state_9.pth +3 -0
- scaler.pt +3 -0
- scheduler.pt +3 -0
- special_tokens_map.json +51 -0
- tokenizer_config.json +66 -0
- trainer_state.json +1204 -0
- training_args.bin +3 -0
- vocab.json +0 -0
config.json
ADDED
|
@@ -0,0 +1,26 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"architectures": [
|
| 3 |
+
"CamembertForMaskedLM"
|
| 4 |
+
],
|
| 5 |
+
"attention_probs_dropout_prob": 0.1,
|
| 6 |
+
"bos_token_id": 0,
|
| 7 |
+
"classifier_dropout": null,
|
| 8 |
+
"eos_token_id": 2,
|
| 9 |
+
"hidden_act": "gelu",
|
| 10 |
+
"hidden_dropout_prob": 0.1,
|
| 11 |
+
"hidden_size": 768,
|
| 12 |
+
"initializer_range": 0.02,
|
| 13 |
+
"intermediate_size": 3072,
|
| 14 |
+
"layer_norm_eps": 1e-12,
|
| 15 |
+
"max_position_embeddings": 514,
|
| 16 |
+
"model_type": "camembert",
|
| 17 |
+
"num_attention_heads": 12,
|
| 18 |
+
"num_hidden_layers": 12,
|
| 19 |
+
"pad_token_id": 1,
|
| 20 |
+
"position_embedding_type": "absolute",
|
| 21 |
+
"torch_dtype": "float32",
|
| 22 |
+
"transformers_version": "4.22.2",
|
| 23 |
+
"type_vocab_size": 1,
|
| 24 |
+
"use_cache": true,
|
| 25 |
+
"vocab_size": 32000
|
| 26 |
+
}
|
merges.txt
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
rng_state_0.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:894d0fc8b9c6882c03f493c377e8494b2de18719c8ed3a4b41dbefa9c39bfed1
|
| 3 |
+
size 14503
|
rng_state_1.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:894d0fc8b9c6882c03f493c377e8494b2de18719c8ed3a4b41dbefa9c39bfed1
|
| 3 |
+
size 14503
|
rng_state_10.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:894d0fc8b9c6882c03f493c377e8494b2de18719c8ed3a4b41dbefa9c39bfed1
|
| 3 |
+
size 14503
|
rng_state_11.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:894d0fc8b9c6882c03f493c377e8494b2de18719c8ed3a4b41dbefa9c39bfed1
|
| 3 |
+
size 14503
|
rng_state_12.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:894d0fc8b9c6882c03f493c377e8494b2de18719c8ed3a4b41dbefa9c39bfed1
|
| 3 |
+
size 14503
|
rng_state_13.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:894d0fc8b9c6882c03f493c377e8494b2de18719c8ed3a4b41dbefa9c39bfed1
|
| 3 |
+
size 14503
|
rng_state_14.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:894d0fc8b9c6882c03f493c377e8494b2de18719c8ed3a4b41dbefa9c39bfed1
|
| 3 |
+
size 14503
|
rng_state_15.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:894d0fc8b9c6882c03f493c377e8494b2de18719c8ed3a4b41dbefa9c39bfed1
|
| 3 |
+
size 14503
|
rng_state_16.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:894d0fc8b9c6882c03f493c377e8494b2de18719c8ed3a4b41dbefa9c39bfed1
|
| 3 |
+
size 14503
|
rng_state_17.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:894d0fc8b9c6882c03f493c377e8494b2de18719c8ed3a4b41dbefa9c39bfed1
|
| 3 |
+
size 14503
|
rng_state_18.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:894d0fc8b9c6882c03f493c377e8494b2de18719c8ed3a4b41dbefa9c39bfed1
|
| 3 |
+
size 14503
|
rng_state_19.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:894d0fc8b9c6882c03f493c377e8494b2de18719c8ed3a4b41dbefa9c39bfed1
|
| 3 |
+
size 14503
|
rng_state_2.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:894d0fc8b9c6882c03f493c377e8494b2de18719c8ed3a4b41dbefa9c39bfed1
|
| 3 |
+
size 14503
|
rng_state_20.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:894d0fc8b9c6882c03f493c377e8494b2de18719c8ed3a4b41dbefa9c39bfed1
|
| 3 |
+
size 14503
|
rng_state_21.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:894d0fc8b9c6882c03f493c377e8494b2de18719c8ed3a4b41dbefa9c39bfed1
|
| 3 |
+
size 14503
|
rng_state_22.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:894d0fc8b9c6882c03f493c377e8494b2de18719c8ed3a4b41dbefa9c39bfed1
|
| 3 |
+
size 14503
|
rng_state_23.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:894d0fc8b9c6882c03f493c377e8494b2de18719c8ed3a4b41dbefa9c39bfed1
|
| 3 |
+
size 14503
|
rng_state_24.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:894d0fc8b9c6882c03f493c377e8494b2de18719c8ed3a4b41dbefa9c39bfed1
|
| 3 |
+
size 14503
|
rng_state_25.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:894d0fc8b9c6882c03f493c377e8494b2de18719c8ed3a4b41dbefa9c39bfed1
|
| 3 |
+
size 14503
|
rng_state_26.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:894d0fc8b9c6882c03f493c377e8494b2de18719c8ed3a4b41dbefa9c39bfed1
|
| 3 |
+
size 14503
|
rng_state_27.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:894d0fc8b9c6882c03f493c377e8494b2de18719c8ed3a4b41dbefa9c39bfed1
|
| 3 |
+
size 14503
|
rng_state_28.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:894d0fc8b9c6882c03f493c377e8494b2de18719c8ed3a4b41dbefa9c39bfed1
|
| 3 |
+
size 14503
|
rng_state_29.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:894d0fc8b9c6882c03f493c377e8494b2de18719c8ed3a4b41dbefa9c39bfed1
|
| 3 |
+
size 14503
|
rng_state_3.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:894d0fc8b9c6882c03f493c377e8494b2de18719c8ed3a4b41dbefa9c39bfed1
|
| 3 |
+
size 14503
|
rng_state_30.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:894d0fc8b9c6882c03f493c377e8494b2de18719c8ed3a4b41dbefa9c39bfed1
|
| 3 |
+
size 14503
|
rng_state_31.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:894d0fc8b9c6882c03f493c377e8494b2de18719c8ed3a4b41dbefa9c39bfed1
|
| 3 |
+
size 14503
|
rng_state_4.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:894d0fc8b9c6882c03f493c377e8494b2de18719c8ed3a4b41dbefa9c39bfed1
