Spaces:
Running
Running
Commit
·
c5ec9df
1
Parent(s):
c8084a3
Upload 9 files
Browse files- config.json +42 -0
- optimizer.pt +3 -0
- preprocessor_config.json +0 -0
- pytorch_model.bin +3 -0
- rng_state.pth +3 -0
- scaler.pt +3 -0
- scheduler.pt +3 -0
- trainer_state.json +1012 -0
- training_args.bin +3 -0
config.json
ADDED
|
@@ -0,0 +1,42 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "openai/whisper-small",
|
| 3 |
+
"activation_dropout": 0.0,
|
| 4 |
+
"activation_function": "gelu",
|
| 5 |
+
"architectures": [
|
| 6 |
+
"WhisperForConditionalGeneration"
|
| 7 |
+
],
|
| 8 |
+
"attention_dropout": 0.0,
|
| 9 |
+
"begin_suppress_tokens": [
|
| 10 |
+
220,
|
| 11 |
+
50257
|
| 12 |
+
],
|
| 13 |
+
"bos_token_id": 50257,
|
| 14 |
+
"d_model": 768,
|
| 15 |
+
"decoder_attention_heads": 12,
|
| 16 |
+
"decoder_ffn_dim": 3072,
|
| 17 |
+
"decoder_layerdrop": 0.0,
|
| 18 |
+
"decoder_layers": 12,
|
| 19 |
+
"decoder_start_token_id": 50258,
|
| 20 |
+
"dropout": 0.0,
|
| 21 |
+
"encoder_attention_heads": 12,
|
| 22 |
+
"encoder_ffn_dim": 3072,
|
| 23 |
+
"encoder_layerdrop": 0.0,
|
| 24 |
+
"encoder_layers": 12,
|
| 25 |
+
"eos_token_id": 50257,
|
| 26 |
+
"forced_decoder_ids": null,
|
| 27 |
+
"init_std": 0.02,
|
| 28 |
+
"is_encoder_decoder": true,
|
| 29 |
+
"max_length": 448,
|
| 30 |
+
"max_source_positions": 1500,
|
| 31 |
+
"max_target_positions": 448,
|
| 32 |
+
"model_type": "whisper",
|
| 33 |
+
"num_hidden_layers": 12,
|
| 34 |
+
"num_mel_bins": 80,
|
| 35 |
+
"pad_token_id": 50257,
|
| 36 |
+
"scale_embedding": false,
|
| 37 |
+
"suppress_tokens": [],
|
| 38 |
+
"torch_dtype": "float32",
|
| 39 |
+
"transformers_version": "4.26.0.dev0",
|
| 40 |
+
"use_cache": true,
|
| 41 |
+
"vocab_size": 51865
|
| 42 |
+
}
|
optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ba9e66c3bf11ee20707d5e9ba25d794c6eae2d6b54690aa7438885b6ed93a601
|
| 3 |
+
size 1934158597
|
preprocessor_config.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
pytorch_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:64544dbbec674b009f24773faa398739897cd558fed26171d62b425810e79179
|
| 3 |
+
size 967099139
|
rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:0a8f1284ab010089d385f786837927af307c3cda0e8170704892449cf9a8ac87
|
| 3 |
+
size 14503
|
scaler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:b286cfb397761804bdb2e957fb89d97d1a18bca02d5d00ad87ad709238a67589
|
| 3 |
+
size 559
|
scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:358d3cc5378fafc026eb4d987f7a358d0f6d6bfc9501f710e91d8952b2fb645b
|
| 3 |
+
size 623
|
trainer_state.json
ADDED
|
@@ -0,0 +1,1012 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 104.12321600170837,
|
| 3 |
+
"best_model_checkpoint": "./whisper-small-ar/checkpoint-4000",
|
| 4 |
+
"epoch": 5.174644243208279,
|
| 5 |
+
"global_step": 4000,
|
| 6 |
+
"is_hyper_param_search": false,
|
| 7 |
+
"is_local_process_zero": true,
|
| 8 |
+