|
| 3 |
+
size 14503
|
rng_state_5.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:894d0fc8b9c6882c03f493c377e8494b2de18719c8ed3a4b41dbefa9c39bfed1
|
| 3 |
+
size 14503
|
rng_state_6.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:894d0fc8b9c6882c03f493c377e8494b2de18719c8ed3a4b41dbefa9c39bfed1
|
| 3 |
+
size 14503
|
rng_state_7.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:894d0fc8b9c6882c03f493c377e8494b2de18719c8ed3a4b41dbefa9c39bfed1
|
| 3 |
+
size 14503
|
rng_state_8.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:894d0fc8b9c6882c03f493c377e8494b2de18719c8ed3a4b41dbefa9c39bfed1
|
| 3 |
+
size 14503
|
rng_state_9.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:894d0fc8b9c6882c03f493c377e8494b2de18719c8ed3a4b41dbefa9c39bfed1
|
| 3 |
+
size 14503
|
scaler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:2170e94f031df4d02a7464802de5e6601bedd6a498154fb84de6a607114971e8
|
| 3 |
+
size 559
|
scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:eea1ee6a90286d7616e47b55d1528529f287f5fc513d5350f9cfc7b7a3fbb5ea
|
| 3 |
+
size 623
|
special_tokens_map.json
ADDED
|
@@ -0,0 +1,51 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": {
|
| 3 |
+
"content": "<s>",
|
| 4 |
+
"lstrip": false,
|
| 5 |
+
"normalized": true,
|
| 6 |
+
"rstrip": false,
|
| 7 |
+
"single_word": false
|
| 8 |
+
},
|
| 9 |
+
"cls_token": {
|
| 10 |
+
"content": "<s>",
|
| 11 |
+
"lstrip": false,
|
| 12 |
+
"normalized": true,
|
| 13 |
+
"rstrip": false,
|
| 14 |
+
"single_word": false
|
| 15 |
+
},
|
| 16 |
+
"eos_token": {
|
| 17 |
+
"content": "</s>",
|
| 18 |
+
"lstrip": false,
|
| 19 |
+
"normalized": true,
|
| 20 |
+
"rstrip": false,
|
| 21 |
+
"single_word": false
|
| 22 |
+
},
|
| 23 |
+
"mask_token": {
|
| 24 |
+
"content": "<mask>",
|
| 25 |
+
"lstrip": true,
|
| 26 |
+
"normalized": true,
|
| 27 |
+
"rstrip": false,
|
| 28 |
+
"single_word": false
|
| 29 |
+
},
|
| 30 |
+
"pad_token": {
|
| 31 |
+
"content": "<pad>",
|
| 32 |
+
"lstrip": false,
|
| 33 |
+
"normalized": true,
|
| 34 |
+
"rstrip": false,
|
| 35 |
+
"single_word": false
|
| 36 |
+
},
|
| 37 |
+
"sep_token": {
|
| 38 |
+
"content": "</s>",
|
| 39 |
+
"lstrip": false,
|
| 40 |
+
"normalized": true,
|
| 41 |
+
"rstrip": false,
|
| 42 |
+
"single_word": false
|
| 43 |
+
},
|
| 44 |
+
"unk_token": {
|
| 45 |
+
"content": "<unk>",
|
| 46 |
+
"lstrip": false,
|
| 47 |
+
"normalized": true,
|
| 48 |
+
"rstrip": false,
|
| 49 |
+
"single_word": false
|
| 50 |
+
}
|
| 51 |
+
}
|
tokenizer_config.json
ADDED
|
@@ -0,0 +1,66 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"add_prefix_space": false,
|
| 3 |
+
"bos_token": {
|
| 4 |
+
"__type": "AddedToken",
|
| 5 |
+
"content": "<s>",
|
| 6 |
+
"lstrip": false,
|
| 7 |
+
"normalized": true,
|
| 8 |
+
"rstrip": false,
|
| 9 |
+
"single_word": false
|
| 10 |
+
},
|
| 11 |
+
"clean_up_tokenization_spaces": true,
|
| 12 |
+
"cls_token": {
|
| 13 |
+
"__type": "AddedToken",
|
| 14 |
+
"content": "<s>",
|
| 15 |
+
"lstrip": false,
|
| 16 |
+
"normalized": true,
|
| 17 |
+
"rstrip": false,
|
| 18 |
+
"single_word": false
|
| 19 |
+
},
|
| 20 |
+
"eos_token": {
|
| 21 |
+
"__type": "AddedToken",
|
| 22 |
+
"content": "</s>",
|
| 23 |
+
"lstrip": false,
|
| 24 |
+
"normalized": true,
|
| 25 |
+
"rstrip": false,
|
| 26 |
+
"single_word": false
|
| 27 |
+
},
|
| 28 |
+
"errors": "replace",
|
| 29 |
+
"mask_token": {
|
| 30 |
+
"__type": "AddedToken",
|
| 31 |
+
"content": "<mask>",
|
| 32 |
+
"lstrip": true,
|
| 33 |
+
"normalized": true,
|
| 34 |
+
"rstrip": false,
|
| 35 |
+
"single_word": false
|
| 36 |
+
},
|
| 37 |
+
"model_max_length": 1000000000000000019884624838656,
|
| 38 |
+
"name_or_path": "./tokenizers/BpeTokenizer_PubMed_Abstracts_CHARS_lowercased_fixed_utf8-roberta/",
|
| 39 |
+
"pad_token": {
|
| 40 |
+
"__type": "AddedToken",
|
| 41 |
+
"content": "<pad>",
|
| 42 |
+
"lstrip": false,
|
| 43 |
+
"normalized": true,
|
| 44 |
+
"rstrip": false,
|
| 45 |
+
"single_word": false
|
| 46 |
+
},
|
| 47 |
+
"sep_token": {
|
| 48 |
+
"__type": "AddedToken",
|
| 49 |
+
"content": "</s>",
|
| 50 |
+
"lstrip": false,
|
| 51 |
+
"normalized": true,
|
| 52 |
+
"rstrip": false,
|
| 53 |
+
"single_word": false
|
| 54 |
+
},
|
| 55 |
+
"special_tokens_map_file": "./tokenizers/BpeTokenizer_PubMed_Abstracts_CHARS_lowercased_fixed_utf8-roberta/special_tokens_map.json",
|
| 56 |
+
"tokenizer_class": "RobertaTokenizer",
|
| 57 |
+
"unk_token": {
|
| 58 |
+
"__type": "AddedToken",
|
| 59 |
+
"content": "<unk>",
|
| 60 |
+
"lstrip": false,
|
| 61 |
+
"normalized": true,
|
| 62 |
+
"rstrip": false,
|
| 63 |
+
"single_word": false
|
| 64 |
+
},
|
| 65 |
+
"use_fast": true
|
| 66 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,1204 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 18.0,
|
| 5 |
+
"global_step": 98676,
|
| 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": 5e-09,
|
| 13 |
+
"loss": 10.532,
|
| 14 |
+
"step": 1
|
| 15 |
+
},
|
| 16 |
+
{
|
| 17 |
+
"epoch": 0.09,
|
| 18 |
+
"learning_rate": 2.5e-06,
|
| 19 |
+
"loss": 9.367,
|
| 20 |
+
"step": 500
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"epoch": 0.18,
|
| 24 |
+
"learning_rate": 5e-06,
|
| 25 |
+
"loss": 7.4277,
|
| 26 |
+
"step": 1000
|
| 27 |
+
},
|
| 28 |
+
{
|
| 29 |
+
"epoch": 0.27,
|
| 30 |
+
"learning_rate": 7.5e-06,
|
| 31 |
+
"loss": 6.1401,
|
| 32 |
+
"step": 1500
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"epoch": 0.36,
|
| 36 |
+
"learning_rate": 1e-05,
|
| 37 |
+
"loss": 5.8709,
|
| 38 |
+
"step": 2000
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.46,
|
| 42 |
+
"learning_rate": 1.25e-05,
|
| 43 |
+
"loss": 5.753,
|
| 44 |
+
"step": 2500
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.55,
|
| 48 |
+
"learning_rate": 1.5e-05,
|
| 49 |
+
"loss": 5.6777,
|
| 50 |
+
"step": 3000
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"epoch": 0.64,
|
| 54 |
+
"learning_rate": 1.75e-05,
|
| 55 |
+
"loss": 5.6151,
|
| 56 |
+
"step": 3500
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"epoch": 0.73,
|
| 60 |
+
"learning_rate": 2e-05,
|
| 61 |
+
"loss": 5.5717,
|
| 62 |
+
"step": 4000
|
| 63 |
+
},
|
| 64 |
+
{
|
| 65 |
+
"epoch": 0.82,
|
| 66 |
+
"learning_rate": 2.25e-05,
|
| 67 |
+
"loss": 5.5305,
|
| 68 |
+
"step": 4500
|
| 69 |
+
},
|
| 70 |
+
{
|
| 71 |
+
"epoch": 0.91,
|
| 72 |
+
"learning_rate": 2.5e-05,
|
| 73 |
+
"loss": 5.4947,
|
| 74 |
+
"step": 5000
|
| 75 |
+
},
|
| 76 |
+
{
|
| 77 |
+
"epoch": 1.0,
|
| 78 |
+
"learning_rate": 2.7500000000000004e-05,
|
| 79 |
+
"loss": 5.4688,
|
| 80 |
+
"step": 5500
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 1.09,
|
| 84 |
+
"learning_rate": 3e-05,
|
| 85 |
+
"loss": 5.4406,
|
| 86 |
+
"step": 6000
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 1.19,
|
| 90 |
+
"learning_rate": 3.2500000000000004e-05,
|
| 91 |
+
"loss": 5.4163,
|
| 92 |
+
"step": 6500
|
| 93 |
+
},
|
| 94 |
+
{
|
| 95 |
+
"epoch": 1.28,
|
| 96 |
+
"learning_rate": 3.5e-05,
|
| 97 |
+
"loss": 5.3942,
|
| 98 |
+
"step": 7000
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 1.37,
|
| 102 |
+
"learning_rate": 3.7500000000000003e-05,
|
| 103 |
+
"loss": 5.3762,
|
| 104 |
+
"step": 7500
|
| 105 |
+
},
|
| 106 |
+
{
|
| 107 |
+
"epoch": 1.46,
|
| 108 |
+
"learning_rate": 4e-05,
|
| 109 |
+
"loss": 5.3524,
|
| 110 |
+
"step": 8000
|
| 111 |
+
},
|
| 112 |
+
{
|
| 113 |
+
"epoch": 1.55,
|
| 114 |
+
"learning_rate": 4.2495e-05,
|
| 115 |
+
"loss": 5.338,
|
| 116 |
+
"step": 8500
|
| 117 |
+
},
|
| 118 |
+
{
|
| 119 |
+
"epoch": 1.64,
|
| 120 |
+
"learning_rate": 4.4995000000000005e-05,
|
| 121 |
+
"loss": 5.3205,
|
| 122 |
+
"step": 9000
|
| 123 |
+
},
|
| 124 |
+
{
|
| 125 |
+
"epoch": 1.73,
|
| 126 |
+
"learning_rate": 4.7495e-05,
|
| 127 |
+
"loss": 5.3096,
|
| 128 |
+
"step": 9500
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 1.82,
|
| 132 |
+
"learning_rate": 4.9995000000000005e-05,
|
| 133 |
+
"loss": 5.2971,
|
| 134 |
+
"step": 10000
|
| 135 |
+
},
|
| 136 |
+
{
|
| 137 |
+
"epoch": 1.92,
|
| 138 |
+
"learning_rate": 4.998859263331501e-05,
|
| 139 |
+
"loss": 5.2822,
|
| 140 |
+
"step": 10500
|
| 141 |
+
},
|
| 142 |
+
{
|
| 143 |
+
"epoch": 2.01,
|
| 144 |
+
"learning_rate": 4.9977139453912406e-05,
|
| 145 |
+
"loss": 5.2716,
|
| 146 |
+
"step": 11000
|
| 147 |
+
},
|
| 148 |
+
{
|
| 149 |
+
"epoch": 2.1,
|
| 150 |
+
"learning_rate": 4.99656862745098e-05,
|
| 151 |
+
"loss": 5.2599,
|
| 152 |
+
"step": 11500
|
| 153 |
+
},
|
| 154 |
+
{
|
| 155 |
+
"epoch": 2.19,
|
| 156 |
+
"learning_rate": 4.9954233095107204e-05,
|
| 157 |
+
"loss": 5.2516,
|
| 158 |
+
"step": 12000
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 2.28,
|
| 162 |
+
"learning_rate": 4.99427799157046e-05,
|
| 163 |
+
"loss": 5.2391,
|
| 164 |
+
"step": 12500
|
| 165 |
+
},
|
| 166 |
+
{
|
| 167 |
+
"epoch": 2.37,
|
| 168 |
+
"learning_rate": 4.9931326736301996e-05,
|
| 169 |
+
"loss": 5.2352,
|
| 170 |
+
"step": 13000
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 2.46,
|
| 174 |
+
"learning_rate": 4.99198964632582e-05,
|
| 175 |
+
"loss": 5.2247,
|
| 176 |
+
"step": 13500
|
| 177 |
+
},
|
| 178 |
+
{
|
| 179 |
+
"epoch": 2.55,
|
| 180 |
+
"learning_rate": 4.99084432838556e-05,
|
| 181 |
+
"loss": 5.2168,
|
| 182 |
+
"step": 14000
|
| 183 |
+
},
|
| 184 |
+
{
|
| 185 |
+
"epoch": 2.65,
|
| 186 |
+
"learning_rate": 4.9896990104453e-05,
|
| 187 |
+
"loss": 5.2091,
|
| 188 |
+
"step": 14500
|
| 189 |
+
},
|
| 190 |
+
{
|
| 191 |
+
"epoch": 2.74,
|
| 192 |
+
"learning_rate": 4.9885536925050395e-05,
|
| 193 |
+
"loss": 5.2037,
|
| 194 |
+
"step": 15000
|
| 195 |
+
},
|
| 196 |
+
{
|
| 197 |
+
"epoch": 2.83,
|
| 198 |
+
"learning_rate": 4.98741066520066e-05,
|
| 199 |
+
"loss": 5.196,
|
| 200 |
+
"step": 15500
|
| 201 |
+
},
|
| 202 |
+
{
|
| 203 |
+
"epoch": 2.92,
|
| 204 |
+
"learning_rate": 4.9862653472603996e-05,
|
| 205 |
+
"loss": 5.1892,
|
| 206 |
+
"step": 16000
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 3.01,
|
| 210 |
+
"learning_rate": 4.985120029320139e-05,
|
| 211 |
+
"loss": 5.1825,
|
| 212 |
+
"step": 16500
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 3.1,
|
| 216 |
+
"learning_rate": 4.98397700201576e-05,
|
| 217 |
+
"loss": 5.1753,
|
| 218 |
+
"step": 17000
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 3.19,
|
| 222 |
+
"learning_rate": 4.9828316840755e-05,
|
| 223 |
+
"loss": 5.1722,
|
| 224 |
+
"step": 17500
|
| 225 |
+
},
|
| 226 |
+
{
|
| 227 |
+
"epoch": 3.28,
|
| 228 |
+
"learning_rate": 4.9816863661352395e-05,
|
| 229 |
+
"loss": 5.1653,
|
| 230 |
+
"step": 18000
|
| 231 |
+
},
|
| 232 |
+
{
|
| 233 |
+
"epoch": 3.37,
|
| 234 |
+
"learning_rate": 4.980541048194979e-05,
|
| 235 |
+