"is_world_process_zero": true,
|
| 9 |
+
"log_history": [
|
| 10 |
+
{
|
| 11 |
+
"epoch": 0.03,
|
| 12 |
+
"learning_rate": 4.2000000000000006e-07,
|
| 13 |
+
"loss": 6.7116,
|
| 14 |
+
"step": 25
|
| 15 |
+
},
|
| 16 |
+
{
|
| 17 |
+
"epoch": 0.06,
|
| 18 |
+
"learning_rate": 9.200000000000001e-07,
|
| 19 |
+
"loss": 5.6008,
|
| 20 |
+
"step": 50
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"epoch": 0.1,
|
| 24 |
+
"learning_rate": 1.42e-06,
|
| 25 |
+
"loss": 4.4409,
|
| 26 |
+
"step": 75
|
| 27 |
+
},
|
| 28 |
+
{
|
| 29 |
+
"epoch": 0.13,
|
| 30 |
+
"learning_rate": 1.9200000000000003e-06,
|
| 31 |
+
"loss": 3.7353,
|
| 32 |
+
"step": 100
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"epoch": 0.16,
|
| 36 |
+
"learning_rate": 2.42e-06,
|
| 37 |
+
"loss": 3.3563,
|
| 38 |
+
"step": 125
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.19,
|
| 42 |
+
"learning_rate": 2.92e-06,
|
| 43 |
+
"loss": 3.3,
|
| 44 |
+
"step": 150
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.23,
|
| 48 |
+
"learning_rate": 3.4200000000000007e-06,
|
| 49 |
+
"loss": 3.2606,
|
| 50 |
+
"step": 175
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"epoch": 0.26,
|
| 54 |
+
"learning_rate": 3.920000000000001e-06,
|
| 55 |
+
"loss": 3.2285,
|
| 56 |
+
"step": 200
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"epoch": 0.29,
|
| 60 |
+
"learning_rate": 4.42e-06,
|
| 61 |
+
"loss": 3.1568,
|
| 62 |
+
"step": 225
|
| 63 |
+
},
|
| 64 |
+
{
|
| 65 |
+
"epoch": 0.32,
|
| 66 |
+
"learning_rate": 4.92e-06,
|
| 67 |
+
"loss": 3.1274,
|
| 68 |
+
"step": 250
|
| 69 |
+
},
|
| 70 |
+
{
|
| 71 |
+
"epoch": 0.36,
|
| 72 |
+
"learning_rate": 5.420000000000001e-06,
|
| 73 |
+
"loss": 3.0402,
|
| 74 |
+
"step": 275
|
| 75 |
+
},
|
| 76 |
+
{
|
| 77 |
+
"epoch": 0.39,
|
| 78 |
+
"learning_rate": 5.92e-06,
|
| 79 |
+
"loss": 3.094,
|
| 80 |
+
"step": 300
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 0.42,
|
| 84 |
+
"learning_rate": 6.42e-06,
|
| 85 |
+
"loss": 2.9766,
|
| 86 |
+
"step": 325
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.45,
|
| 90 |
+
"learning_rate": 6.92e-06,
|
| 91 |
+
"loss": 2.6614,
|
| 92 |
+
"step": 350
|
| 93 |
+
},
|
| 94 |
+
{
|
| 95 |
+
"epoch": 0.49,
|
| 96 |
+
"learning_rate": 7.420000000000001e-06,
|
| 97 |
+
"loss": 2.5191,
|
| 98 |
+
"step": 375
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 0.52,
|
| 102 |
+
"learning_rate": 7.92e-06,
|
| 103 |
+
"loss": 2.4224,
|
| 104 |
+
"step": 400
|
| 105 |
+
},
|
| 106 |
+
{
|
| 107 |
+
"epoch": 0.55,
|
| 108 |
+
"learning_rate": 8.42e-06,
|
| 109 |
+
"loss": 2.4196,
|
| 110 |
+
"step": 425
|
| 111 |
+
},
|
| 112 |
+
{
|
| 113 |
+
"epoch": 0.58,
|
| 114 |
+
"learning_rate": 8.920000000000001e-06,
|
| 115 |
+
"loss": 2.3527,
|
| 116 |
+
"step": 450
|
| 117 |
+
},
|
| 118 |
+
{
|
| 119 |
+
"epoch": 0.61,
|
| 120 |
+
"learning_rate": 9.42e-06,
|
| 121 |
+
"loss": 2.3644,
|