"loss": 5.1603,
|
| 236 |
+
"step": 18500
|
| 237 |
+
},
|
| 238 |
+
{
|
| 239 |
+
"epoch": 3.47,
|
| 240 |
+
"learning_rate": 4.9793980208905996e-05,
|
| 241 |
+
"loss": 5.1547,
|
| 242 |
+
"step": 19000
|
| 243 |
+
},
|
| 244 |
+
{
|
| 245 |
+
"epoch": 3.56,
|
| 246 |
+
"learning_rate": 4.97825270295034e-05,
|
| 247 |
+
"loss": 5.151,
|
| 248 |
+
"step": 19500
|
| 249 |
+
},
|
| 250 |
+
{
|
| 251 |
+
"epoch": 3.65,
|
| 252 |
+
"learning_rate": 4.9771073850100794e-05,
|
| 253 |
+
"loss": 5.1468,
|
| 254 |
+
"step": 20000
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 3.74,
|
| 258 |
+
"learning_rate": 4.975962067069819e-05,
|
| 259 |
+
"loss": 5.1434,
|
| 260 |
+
"step": 20500
|
| 261 |
+
},
|
| 262 |
+
{
|
| 263 |
+
"epoch": 3.83,
|
| 264 |
+
"learning_rate": 4.974816749129559e-05,
|
| 265 |
+
"loss": 5.1388,
|
| 266 |
+
"step": 21000
|
| 267 |
+
},
|
| 268 |
+
{
|
| 269 |
+
"epoch": 3.92,
|
| 270 |
+
"learning_rate": 4.973671431189299e-05,
|
| 271 |
+
"loss": 5.133,
|
| 272 |
+
"step": 21500
|
| 273 |
+
},
|
| 274 |
+
{
|
| 275 |
+
"epoch": 4.01,
|
| 276 |
+
"learning_rate": 4.9725284038849186e-05,
|
| 277 |
+
"loss": 5.1264,
|
| 278 |
+
"step": 22000
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 4.1,
|
| 282 |
+
"learning_rate": 4.971383085944659e-05,
|
| 283 |
+
"loss": 5.1248,
|
| 284 |
+
"step": 22500
|
| 285 |
+
},
|
| 286 |
+
{
|
| 287 |
+
"epoch": 4.2,
|
| 288 |
+
"learning_rate": 4.9702377680043984e-05,
|
| 289 |
+
"loss": 5.1185,
|
| 290 |
+
"step": 23000
|
| 291 |
+
},
|
| 292 |
+
{
|
| 293 |
+
"epoch": 4.29,
|
| 294 |
+
"learning_rate": 4.969092450064138e-05,
|
| 295 |
+
"loss": 5.1139,
|
| 296 |
+
"step": 23500
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 4.38,
|
| 300 |
+
"learning_rate": 4.967947132123878e-05,
|
| 301 |
+
"loss": 5.1114,
|
| 302 |
+
"step": 24000
|
| 303 |
+
},
|
| 304 |
+
{
|
| 305 |
+
"epoch": 4.47,
|
| 306 |
+
"learning_rate": 4.966806395455378e-05,
|
| 307 |
+
"loss": 5.1084,
|
| 308 |
+
"step": 24500
|
| 309 |
+
},
|
| 310 |
+
{
|
| 311 |
+
"epoch": 4.56,
|
| 312 |
+
"learning_rate": 4.9656610775151186e-05,
|
| 313 |
+
"loss": 5.1041,
|
| 314 |
+
"step": 25000
|
| 315 |
+
},
|
| 316 |
+
{
|
| 317 |
+
"epoch": 4.65,
|
| 318 |
+
"learning_rate": 4.964515759574858e-05,
|
| 319 |
+
"loss": 5.102,
|
| 320 |
+
"step": 25500
|
| 321 |
+
},
|
| 322 |
+
{
|
| 323 |
+
"epoch": 4.74,
|
| 324 |
+
"learning_rate": 4.963370441634598e-05,
|
| 325 |
+
"loss": 5.1012,
|
| 326 |
+
"step": 26000
|
| 327 |
+
},
|
| 328 |
+
{
|
| 329 |
+
"epoch": 4.83,
|
| 330 |
+
"learning_rate": 4.962225123694338e-05,
|
| 331 |
+
"loss": 5.0961,
|
| 332 |
+
"step": 26500
|
| 333 |
+
},
|
| 334 |
+
{
|
| 335 |
+
"epoch": 4.93,
|
| 336 |
+
"learning_rate": 4.9610798057540775e-05,
|
| 337 |
+
"loss": 5.0918,
|
| 338 |
+
"step": 27000
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 5.02,
|
| 342 |
+
"learning_rate": 4.959934487813817e-05,
|
| 343 |
+
"loss": 5.0823,
|
| 344 |
+
"step": 27500
|
| 345 |
+
},
|
| 346 |
+
{
|
| 347 |
+
"epoch": 5.11,
|
| 348 |
+
"learning_rate": 4.958789169873557e-05,
|
| 349 |
+
"loss": 4.7898,
|
| 350 |
+
"step": 28000
|
| 351 |
+
},
|
| 352 |
+
{
|
| 353 |
+
"epoch": 5.2,
|
| 354 |
+
"learning_rate": 4.957643851933297e-05,
|
| 355 |
+
"loss": 4.47,
|
| 356 |
+
"step": 28500
|
| 357 |
+
},
|
| 358 |
+
{
|
| 359 |
+
"epoch": 5.29,
|
| 360 |
+
"learning_rate": 4.9564985339930364e-05,
|
| 361 |
+
"loss": 4.1839,
|
| 362 |
+
"step": 29000
|
| 363 |
+
},
|
| 364 |
+
{
|
| 365 |
+
"epoch": 5.38,
|
| 366 |
+
"learning_rate": 4.9553532160527766e-05,
|
| 367 |
+
"loss": 3.9283,
|
| 368 |
+
"step": 29500
|
| 369 |
+
},
|
| 370 |
+
{
|
| 371 |
+
"epoch": 5.47,
|
| 372 |
+
"learning_rate": 4.9542101887483965e-05,
|
| 373 |
+
"loss": 3.3536,
|
| 374 |
+
"step": 30000
|
| 375 |
+
},
|
| 376 |
+
{
|
| 377 |
+
"epoch": 5.56,
|
| 378 |
+
"learning_rate": 4.953067161444017e-05,
|
| 379 |
+
"loss": 2.7553,
|
| 380 |
+
"step": 30500
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 5.65,
|
| 384 |
+
"learning_rate": 4.9519218435037566e-05,
|
| 385 |
+
"loss": 2.3501,
|
| 386 |
+
"step": 31000
|
| 387 |
+
},
|
| 388 |
+
{
|
| 389 |
+
"epoch": 5.75,
|
| 390 |
+
"learning_rate": 4.950776525563497e-05,
|
| 391 |
+
"loss": 1.9139,
|
| 392 |
+
"step": 31500
|
| 393 |
+
},
|
| 394 |
+
{
|
| 395 |
+
"epoch": 5.84,
|
| 396 |
+
"learning_rate": 4.9496312076232364e-05,
|
| 397 |
+
"loss": 1.6857,
|
| 398 |
+
"step": 32000
|
| 399 |
+
},
|
| 400 |
+
{
|
| 401 |
+
"epoch": 5.93,
|
| 402 |
+
"learning_rate": 4.948485889682976e-05,
|
| 403 |
+
"loss": 1.5418,
|
| 404 |
+
"step": 32500
|
| 405 |
+
},
|
| 406 |
+
{
|
| 407 |
+
"epoch": 6.02,
|
| 408 |
+
"learning_rate": 4.947340571742716e-05,
|
| 409 |
+
"loss": 1.4448,
|
| 410 |
+
"step": 33000
|
| 411 |
+
},
|
| 412 |
+
{
|
| 413 |
+
"epoch": 6.11,
|
| 414 |
+
"learning_rate": 4.946195253802456e-05,
|
| 415 |
+
"loss": 1.3729,
|
| 416 |
+
"step": 33500
|
| 417 |
+
},
|
| 418 |
+
{
|
| 419 |
+
"epoch": 6.2,
|
| 420 |
+
"learning_rate": 4.945049935862195e-05,
|
| 421 |
+
"loss": 1.3178,
|
| 422 |
+
"step": 34000
|
| 423 |
+
},
|
| 424 |
+
{
|
| 425 |
+
"epoch": 6.29,
|
| 426 |
+
"learning_rate": 4.9439046179219356e-05,
|
| 427 |
+
"loss": 1.2557,
|
| 428 |
+
"step": 34500
|
| 429 |
+
},
|
| 430 |
+
{
|
| 431 |
+
"epoch": 6.38,
|
| 432 |
+
"learning_rate": 4.942759299981675e-05,