| 122 |
+
"step": 475
|
| 123 |
+
},
|
| 124 |
+
{
|
| 125 |
+
"epoch": 0.65,
|
| 126 |
+
"learning_rate": 9.920000000000002e-06,
|
| 127 |
+
"loss": 2.2919,
|
| 128 |
+
"step": 500
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.68,
|
| 132 |
+
"learning_rate": 9.940000000000001e-06,
|
| 133 |
+
"loss": 2.2907,
|
| 134 |
+
"step": 525
|
| 135 |
+
},
|
| 136 |
+
{
|
| 137 |
+
"epoch": 0.71,
|
| 138 |
+
"learning_rate": 9.86857142857143e-06,
|
| 139 |
+
"loss": 2.2864,
|
| 140 |
+
"step": 550
|
| 141 |
+
},
|
| 142 |
+
{
|
| 143 |
+
"epoch": 0.74,
|
| 144 |
+
"learning_rate": 9.797142857142858e-06,
|
| 145 |
+
"loss": 2.3032,
|
| 146 |
+
"step": 575
|
| 147 |
+
},
|
| 148 |
+
{
|
| 149 |
+
"epoch": 0.78,
|
| 150 |
+
"learning_rate": 9.725714285714287e-06,
|
| 151 |
+
"loss": 2.3143,
|
| 152 |
+
"step": 600
|
| 153 |
+
},
|
| 154 |
+
{
|
| 155 |
+
"epoch": 0.81,
|
| 156 |
+
"learning_rate": 9.654285714285716e-06,
|
| 157 |
+
"loss": 2.2779,
|
| 158 |
+
"step": 625
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 0.84,
|
| 162 |
+
"learning_rate": 9.582857142857143e-06,
|
| 163 |
+
"loss": 2.2326,
|
| 164 |
+
"step": 650
|
| 165 |
+
},
|
| 166 |
+
{
|
| 167 |
+
"epoch": 0.87,
|
| 168 |
+
"learning_rate": 9.511428571428572e-06,
|
| 169 |
+
"loss": 2.2106,
|
| 170 |
+
"step": 675
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 0.91,
|
| 174 |
+
"learning_rate": 9.440000000000001e-06,
|
| 175 |
+
"loss": 2.2268,
|
| 176 |
+
"step": 700
|
| 177 |
+
},
|
| 178 |
+
{
|
| 179 |
+
"epoch": 0.94,
|
| 180 |
+
"learning_rate": 9.368571428571428e-06,
|
| 181 |
+
"loss": 2.2314,
|
| 182 |
+
"step": 725
|
| 183 |
+
},
|
| 184 |
+
{
|
| 185 |
+
"epoch": 0.97,
|
| 186 |
+
"learning_rate": 9.297142857142857e-06,
|
| 187 |
+
"loss": 2.1607,
|
| 188 |
+
"step": 750
|
| 189 |
+
},
|
| 190 |
+
{
|
| 191 |
+
"epoch": 1.0,
|
| 192 |
+
"learning_rate": 9.225714285714286e-06,
|
| 193 |
+
"loss": 2.2081,
|
| 194 |
+
"step": 775
|
| 195 |
+
},
|
| 196 |
+
{
|
| 197 |
+
"epoch": 1.03,
|
| 198 |
+
"learning_rate": 9.154285714285715e-06,
|
| 199 |
+
"loss": 1.9049,
|
| 200 |
+
"step": 800
|
| 201 |
+
},
|
| 202 |
+
{
|
| 203 |
+
"epoch": 1.07,
|
| 204 |
+
"learning_rate": 9.082857142857143e-06,
|
| 205 |
+
"loss": 1.9724,
|
| 206 |
+
"step": 825
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 1.1,
|
| 210 |
+
"learning_rate": 9.011428571428572e-06,
|
| 211 |
+
"loss": 1.8894,
|
| 212 |
+
"step": 850
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 1.13,
|
| 216 |
+
"learning_rate": 8.94e-06,
|
| 217 |
+
"loss": 1.9267,
|
| 218 |
+
"step": 875
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 1.16,
|
| 222 |
+
"learning_rate": 8.86857142857143e-06,
|
| 223 |
+
"loss": 1.8992,
|
| 224 |
+
"step": 900
|
| 225 |
+
},
|
| 226 |
+
{
|
| 227 |
+
"epoch": 1.2,
|
| 228 |
+
"learning_rate": 8.797142857142857e-06,
|
| 229 |
+
"loss": 1.954,
|
| 230 |
+
"step": 925
|
| 231 |
+
},
|
| 232 |
+
{
|
| 233 |
+
"epoch": 1.23,
|
| 234 |
+
"learning_rate": 8.725714285714286e-06,
|
| 235 |
+
"loss": 1.9228,
|
| 236 |
+
"step": 950
|
| 237 |
+
},
|
| 238 |
+
{
|
| 239 |
+
"epoch": 1.26,
|
| 240 |
+
"learning_rate": 8.654285714285715e-06,
|
| 241 |
+
"loss": 1.9127,
|
| 242 |
+
"step": 975
|
| 243 |
+
},
|
| 244 |
+
{
|
| 245 |
+
"epoch": 1.29,
|
| 246 |
+
"learning_rate": 8.582857142857144e-06,
|
| 247 |
+
"loss": 1.9129,
|
| 248 |
+
"step": 1000
|
| 249 |
+
},
|
| 250 |
+
{
|
| 251 |
+
"epoch": 1.29,
|
| 252 |
+
"eval_loss": 2.144951581954956,
|
| 253 |
+
"eval_runtime": 1662.5185,
|
| 254 |
+
"eval_samples_per_second": 1.86,
|
| 255 |
+
"eval_steps_per_second": 0.233,
|
| 256 |
+
"eval_wer": 130.23632416272198,
|
| 257 |
+
"step": 1000
|
| 258 |
+
},
|
| 259 |
+
{
|
| 260 |
+
"epoch": 1.33,
|
| 261 |
+
"learning_rate": 8.511428571428571e-06,
|
| 262 |
+
"loss": 1.8509,
|
| 263 |
+
"step": 1025
|
| 264 |
+
},
|
| 265 |
+
{
|
| 266 |
+
"epoch": 1.36,
|
| 267 |
+
"learning_rate": 8.44e-06,
|
| 268 |
+
"loss": 1.8959,
|
| 269 |
+
"step": 1050
|
| 270 |
+
},
|
| 271 |
+
{
|
| 272 |
+
"epoch": 1.39,
|
| 273 |
+
"learning_rate": 8.36857142857143e-06,
|
| 274 |
+
"loss": 1.8961,
|
| 275 |
+
"step": 1075
|
| 276 |
+
},
|
| 277 |
+
{
|
| 278 |
+
"epoch": 1.42,
|
| 279 |
+
"learning_rate": 8.297142857142859e-06,
|
| 280 |
+
"loss": 1.9555,
|
| 281 |
+
"step": 1100
|
| 282 |
+
},
|
| 283 |
+
{
|
| 284 |
+
"epoch": 1.46,
|
| 285 |
+
"learning_rate": 8.225714285714288e-06,
|
| 286 |
+
"loss": 1.842,
|
| 287 |
+
"step": 1125
|
| 288 |
+
},
|
| 289 |
+
{
|
| 290 |
+
"epoch": 1.49,
|
| 291 |
+
"learning_rate": 8.154285714285715e-06,
|
| 292 |
+
"loss": 1.8813,
|
| 293 |
+
"step": 1150
|
| 294 |
+
},
|
| 295 |
+
{
|
| 296 |
+
"epoch": 1.52,
|
| 297 |
+
"learning_rate": 8.082857142857144e-06,
|
| 298 |
+
"loss": 1.9007,
|
| 299 |
+
"step": 1175
|
| 300 |
+
},
|
| 301 |
+
{
|
| 302 |
+
"epoch": 1.55,
|
| 303 |
+
"learning_rate": 8.011428571428573e-06,
|
| 304 |
+
"loss": 1.8851,
|
| 305 |
+
"step": 1200
|
| 306 |
+
},
|
| 307 |
+
{
|
| 308 |
+
"epoch": 1.58,
|
| 309 |
+
"learning_rate": 7.94e-06,
|
| 310 |
+
"loss": 1.8445,
|
| 311 |
+
"step": 1225
|
| 312 |
+
},
|
| 313 |
+
{
|
| 314 |
+
"epoch": 1.62,
|
| 315 |
+
"learning_rate": 7.86857142857143e-06,
|
| 316 |
+
"loss": 1.932,
|
| 317 |
+
"step": 1250
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 1.65,
|
| 321 |
+
"learning_rate": 7.797142857142858e-06,
|
| 322 |
+
"loss": 1.9009,
|
| 323 |
+
"step": 1275
|
| 324 |
+
},
|
| 325 |
+
{
|
| 326 |
+
"epoch": 1.68,
|
| 327 |
+
"learning_rate": 7.725714285714286e-06,
|
| 328 |
+
"loss": 1.8665,
|
| 329 |
+
"step": 1300
|
| 330 |
+
},
|
| 331 |
+
{
|
| 332 |
+
"epoch": 1.71,
|
| 333 |
+
"learning_rate": 7.654285714285715e-06,
|
| 334 |
+
"loss": 1.862,
|
| 335 |
+
"step": 1325
|
| 336 |
+
},
|
| 337 |
+
{
|
| 338 |
+
"epoch": 1.75,
|
| 339 |
+