|
| 433 |
+
"loss": 1.1956,
|
| 434 |
+
"step": 35000
|
| 435 |
+
},
|
| 436 |
+
{
|
| 437 |
+
"epoch": 6.48,
|
| 438 |
+
"learning_rate": 4.941613982041415e-05,
|
| 439 |
+
"loss": 1.1306,
|
| 440 |
+
"step": 35500
|
| 441 |
+
},
|
| 442 |
+
{
|
| 443 |
+
"epoch": 6.57,
|
| 444 |
+
"learning_rate": 4.940470954737035e-05,
|
| 445 |
+
"loss": 1.0845,
|
| 446 |
+
"step": 36000
|
| 447 |
+
},
|
| 448 |
+
{
|
| 449 |
+
"epoch": 6.66,
|
| 450 |
+
"learning_rate": 4.939325636796775e-05,
|
| 451 |
+
"loss": 1.0483,
|
| 452 |
+
"step": 36500
|
| 453 |
+
},
|
| 454 |
+
{
|
| 455 |
+
"epoch": 6.75,
|
| 456 |
+
"learning_rate": 4.9381803188565143e-05,
|
| 457 |
+
"loss": 1.0169,
|
| 458 |
+
"step": 37000
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 6.84,
|
| 462 |
+
"learning_rate": 4.9370350009162546e-05,
|
| 463 |
+
"loss": 0.9886,
|
| 464 |
+
"step": 37500
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 6.93,
|
| 468 |
+
"learning_rate": 4.9358919736118744e-05,
|
| 469 |
+
"loss": 0.966,
|
| 470 |
+
"step": 38000
|
| 471 |
+
},
|
| 472 |
+
{
|
| 473 |
+
"epoch": 7.02,
|
| 474 |
+
"learning_rate": 4.934746655671615e-05,
|
| 475 |
+
"loss": 0.9429,
|
| 476 |
+
"step": 38500
|
| 477 |
+
},
|
| 478 |
+
{
|
| 479 |
+
"epoch": 7.11,
|
| 480 |
+
"learning_rate": 4.933601337731354e-05,
|
| 481 |
+
"loss": 0.9261,
|
| 482 |
+
"step": 39000
|
| 483 |
+
},
|
| 484 |
+
{
|
| 485 |
+
"epoch": 7.21,
|
| 486 |
+
"learning_rate": 4.932456019791094e-05,
|
| 487 |
+
"loss": 0.908,
|
| 488 |
+
"step": 39500
|
| 489 |
+
},
|
| 490 |
+
{
|
| 491 |
+
"epoch": 7.3,
|
| 492 |
+
"learning_rate": 4.9313129924867143e-05,
|
| 493 |
+
"loss": 0.8914,
|
| 494 |
+
"step": 40000
|
| 495 |
+
},
|
| 496 |
+
{
|
| 497 |
+
"epoch": 7.39,
|
| 498 |
+
"learning_rate": 4.930167674546454e-05,
|
| 499 |
+
"loss": 0.8789,
|
| 500 |
+
"step": 40500
|
| 501 |
+
},
|
| 502 |
+
{
|
| 503 |
+
"epoch": 7.48,
|
| 504 |
+
"learning_rate": 4.929022356606194e-05,
|
| 505 |
+
"loss": 0.8643,
|
| 506 |
+
"step": 41000
|
| 507 |
+
},
|
| 508 |
+
{
|
| 509 |
+
"epoch": 7.57,
|
| 510 |
+
"learning_rate": 4.927877038665934e-05,
|
| 511 |
+
"loss": 0.8531,
|
| 512 |
+
"step": 41500
|
| 513 |
+
},
|
| 514 |
+
{
|
| 515 |
+
"epoch": 7.66,
|
| 516 |
+
"learning_rate": 4.926731720725673e-05,
|
| 517 |
+
"loss": 0.8418,
|
| 518 |
+
"step": 42000
|
| 519 |
+
},
|
| 520 |
+
{
|
| 521 |
+
"epoch": 7.75,
|
| 522 |
+
"learning_rate": 4.925588693421294e-05,
|
| 523 |
+
"loss": 0.83,
|
| 524 |
+
"step": 42500
|
| 525 |
+
},
|
| 526 |
+
{
|
| 527 |
+
"epoch": 7.84,
|
| 528 |
+
"learning_rate": 4.924443375481034e-05,
|
| 529 |
+
"loss": 0.8201,
|
| 530 |
+
"step": 43000
|
| 531 |
+
},
|
| 532 |
+
{
|
| 533 |
+
"epoch": 7.94,
|
| 534 |
+
"learning_rate": 4.9232980575407736e-05,
|
| 535 |
+
"loss": 0.8115,
|
| 536 |
+
"step": 43500
|
| 537 |
+
},
|
| 538 |
+
{
|
| 539 |
+
"epoch": 8.03,
|
| 540 |
+
"learning_rate": 4.922152739600513e-05,
|
| 541 |
+
"loss": 0.803,
|
| 542 |
+
"step": 44000
|
| 543 |
+
},
|
| 544 |
+
{
|
| 545 |
+
"epoch": 8.12,
|
| 546 |
+
"learning_rate": 4.9210074216602534e-05,
|
| 547 |
+
"loss": 0.7935,
|
| 548 |
+
"step": 44500
|
| 549 |
+
},
|
| 550 |
+
{
|
| 551 |
+
"epoch": 8.21,
|
| 552 |
+
"learning_rate": 4.919862103719993e-05,
|
| 553 |
+
"loss": 0.7857,
|
| 554 |
+
"step": 45000
|
| 555 |
+
},
|
| 556 |
+
{
|
| 557 |
+
"epoch": 8.3,
|
| 558 |
+
"learning_rate": 4.9187190764156135e-05,
|
| 559 |
+
"loss": 0.7753,
|
| 560 |
+
"step": 45500
|
| 561 |
+
},
|
| 562 |
+
{
|
| 563 |
+
"epoch": 8.39,
|
| 564 |
+
"learning_rate": 4.917573758475353e-05,
|
| 565 |
+
"loss": 0.7683,
|
| 566 |
+
"step": 46000
|
| 567 |
+
},
|
| 568 |
+
{
|
| 569 |
+
"epoch": 8.48,
|
| 570 |
+
"learning_rate": 4.916428440535093e-05,
|
| 571 |
+
"loss": 0.761,
|
| 572 |
+
"step": 46500
|
| 573 |
+
},
|
| 574 |
+
{
|
| 575 |
+
"epoch": 8.57,
|
| 576 |
+
"learning_rate": 4.915283122594833e-05,
|
| 577 |
+
"loss": 0.7549,
|
| 578 |
+
"step": 47000
|
| 579 |
+
},
|
| 580 |
+
{
|
| 581 |
+
"epoch": 8.66,
|
| 582 |
+
"learning_rate": 4.914140095290453e-05,
|
| 583 |
+
"loss": 0.7478,
|
| 584 |
+
"step": 47500
|
| 585 |
+
},
|
| 586 |
+
{
|
| 587 |
+
"epoch": 8.76,
|
| 588 |
+
"learning_rate": 4.912994777350193e-05,
|
| 589 |
+
"loss": 0.7404,
|
| 590 |
+
"step": 48000
|
| 591 |
+
},
|
| 592 |
+
{
|
| 593 |
+
"epoch": 8.85,
|
| 594 |
+
"learning_rate": 4.9118494594099325e-05,
|
| 595 |
+
"loss": 0.7335,
|
| 596 |
+
"step": 48500
|
| 597 |
+
},
|
| 598 |
+
{
|
| 599 |
+
"epoch": 8.94,
|
| 600 |
+
"learning_rate": 4.910704141469672e-05,
|
| 601 |
+
"loss": 0.7283,
|
| 602 |
+
"step": 49000
|
| 603 |
+
},
|
| 604 |
+
{
|
| 605 |
+
"epoch": 9.03,
|
| 606 |
+
"learning_rate": 4.9095611141652926e-05,
|
| 607 |
+
"loss": 0.7231,
|
| 608 |
+
"step": 49500
|
| 609 |
+
},
|
| 610 |
+
{
|
| 611 |
+
"epoch": 9.12,
|
| 612 |
+
"learning_rate": 4.908415796225032e-05,
|
| 613 |
+
"loss": 0.7156,
|
| 614 |
+
"step": 50000
|
| 615 |
+
},
|
| 616 |
+
{
|
| 617 |
+
"epoch": 9.21,
|
| 618 |
+
"learning_rate": 4.9072704782847724e-05,
|
| 619 |
+
"loss": 0.7099,
|
| 620 |
+
"step": 50500
|
| 621 |
+
},
|
| 622 |
+
{
|
| 623 |
+
"epoch": 9.3,
|
| 624 |
+
"learning_rate": 4.906125160344512e-05,
|
| 625 |
+
"loss": 0.7071,
|
| 626 |
+
"step": 51000
|
| 627 |
+
},
|
| 628 |
+
{
|
| 629 |
+
"epoch": 9.39,
|
| 630 |
+
"learning_rate": 4.9049798424042515e-05,