"learning_rate": 7.5828571428571444e-06,
|
| 340 |
+
"loss": 1.8806,
|
| 341 |
+
"step": 1350
|
| 342 |
+
},
|
| 343 |
+
{
|
| 344 |
+
"epoch": 1.78,
|
| 345 |
+
"learning_rate": 7.511428571428572e-06,
|
| 346 |
+
"loss": 1.8959,
|
| 347 |
+
"step": 1375
|
| 348 |
+
},
|
| 349 |
+
{
|
| 350 |
+
"epoch": 1.81,
|
| 351 |
+
"learning_rate": 7.440000000000001e-06,
|
| 352 |
+
"loss": 1.8419,
|
| 353 |
+
"step": 1400
|
| 354 |
+
},
|
| 355 |
+
{
|
| 356 |
+
"epoch": 1.84,
|
| 357 |
+
"learning_rate": 7.36857142857143e-06,
|
| 358 |
+
"loss": 1.8484,
|
| 359 |
+
"step": 1425
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 1.88,
|
| 363 |
+
"learning_rate": 7.297142857142858e-06,
|
| 364 |
+
"loss": 1.8625,
|
| 365 |
+
"step": 1450
|
| 366 |
+
},
|
| 367 |
+
{
|
| 368 |
+
"epoch": 1.91,
|
| 369 |
+
"learning_rate": 7.225714285714286e-06,
|
| 370 |
+
"loss": 1.8971,
|
| 371 |
+
"step": 1475
|
| 372 |
+
},
|
| 373 |
+
{
|
| 374 |
+
"epoch": 1.94,
|
| 375 |
+
"learning_rate": 7.154285714285715e-06,
|
| 376 |
+
"loss": 1.8816,
|
| 377 |
+
"step": 1500
|
| 378 |
+
},
|
| 379 |
+
{
|
| 380 |
+
"epoch": 1.97,
|
| 381 |
+
"learning_rate": 7.082857142857143e-06,
|
| 382 |
+
"loss": 1.8283,
|
| 383 |
+
"step": 1525
|
| 384 |
+
},
|
| 385 |
+
{
|
| 386 |
+
"epoch": 2.01,
|
| 387 |
+
"learning_rate": 7.011428571428572e-06,
|
| 388 |
+
"loss": 1.7663,
|
| 389 |
+
"step": 1550
|
| 390 |
+
},
|
| 391 |
+
{
|
| 392 |
+
"epoch": 2.04,
|
| 393 |
+
"learning_rate": 6.9400000000000005e-06,
|
| 394 |
+
"loss": 1.5536,
|
| 395 |
+
"step": 1575
|
| 396 |
+
},
|
| 397 |
+
{
|
| 398 |
+
"epoch": 2.07,
|
| 399 |
+
"learning_rate": 6.868571428571429e-06,
|
| 400 |
+
"loss": 1.4843,
|
| 401 |
+
"step": 1600
|
| 402 |
+
},
|
| 403 |
+
{
|
| 404 |
+
"epoch": 2.1,
|
| 405 |
+
"learning_rate": 6.797142857142858e-06,
|
| 406 |
+
"loss": 1.4662,
|
| 407 |
+
"step": 1625
|
| 408 |
+
},
|
| 409 |
+
{
|
| 410 |
+
"epoch": 2.13,
|
| 411 |
+
"learning_rate": 6.725714285714287e-06,
|
| 412 |
+
"loss": 1.4634,
|
| 413 |
+
"step": 1650
|
| 414 |
+
},
|
| 415 |
+
{
|
| 416 |
+
"epoch": 2.17,
|
| 417 |
+
"learning_rate": 6.654285714285716e-06,
|
| 418 |
+
"loss": 1.5186,
|
| 419 |
+
"step": 1675
|
| 420 |
+
},
|
| 421 |
+
{
|
| 422 |
+
"epoch": 2.2,
|
| 423 |
+
"learning_rate": 6.582857142857143e-06,
|
| 424 |
+
"loss": 1.5014,
|
| 425 |
+
"step": 1700
|
| 426 |
+
},
|
| 427 |
+
{
|
| 428 |
+
"epoch": 2.23,
|
| 429 |
+
"learning_rate": 6.511428571428572e-06,
|
| 430 |
+
"loss": 1.5121,
|
| 431 |
+
"step": 1725
|
| 432 |
+
},
|
| 433 |
+
{
|
| 434 |
+
"epoch": 2.26,
|
| 435 |
+
"learning_rate": 6.440000000000001e-06,
|
| 436 |
+
"loss": 1.4588,
|
| 437 |
+
"step": 1750
|
| 438 |
+
},
|
| 439 |
+
{
|
| 440 |
+
"epoch": 2.3,
|
| 441 |
+
"learning_rate": 6.368571428571429e-06,
|
| 442 |
+
"loss": 1.5382,
|
| 443 |
+
"step": 1775
|
| 444 |
+
},
|
| 445 |
+
{
|
| 446 |
+
"epoch": 2.33,
|
| 447 |
+
"learning_rate": 6.297142857142857e-06,
|
| 448 |
+
"loss": 1.5234,
|
| 449 |
+
"step": 1800
|
| 450 |
+
},
|
| 451 |
+
{
|
| 452 |
+
"epoch": 2.36,
|
| 453 |
+
"learning_rate": 6.225714285714286e-06,
|
| 454 |
+
"loss": 1.4473,
|
| 455 |
+
"step": 1825
|
| 456 |
+
},
|
| 457 |
+
{
|
| 458 |
+
"epoch": 2.39,
|
| 459 |
+
"learning_rate": 6.1542857142857145e-06,
|
| 460 |
+
"loss": 1.4842,
|
| 461 |
+
"step": 1850
|
| 462 |
+
},
|
| 463 |
+
{
|
| 464 |
+
"epoch": 2.43,
|
| 465 |
+
"learning_rate": 6.0828571428571435e-06,
|
| 466 |
+
"loss": 1.5178,
|
| 467 |
+
"step": 1875
|
| 468 |
+
},
|
| 469 |
+
{
|
| 470 |
+
"epoch": 2.46,
|
| 471 |
+
"learning_rate": 6.011428571428572e-06,
|
| 472 |
+
"loss": 1.5751,
|
| 473 |
+
"step": 1900
|
| 474 |
+
},
|
| 475 |
+
{
|
| 476 |
+
"epoch": 2.49,
|
| 477 |
+
"learning_rate": 5.94e-06,
|
| 478 |
+
"loss": 1.5134,
|
| 479 |
+
"step": 1925
|
| 480 |
+
},
|
| 481 |
+
{
|
| 482 |
+
"epoch": 2.52,
|
| 483 |
+
"learning_rate": 5.868571428571429e-06,
|
| 484 |
+
"loss": 1.5102,
|
| 485 |
+
"step": 1950
|
| 486 |
+
},
|
| 487 |
+
{
|
| 488 |
+
"epoch": 2.55,
|
| 489 |
+
"learning_rate": 5.797142857142858e-06,
|
| 490 |
+
"loss": 1.5782,
|
| 491 |
+
"step": 1975
|
| 492 |
+
},
|
| 493 |
+
{
|
| 494 |
+
"epoch": 2.59,
|
| 495 |
+
"learning_rate": 5.725714285714287e-06,
|
| 496 |
+
"loss": 1.5301,
|
| 497 |
+
"step": 2000
|
| 498 |
+
},
|
| 499 |
+
{
|
| 500 |
+
"epoch": 2.59,
|
| 501 |
+
"eval_loss": 2.0798511505126953,
|
| 502 |
+
"eval_runtime": 1458.9783,
|
| 503 |
+
"eval_samples_per_second": 2.119,
|
| 504 |
+
"eval_steps_per_second": 0.265,
|
| 505 |
+
"eval_wer": 109.43695056411717,
|
| 506 |
+
"step": 2000
|
| 507 |
+
},
|
| 508 |
+
{
|
| 509 |
+
"epoch": 2.62,
|
| 510 |
+
"learning_rate": 5.654285714285714e-06,
|
| 511 |
+
"loss": 1.5363,
|
| 512 |
+
"step": 2025
|
| 513 |
+
},
|
| 514 |
+
{
|
| 515 |
+
"epoch": 2.65,
|
| 516 |
+
"learning_rate": 5.582857142857143e-06,
|
| 517 |
+
"loss": 1.481,
|
| 518 |
+
"step": 2050
|
| 519 |
+
},
|
| 520 |
+
{
|
| 521 |
+
"epoch": 2.68,
|
| 522 |
+
"learning_rate": 5.511428571428572e-06,
|
| 523 |
+
"loss": 1.4956,
|
| 524 |
+
"step": 2075
|
| 525 |
+
},
|
| 526 |
+
{
|
| 527 |
+
"epoch": 2.72,
|
| 528 |
+
"learning_rate": 5.4400000000000004e-06,
|
| 529 |
+
"loss": 1.4914,
|
| 530 |
+
"step": 2100
|
| 531 |
+
},
|
| 532 |
+
{
|
| 533 |
+
"epoch": 2.75,
|
| 534 |
+
"learning_rate": 5.368571428571429e-06,
|
| 535 |
+
"loss": 1.5007,
|
| 536 |
+
"step": 2125
|
| 537 |
+
},
|
| 538 |
+
{
|
| 539 |
+
"epoch": 2.78,
|
| 540 |
+
"learning_rate": 5.297142857142858e-06,
|
| 541 |
+
"loss": 1.4407,
|
| 542 |
+
"step": 2150
|
| 543 |
+
},
|
| 544 |
+
{
|
| 545 |
+
"epoch": 2.81,
|
| 546 |
+
"learning_rate": 5.225714285714286e-06,
|
| 547 |
+
"loss": 1.5283,
|
| 548 |
+
"step": 2175
|
| 549 |
+
},
|
| 550 |
+
{
|
| 551 |
+
"epoch": 2.85,
|
| 552 |
+
"learning_rate": 5.154285714285715e-06,
|
| 553 |
+
"loss": 1.4745,
|