|
| 631 |
+
"loss": 0.702,
|
| 632 |
+
"step": 51500
|
| 633 |
+
},
|
| 634 |
+
{
|
| 635 |
+
"epoch": 9.49,
|
| 636 |
+
"learning_rate": 4.903836815099872e-05,
|
| 637 |
+
"loss": 0.6982,
|
| 638 |
+
"step": 52000
|
| 639 |
+
},
|
| 640 |
+
{
|
| 641 |
+
"epoch": 9.58,
|
| 642 |
+
"learning_rate": 4.9026914971596116e-05,
|
| 643 |
+
"loss": 0.6933,
|
| 644 |
+
"step": 52500
|
| 645 |
+
},
|
| 646 |
+
{
|
| 647 |
+
"epoch": 9.67,
|
| 648 |
+
"learning_rate": 4.901546179219352e-05,
|
| 649 |
+
"loss": 0.6901,
|
| 650 |
+
"step": 53000
|
| 651 |
+
},
|
| 652 |
+
{
|
| 653 |
+
"epoch": 9.76,
|
| 654 |
+
"learning_rate": 4.9004008612790914e-05,
|
| 655 |
+
"loss": 0.6854,
|
| 656 |
+
"step": 53500
|
| 657 |
+
},
|
| 658 |
+
{
|
| 659 |
+
"epoch": 9.85,
|
| 660 |
+
"learning_rate": 4.899255543338831e-05,
|
| 661 |
+
"loss": 0.6813,
|
| 662 |
+
"step": 54000
|
| 663 |
+
},
|
| 664 |
+
{
|
| 665 |
+
"epoch": 9.94,
|
| 666 |
+
"learning_rate": 4.8981125160344515e-05,
|
| 667 |
+
"loss": 0.6762,
|
| 668 |
+
"step": 54500
|
| 669 |
+
},
|
| 670 |
+
{
|
| 671 |
+
"epoch": 10.03,
|
| 672 |
+
"learning_rate": 4.896967198094191e-05,
|
| 673 |
+
"loss": 0.6708,
|
| 674 |
+
"step": 55000
|
| 675 |
+
},
|
| 676 |
+
{
|
| 677 |
+
"epoch": 10.12,
|
| 678 |
+
"learning_rate": 4.8958218801539307e-05,
|
| 679 |
+
"loss": 0.6687,
|
| 680 |
+
"step": 55500
|
| 681 |
+
},
|
| 682 |
+
{
|
| 683 |
+
"epoch": 10.22,
|
| 684 |
+
"learning_rate": 4.894676562213671e-05,
|
| 685 |
+
"loss": 0.6641,
|
| 686 |
+
"step": 56000
|
| 687 |
+
},
|
| 688 |
+
{
|
| 689 |
+
"epoch": 10.31,
|
| 690 |
+
"learning_rate": 4.8935312442734105e-05,
|
| 691 |
+
"loss": 0.6594,
|
| 692 |
+
"step": 56500
|
| 693 |
+
},
|
| 694 |
+
{
|
| 695 |
+
"epoch": 10.4,
|
| 696 |
+
"learning_rate": 4.892388216969031e-05,
|
| 697 |
+
"loss": 0.6578,
|
| 698 |
+
"step": 57000
|
| 699 |
+
},
|
| 700 |
+
{
|
| 701 |
+
"epoch": 10.49,
|
| 702 |
+
"learning_rate": 4.8912428990287706e-05,
|
| 703 |
+
"loss": 0.6542,
|
| 704 |
+
"step": 57500
|
| 705 |
+
},
|
| 706 |
+
{
|
| 707 |
+
"epoch": 10.58,
|
| 708 |
+
"learning_rate": 4.89009758108851e-05,
|
| 709 |
+
"loss": 0.6509,
|
| 710 |
+
"step": 58000
|
| 711 |
+
},
|
| 712 |
+
{
|
| 713 |
+
"epoch": 10.67,
|
| 714 |
+
"learning_rate": 4.8889522631482504e-05,
|
| 715 |
+
"loss": 0.6471,
|
| 716 |
+
"step": 58500
|
| 717 |
+
},
|
| 718 |
+
{
|
| 719 |
+
"epoch": 10.76,
|
| 720 |
+
"learning_rate": 4.88780694520799e-05,
|
| 721 |
+
"loss": 0.6463,
|
| 722 |
+
"step": 59000
|
| 723 |
+
},
|
| 724 |
+
{
|
| 725 |
+
"epoch": 10.85,
|
| 726 |
+
"learning_rate": 4.8866616272677295e-05,
|
| 727 |
+
"loss": 0.6418,
|
| 728 |
+
"step": 59500
|
| 729 |
+
},
|
| 730 |
+
{
|
| 731 |
+
"epoch": 10.94,
|
| 732 |
+
"learning_rate": 4.88551859996335e-05,
|
| 733 |
+
"loss": 0.6391,
|
| 734 |
+
"step": 60000
|
| 735 |
+
},
|
| 736 |
+
{
|
| 737 |
+
"epoch": 11.04,
|
| 738 |
+
"learning_rate": 4.8843732820230896e-05,
|
| 739 |
+
"loss": 0.6354,
|
| 740 |
+
"step": 60500
|
| 741 |
+
},
|
| 742 |
+
{
|
| 743 |
+
"epoch": 11.13,
|
| 744 |
+
"learning_rate": 4.88322796408283e-05,
|
| 745 |
+
"loss": 0.6327,
|
| 746 |
+
"step": 61000
|
| 747 |
+
},
|
| 748 |
+
{
|
| 749 |
+
"epoch": 11.22,
|
| 750 |
+
"learning_rate": 4.8820826461425694e-05,
|
| 751 |
+
"loss": 0.6292,
|
| 752 |
+
"step": 61500
|
| 753 |
+
},
|
| 754 |
+
{
|
| 755 |
+
"epoch": 11.31,
|
| 756 |
+
"learning_rate": 4.88094190947407e-05,
|
| 757 |
+
"loss": 0.6258,
|
| 758 |
+
"step": 62000
|
| 759 |
+
},
|
| 760 |
+
{
|
| 761 |
+
"epoch": 11.4,
|
| 762 |
+
"learning_rate": 4.879798882169691e-05,
|
| 763 |
+
"loss": 0.6257,
|
| 764 |
+
"step": 62500
|
| 765 |
+
},
|
| 766 |
+
{
|
| 767 |
+
"epoch": 11.49,
|
| 768 |
+
"learning_rate": 4.87865356422943e-05,
|
| 769 |
+
"loss": 0.6221,
|
| 770 |
+
"step": 63000
|
| 771 |
+
},
|
| 772 |
+
{
|
| 773 |
+
"epoch": 11.58,
|
| 774 |
+
"learning_rate": 4.8775082462891706e-05,
|
| 775 |
+
"loss": 0.618,
|
| 776 |
+
"step": 63500
|
| 777 |
+
},
|
| 778 |
+
{
|
| 779 |
+
"epoch": 11.67,
|
| 780 |
+
"learning_rate": 4.87636292834891e-05,
|
| 781 |
+
"loss": 0.6156,
|
| 782 |
+
"step": 64000
|
| 783 |
+
},
|
| 784 |
+
{
|
| 785 |
+
"epoch": 11.77,
|
| 786 |
+
"learning_rate": 4.87521761040865e-05,
|
| 787 |
+
"loss": 0.614,
|
| 788 |
+
"step": 64500
|
| 789 |
+
},
|
| 790 |
+
{
|
| 791 |
+
"epoch": 11.86,
|
| 792 |
+
"learning_rate": 4.87407229246839e-05,
|
| 793 |
+
"loss": 0.612,
|
| 794 |
+
"step": 65000
|
| 795 |
+
},
|
| 796 |
+
{
|
| 797 |
+
"epoch": 11.95,
|
| 798 |
+
"learning_rate": 4.8729269745281295e-05,
|
| 799 |
+
"loss": 0.6096,
|
| 800 |
+
"step": 65500
|
| 801 |
+
},
|
| 802 |
+
{
|
| 803 |
+
"epoch": 12.04,
|
| 804 |
+
"learning_rate": 4.871781656587869e-05,
|
| 805 |
+
"loss": 0.6073,
|
| 806 |
+
"step": 66000
|
| 807 |
+
},
|
| 808 |
+
{
|
| 809 |
+
"epoch": 12.13,
|
| 810 |
+
"learning_rate": 4.8706386292834896e-05,
|
| 811 |
+
"loss": 0.6039,
|
| 812 |
+
"step": 66500
|
| 813 |
+
},
|
| 814 |
+
{
|
| 815 |
+
"epoch": 12.22,
|
| 816 |
+
"learning_rate": 4.869493311343229e-05,
|
| 817 |
+
"loss": 0.6033,
|
| 818 |
+
"step": 67000
|
| 819 |
+
},
|
| 820 |
+
{
|
| 821 |
+
"epoch": 12.31,
|
| 822 |
+
"learning_rate": 4.8683479934029694e-05,
|
| 823 |
+
"loss": 0.6005,
|
| 824 |
+
"step": 67500
|
| 825 |
+
},
|
| 826 |
+
{
|
| 827 |
+
"epoch": 12.4,
|
| 828 |
+
"learning_rate": 4.867202675462709e-05,
|
| 829 |
+
"loss": 0.5971,
|
| 830 |
+
"step": 68000
|
| 831 |
+
},
|
| 832 |
+
{
|
| 833 |
+
"epoch": 12.5,
|
| 834 |
+
"learning_rate": 4.866059648158329e-05,
|
| 835 |
+
"loss": 0.5933,
|
| 836 |
+
"step": 68500
|
| 837 |
+
},
|
| 838 |
+
{
|
| 839 |
+
"epoch": 12.59,
|
| 840 |
+
"learning_rate": 4.864914330218069e-05,
|
| 841 |
+
"loss": 0.5947,
|
| 842 |
+
"step": 69000
|
| 843 |
+
},
|
| 844 |
+
{
|
| 845 |
+
"epoch": 12.68,
|
| 846 |
+
"learning_rate": 4.8637690122778086e-05,
|
| 847 |
+
"loss": 0.5928,
|
| 848 |
+
"step": 69500
|
| 849 |
+
},
|
| 850 |
+
{
|
| 851 |
+
"epoch": 12.77,
|
| 852 |
+
"learning_rate": 4.862623694337549e-05,
|
| 853 |
+
"loss": 0.5897,
|
| 854 |
+
"step": 70000
|
| 855 |
+
},
|
| 856 |
+
{
|
| 857 |
+
"epoch": 12.86,
|
| 858 |
+
"learning_rate": 4.8614783763972884e-05,
|
| 859 |
+
"loss": 0.588,
|
| 860 |
+
"step": 70500
|
| 861 |
+
},
|
| 862 |
+
{
|
| 863 |
+
"epoch": 12.95,
|
| 864 |
+
"learning_rate": 4.860333058457028e-05,
|
| 865 |
+
"loss": 0.5862,
|
| 866 |
+
"step": 71000
|
| 867 |
+
},
|
| 868 |
+
{
|
| 869 |
+
"epoch": 13.04,
|
| 870 |
+
"learning_rate": 4.859187740516768e-05,
|
| 871 |
+
"loss": 0.5835,
|
| 872 |
+
"step": 71500
|
| 873 |
+
},
|
| 874 |
+
{
|
| 875 |
+
"epoch": 13.13,
|
| 876 |
+
"learning_rate": 4.858042422576508e-05,
|
| 877 |
+
"loss": 0.5827,
|
| 878 |
+
"step": 72000
|
| 879 |
+
},
|
| 880 |
+
{
|
| 881 |
+
"epoch": 13.23,
|
| 882 |
+
"learning_rate": 4.8569016859080086e-05,
|
| 883 |
+
"loss": 0.581,
|
| 884 |
+
"step": 72500
|
| 885 |
+
},
|
| 886 |
+
{
|
| 887 |
+
"epoch": 13.32,
|
| 888 |
+
"learning_rate": 4.855756367967748e-05,
|
| 889 |
+
"loss": 0.5791,
|
| 890 |
+
"step": 73000
|
| 891 |
+
},
|
| 892 |
+
{
|
| 893 |
+
"epoch": 13.41,
|
| 894 |
+
"learning_rate": 4.854611050027488e-05,
|
| 895 |
+
"loss": 0.5764,
|
| 896 |
+
"step": 73500
|
| 897 |
+
},
|
| 898 |
+
{
|
| 899 |
+
"epoch": 13.5,
|
| 900 |
+
"learning_rate": 4.853465732087228e-05,
|
| 901 |
+
"loss": 0.5749,
|
| 902 |
+
"step": 74000
|
| 903 |
+
},
|
| 904 |
+
{
|
| 905 |
+
"epoch": 13.59,
|
| 906 |
+
"learning_rate": 4.8523204141469675e-05,
|
| 907 |
+
"loss": 0.5747,
|
| 908 |
+
"step": 74500
|
| 909 |
+
},
|
| 910 |
+
{
|
| 911 |
+
"epoch": 13.68,
|
| 912 |
+
"learning_rate": 4.8511773868425874e-05,
|
| 913 |
+
"loss": 0.5717,
|
| 914 |
+
"step": 75000
|
| 915 |
+
},
|
| 916 |
+
{
|
| 917 |
+
"epoch": 13.77,
|
| 918 |
+
"learning_rate": 4.8500320689023276e-05,
|
| 919 |
+
"loss": 0.5716,
|
| 920 |
+
"step": 75500
|
| 921 |
+
},
|
| 922 |
+
{
|
| 923 |
+
"epoch": 13.86,
|
| 924 |
+
"learning_rate": 4.848886750962067e-05,
|
| 925 |
+
"loss": 0.5691,
|
| 926 |
+
"step": 76000
|
| 927 |
+
},
|
| 928 |
+
{
|
| 929 |
+
"epoch": 13.95,
|
| 930 |
+
"learning_rate": 4.847741433021807e-05,
|
| 931 |
+
"loss": 0.5661,
|
| 932 |
+
"step": 76500
|
| 933 |
+
},
|
| 934 |
+
{
|
| 935 |
+
"epoch": 14.05,
|
| 936 |
+
"learning_rate": 4.846596115081547e-05,
|
| 937 |
+
"loss": 0.565,
|
| 938 |
+
"step": 77000
|
| 939 |
+
},
|
| 940 |
+
{
|
| 941 |
+
"epoch": 14.14,
|
| 942 |
+
"learning_rate": 4.845453087777167e-05,
|
| 943 |
+
"loss": 0.5627,
|
| 944 |
+
"step": 77500
|
| 945 |
+
},
|
| 946 |
+
{
|
| 947 |
+
"epoch": 14.23,
|
| 948 |
+
"learning_rate": 4.844307769836907e-05,
|
| 949 |
+
"loss": 0.563,
|
| 950 |
+
"step": 78000
|
| 951 |
+
},
|
| 952 |
+
{
|
| 953 |
+
"epoch": 14.32,
|
| 954 |
+
"learning_rate": 4.8431624518966466e-05,
|
| 955 |
+
"loss": 0.561,
|
| 956 |
+
"step": 78500
|
| 957 |
+
},
|
| 958 |
+
{
|
| 959 |
+
"epoch": 14.41,
|
| 960 |
+
"learning_rate": 4.842017133956386e-05,
|
| 961 |
+
"loss": 0.5599,
|
| 962 |
+
"step": 79000
|
| 963 |
+
},
|
| 964 |
+
{
|
| 965 |
+
"epoch": 14.5,
|
| 966 |
+
"learning_rate": 4.8408718160161264e-05,
|
| 967 |
+
"loss": 0.5573,
|
| 968 |
+
"step": 79500
|
| 969 |
+
},
|
| 970 |
+
{
|
| 971 |
+
"epoch": 14.59,
|
| 972 |
+
"learning_rate": 4.839726498075866e-05,
|
| 973 |
+
"loss": 0.5558,
|
| 974 |
+
"step": 80000
|
| 975 |
+
},
|
| 976 |
+
{
|
| 977 |
+
"epoch": 14.68,
|
| 978 |
+
"learning_rate": 4.8385834707714865e-05,
|
| 979 |
+
"loss": 0.5559,
|
| 980 |
+
"step": 80500
|
| 981 |
+
},
|
| 982 |
+
{
|
| 983 |
+
"epoch": 14.78,
|
| 984 |
+
"learning_rate": 4.837438152831226e-05,
|
| 985 |
+
"loss": 0.5546,
|
| 986 |
+
"step": 81000
|
| 987 |
+
},
|
| 988 |
+
{
|
| 989 |
+
"epoch": 14.87,
|
| 990 |
+
"learning_rate": 4.8362928348909656e-05,
|
| 991 |
+
"loss": 0.5531,
|
| 992 |
+
"step": 81500
|
| 993 |
+
},
|
| 994 |
+
{
|
| 995 |
+
"epoch": 14.96,
|
| 996 |
+
"learning_rate": 4.835147516950706e-05,
|
| 997 |
+
"loss": 0.5493,
|
| 998 |
+
"step": 82000
|
| 999 |
+
},
|
| 1000 |
+
{
|
| 1001 |
+
"epoch": 15.05,
|
| 1002 |
+
"learning_rate": 4.8340021990104454e-05,
|
| 1003 |
+
"loss": 0.5483,
|
| 1004 |
+
"step": 82500
|
| 1005 |
+
},
|
| 1006 |
+
{
|
| 1007 |
+
"epoch": 15.14,
|
| 1008 |
+
"learning_rate": 4.832859171706065e-05,
|
| 1009 |
+
"loss": 0.548,
|
| 1010 |
+
"step": 83000
|
| 1011 |
+
},
|
| 1012 |
+
{
|
| 1013 |
+
"epoch": 15.23,
|
| 1014 |
+
"learning_rate": 4.8317138537658055e-05,
|
| 1015 |
+
"loss": 0.5462,
|
| 1016 |
+
"step": 83500
|
| 1017 |
+
},
|
| 1018 |
+
{
|
| 1019 |
+
"epoch": 15.32,
|
| 1020 |
+
"learning_rate": 4.830568535825545e-05,
|
| 1021 |
+
"loss": 0.5454,
|
| 1022 |
+
"step": 84000