| 554 |
+
"step": 2200
|
| 555 |
+
},
|
| 556 |
+
{
|
| 557 |
+
"epoch": 2.88,
|
| 558 |
+
"learning_rate": 5.082857142857144e-06,
|
| 559 |
+
"loss": 1.5076,
|
| 560 |
+
"step": 2225
|
| 561 |
+
},
|
| 562 |
+
{
|
| 563 |
+
"epoch": 2.91,
|
| 564 |
+
"learning_rate": 5.011428571428571e-06,
|
| 565 |
+
"loss": 1.5108,
|
| 566 |
+
"step": 2250
|
| 567 |
+
},
|
| 568 |
+
{
|
| 569 |
+
"epoch": 2.94,
|
| 570 |
+
"learning_rate": 4.94e-06,
|
| 571 |
+
"loss": 1.5575,
|
| 572 |
+
"step": 2275
|
| 573 |
+
},
|
| 574 |
+
{
|
| 575 |
+
"epoch": 2.98,
|
| 576 |
+
"learning_rate": 4.868571428571429e-06,
|
| 577 |
+
"loss": 1.4867,
|
| 578 |
+
"step": 2300
|
| 579 |
+
},
|
| 580 |
+
{
|
| 581 |
+
"epoch": 3.01,
|
| 582 |
+
"learning_rate": 4.797142857142857e-06,
|
| 583 |
+
"loss": 1.4243,
|
| 584 |
+
"step": 2325
|
| 585 |
+
},
|
| 586 |
+
{
|
| 587 |
+
"epoch": 3.04,
|
| 588 |
+
"learning_rate": 4.725714285714286e-06,
|
| 589 |
+
"loss": 1.2097,
|
| 590 |
+
"step": 2350
|
| 591 |
+
},
|
| 592 |
+
{
|
| 593 |
+
"epoch": 3.07,
|
| 594 |
+
"learning_rate": 4.6542857142857145e-06,
|
| 595 |
+
"loss": 1.1907,
|
| 596 |
+
"step": 2375
|
| 597 |
+
},
|
| 598 |
+
{
|
| 599 |
+
"epoch": 3.1,
|
| 600 |
+
"learning_rate": 4.5828571428571435e-06,
|
| 601 |
+
"loss": 1.1848,
|
| 602 |
+
"step": 2400
|
| 603 |
+
},
|
| 604 |
+
{
|
| 605 |
+
"epoch": 3.14,
|
| 606 |
+
"learning_rate": 4.511428571428572e-06,
|
| 607 |
+
"loss": 1.1698,
|
| 608 |
+
"step": 2425
|
| 609 |
+
},
|
| 610 |
+
{
|
| 611 |
+
"epoch": 3.17,
|
| 612 |
+
"learning_rate": 4.440000000000001e-06,
|
| 613 |
+
"loss": 1.2308,
|
| 614 |
+
"step": 2450
|
| 615 |
+
},
|
| 616 |
+
{
|
| 617 |
+
"epoch": 3.2,
|
| 618 |
+
"learning_rate": 4.368571428571429e-06,
|
| 619 |
+
"loss": 1.2729,
|
| 620 |
+
"step": 2475
|
| 621 |
+
},
|
| 622 |
+
{
|
| 623 |
+
"epoch": 3.23,
|
| 624 |
+
"learning_rate": 4.297142857142858e-06,
|
| 625 |
+
"loss": 1.2145,
|
| 626 |
+
"step": 2500
|
| 627 |
+
},
|
| 628 |
+
{
|
| 629 |
+
"epoch": 3.27,
|
| 630 |
+
"learning_rate": 4.225714285714286e-06,
|
| 631 |
+
"loss": 1.2325,
|
| 632 |
+
"step": 2525
|
| 633 |
+
},
|
| 634 |
+
{
|
| 635 |
+
"epoch": 3.3,
|
| 636 |
+
"learning_rate": 4.154285714285714e-06,
|
| 637 |
+
"loss": 1.2873,
|
| 638 |
+
"step": 2550
|
| 639 |
+
},
|
| 640 |
+
{
|
| 641 |
+
"epoch": 3.33,
|
| 642 |
+
"learning_rate": 4.082857142857143e-06,
|
| 643 |
+
"loss": 1.2272,
|
| 644 |
+
"step": 2575
|
| 645 |
+
},
|
| 646 |
+
{
|
| 647 |
+
"epoch": 3.36,
|
| 648 |
+
"learning_rate": 4.011428571428571e-06,
|
| 649 |
+
"loss": 1.2025,
|
| 650 |
+
"step": 2600
|
| 651 |
+
},
|
| 652 |
+
{
|
| 653 |
+
"epoch": 3.4,
|
| 654 |
+
"learning_rate": 3.94e-06,
|
| 655 |
+
"loss": 1.2726,
|
| 656 |
+
"step": 2625
|
| 657 |
+
},
|
| 658 |
+
{
|
| 659 |
+
"epoch": 3.43,
|
| 660 |
+
"learning_rate": 3.8685714285714286e-06,
|
| 661 |
+
"loss": 1.2182,
|
| 662 |
+
"step": 2650
|
| 663 |
+
},
|
| 664 |