|
| 1023 |
+
},
|
| 1024 |
+
{
|
| 1025 |
+
"epoch": 15.41,
|
| 1026 |
+
"learning_rate": 4.829423217885285e-05,
|
| 1027 |
+
"loss": 0.5428,
|
| 1028 |
+
"step": 84500
|
| 1029 |
+
},
|
| 1030 |
+
{
|
| 1031 |
+
"epoch": 15.51,
|
| 1032 |
+
"learning_rate": 4.828277899945025e-05,
|
| 1033 |
+
"loss": 0.5428,
|
| 1034 |
+
"step": 85000
|
| 1035 |
+
},
|
| 1036 |
+
{
|
| 1037 |
+
"epoch": 15.6,
|
| 1038 |
+
"learning_rate": 4.8271325820047645e-05,
|
| 1039 |
+
"loss": 0.5424,
|
| 1040 |
+
"step": 85500
|
| 1041 |
+
},
|
| 1042 |
+
{
|
| 1043 |
+
"epoch": 15.69,
|
| 1044 |
+
"learning_rate": 4.825987264064504e-05,
|
| 1045 |
+
"loss": 0.5421,
|
| 1046 |
+
"step": 86000
|
| 1047 |
+
},
|
| 1048 |
+
{
|
| 1049 |
+
"epoch": 15.78,
|
| 1050 |
+
"learning_rate": 4.824841946124244e-05,
|
| 1051 |
+
"loss": 0.5382,
|
| 1052 |
+
"step": 86500
|
| 1053 |
+
},
|
| 1054 |
+
{
|
| 1055 |
+
"epoch": 15.87,
|
| 1056 |
+
"learning_rate": 4.823698918819865e-05,
|
| 1057 |
+
"loss": 0.5373,
|
| 1058 |
+
"step": 87000
|
| 1059 |
+
},
|
| 1060 |
+
{
|
| 1061 |
+
"epoch": 15.96,
|
| 1062 |
+
"learning_rate": 4.8225536008796044e-05,
|
| 1063 |
+
"loss": 0.5359,
|
| 1064 |
+
"step": 87500
|
| 1065 |
+
},
|
| 1066 |
+
{
|
| 1067 |
+
"epoch": 16.05,
|
| 1068 |
+
"learning_rate": 4.821408282939344e-05,
|
| 1069 |
+
"loss": 0.5357,
|
| 1070 |
+
"step": 88000
|
| 1071 |
+
},
|
| 1072 |
+
{
|
| 1073 |
+
"epoch": 16.14,
|
| 1074 |
+
"learning_rate": 4.8202629649990835e-05,
|
| 1075 |
+
"loss": 0.5351,
|
| 1076 |
+
"step": 88500
|
| 1077 |
+
},
|
| 1078 |
+
{
|
| 1079 |
+
"epoch": 16.23,
|
| 1080 |
+
"learning_rate": 4.819117647058824e-05,
|
| 1081 |
+
"loss": 0.5317,
|
| 1082 |
+
"step": 89000
|
| 1083 |
+
},
|
| 1084 |
+
{
|
| 1085 |
+
"epoch": 16.33,
|
| 1086 |
+
"learning_rate": 4.817972329118563e-05,
|
| 1087 |
+
"loss": 0.5333,
|
| 1088 |
+
"step": 89500
|
| 1089 |
+
},
|
| 1090 |
+
{
|
| 1091 |
+
"epoch": 16.42,
|
| 1092 |
+
"learning_rate": 4.816829301814184e-05,
|
| 1093 |
+
"loss": 0.5309,
|
| 1094 |
+
"step": 90000
|
| 1095 |
+
},
|
| 1096 |
+
{
|
| 1097 |
+
"epoch": 16.51,
|
| 1098 |
+
"learning_rate": 4.815683983873924e-05,
|
| 1099 |
+
"loss": 0.5308,
|
| 1100 |
+
"step": 90500
|
| 1101 |
+
},
|
| 1102 |
+
{
|
| 1103 |
+
"epoch": 16.6,
|
| 1104 |
+
"learning_rate": 4.8145386659336636e-05,
|
| 1105 |
+
"loss": 0.5289,
|
| 1106 |
+
"step": 91000
|
| 1107 |
+
},
|
| 1108 |
+
{
|
| 1109 |
+
"epoch": 16.69,
|
| 1110 |
+
"learning_rate": 4.813393347993403e-05,
|
| 1111 |
+
"loss": 0.5275,
|
| 1112 |
+
"step": 91500
|
| 1113 |
+
},
|
| 1114 |
+
{
|
| 1115 |
+
"epoch": 16.78,
|
| 1116 |
+
"learning_rate": 4.812248030053143e-05,
|
| 1117 |
+
"loss": 0.5269,
|
| 1118 |
+
"step": 92000
|
| 1119 |
+
},
|
| 1120 |
+
{
|
| 1121 |
+
"epoch": 16.87,
|
| 1122 |
+
"learning_rate": 4.811102712112883e-05,
|
| 1123 |
+
"loss": 0.5248,
|
| 1124 |
+
"step": 92500
|
| 1125 |
+
},
|
| 1126 |
+
{
|
| 1127 |
+
"epoch": 16.96,
|
| 1128 |
+
"learning_rate": 4.8099573941726225e-05,
|
| 1129 |
+
"loss": 0.5242,
|
| 1130 |
+
"step": 93000
|
| 1131 |
+
},
|
| 1132 |
+
{
|
| 1133 |
+
"epoch": 17.06,
|
| 1134 |
+
"learning_rate": 4.808814366868243e-05,
|
| 1135 |
+
"loss": 0.5238,
|
| 1136 |
+
"step": 93500
|
| 1137 |
+
},
|
| 1138 |
+
{
|
| 1139 |
+
"epoch": 17.15,
|
| 1140 |
+
"learning_rate": 4.8076690489279826e-05,
|
| 1141 |
+
"loss": 0.5239,
|
| 1142 |
+
"step": 94000
|
| 1143 |
+
},
|
| 1144 |
+
{
|
| 1145 |
+
"epoch": 17.24,
|
| 1146 |
+
"learning_rate": 4.806523730987723e-05,
|
| 1147 |
+
"loss": 0.5221,
|
| 1148 |
+
"step": 94500
|
| 1149 |
+
},
|
| 1150 |
+
{
|
| 1151 |
+
"epoch": 17.33,
|
| 1152 |
+
"learning_rate": 4.8053784130474624e-05,
|
| 1153 |
+
"loss": 0.52,
|
| 1154 |
+
"step": 95000
|
| 1155 |
+
},
|
| 1156 |
+
{
|
| 1157 |
+
"epoch": 17.42,
|
| 1158 |
+
"learning_rate": 4.804233095107202e-05,
|
| 1159 |
+
"loss": 0.5184,
|
| 1160 |
+
"step": 95500
|
| 1161 |
+
},
|
| 1162 |
+
{
|
| 1163 |
+
"epoch": 17.51,
|
| 1164 |
+
"learning_rate": 4.8030900678028225e-05,
|
| 1165 |
+
"loss": 0.5186,
|
| 1166 |
+
"step": 96000
|
| 1167 |
+
},
|
| 1168 |
+
{
|
| 1169 |
+
"epoch": 17.6,
|
| 1170 |
+
"learning_rate": 4.801944749862562e-05,
|
| 1171 |
+
"loss": 0.5176,
|
| 1172 |
+
"step": 96500
|
| 1173 |
+
},
|
| 1174 |
+
{
|
| 1175 |
+
"epoch": 17.69,
|
| 1176 |
+
"learning_rate": 4.8008017225581826e-05,
|
| 1177 |
+
"loss": 0.5182,
|
| 1178 |
+
"step": 97000
|
| 1179 |
+
},
|
| 1180 |
+
{
|
| 1181 |
+
"epoch": 17.79,
|
| 1182 |
+
"learning_rate": 4.799656404617922e-05,
|
| 1183 |
+
"loss": 0.5148,
|
| 1184 |
+
"step": 97500
|
| 1185 |
+
},
|
| 1186 |
+
{
|
| 1187 |
+
"epoch": 17.88,
|
| 1188 |
+
"learning_rate": 4.798511086677662e-05,
|
| 1189 |
+
"loss": 0.5157,
|
| 1190 |
+
"step": 98000
|
| 1191 |
+
},
|
| 1192 |
+
{
|
| 1193 |
+
"epoch": 17.97,
|
| 1194 |
+
"learning_rate": 4.797365768737402e-05,
|
| 1195 |
+
"loss": 0.5131,
|
| 1196 |
+
"step": 98500
|
| 1197 |
+
}
|
| 1198 |
+
],
|
| 1199 |
+
"max_steps": 2192800,
|
| 1200 |
+
"num_train_epochs": 400,
|
| 1201 |
+
"total_flos": 2.659426878192668e+19,
|
| 1202 |
+
"trial_name": null,
|
| 1203 |
+
"trial_params": null
|
| 1204 |
+
}
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:79a718c3a61e81a142391d34ecf663b156f8aa2893666ca2e09131bbe374b4b5
|
| 3 |
+
size 3503
|
vocab.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|