+
{
|
| 665 |
+
"epoch": 3.46,
|
| 666 |
+
"learning_rate": 3.7971428571428576e-06,
|
| 667 |
+
"loss": 1.25,
|
| 668 |
+
"step": 2675
|
| 669 |
+
},
|
| 670 |
+
{
|
| 671 |
+
"epoch": 3.49,
|
| 672 |
+
"learning_rate": 3.7257142857142857e-06,
|
| 673 |
+
"loss": 1.2539,
|
| 674 |
+
"step": 2700
|
| 675 |
+
},
|
| 676 |
+
{
|
| 677 |
+
"epoch": 3.53,
|
| 678 |
+
"learning_rate": 3.6542857142857148e-06,
|
| 679 |
+
"loss": 1.2184,
|
| 680 |
+
"step": 2725
|
| 681 |
+
},
|
| 682 |
+
{
|
| 683 |
+
"epoch": 3.56,
|
| 684 |
+
"learning_rate": 3.582857142857143e-06,
|
| 685 |
+
"loss": 1.2406,
|
| 686 |
+
"step": 2750
|
| 687 |
+
},
|
| 688 |
+
{
|
| 689 |
+
"epoch": 3.59,
|
| 690 |
+
"learning_rate": 3.511428571428572e-06,
|
| 691 |
+
"loss": 1.2307,
|
| 692 |
+
"step": 2775
|
| 693 |
+
},
|
| 694 |
+
{
|
| 695 |
+
"epoch": 3.62,
|
| 696 |
+
"learning_rate": 3.44e-06,
|
| 697 |
+
"loss": 1.1811,
|
| 698 |
+
"step": 2800
|
| 699 |
+
},
|
| 700 |
+
{
|
| 701 |
+
"epoch": 3.65,
|
| 702 |
+
"learning_rate": 3.3685714285714287e-06,
|
| 703 |
+
"loss": 1.227,
|
| 704 |
+
"step": 2825
|
| 705 |
+
},
|
| 706 |
+
{
|
| 707 |
+
"epoch": 3.69,
|
| 708 |
+
"learning_rate": 3.2971428571428577e-06,
|
| 709 |
+
"loss": 1.267,
|
| 710 |
+
"step": 2850
|
| 711 |
+
},
|
| 712 |
+
{
|
| 713 |
+
"epoch": 3.72,
|
| 714 |
+
"learning_rate": 3.225714285714286e-06,
|
| 715 |
+
"loss": 1.2616,
|
| 716 |
+
"step": 2875
|
| 717 |
+
},
|
| 718 |
+
{
|
| 719 |
+
"epoch": 3.75,
|
| 720 |
+
"learning_rate": 3.154285714285715e-06,
|
| 721 |
+
"loss": 1.2183,
|
| 722 |
+
"step": 2900
|
| 723 |
+
},
|
| 724 |
+
{
|
| 725 |
+
"epoch": 3.78,
|
| 726 |
+
"learning_rate": 3.082857142857143e-06,
|
| 727 |
+
"loss": 1.2221,
|
| 728 |
+
"step": 2925
|
| 729 |
+
},
|
| 730 |
+
{
|
| 731 |
+
"epoch": 3.82,
|
| 732 |
+
"learning_rate": 3.0114285714285716e-06,
|
| 733 |
+
"loss": 1.1916,
|
| 734 |
+
"step": 2950
|
| 735 |
+
},
|
| 736 |
+
{
|
| 737 |
+
"epoch": 3.85,
|
| 738 |
+
"learning_rate": 2.9400000000000002e-06,
|
| 739 |
+
"loss": 1.1972,
|
| 740 |
+
"step": 2975
|
| 741 |
+
},
|
| 742 |
+
{
|
| 743 |
+
"epoch": 3.88,
|
| 744 |
+
"learning_rate": 2.868571428571429e-06,
|
| 745 |
+
"loss": 1.1928,
|
| 746 |
+
"step": 3000
|
| 747 |
+
},
|
| 748 |
+
{
|
| 749 |
+
"epoch": 3.88,
|
| 750 |
+
"eval_loss": 2.1198830604553223,
|
| 751 |
+
"eval_runtime": 1449.4974,
|
| 752 |
+
"eval_samples_per_second": 2.133,
|
| 753 |
+
"eval_steps_per_second": 0.267,
|
| 754 |
+
"eval_wer": 107.93501085525146,
|
| 755 |
+
"step": 3000
|
| 756 |
+
},
|
| 757 |
+
{
|
| 758 |
+
"epoch": 3.91,
|
| 759 |
+
"learning_rate": 2.797142857142857e-06,
|
| 760 |
+
"loss": 1.2472,
|
| 761 |
+
"step": 3025
|
| 762 |
+
},
|
| 763 |
+
{
|
| 764 |
+
"epoch": 3.95,
|
| 765 |
+
"learning_rate": 2.725714285714286e-06,
|
| 766 |
+
"loss": 1.2415,
|
| 767 |
+
"step": 3050
|
| 768 |
+
},
|
| 769 |
+
{
|
| 770 |
+
"epoch": 3.98,
|
| 771 |
+
"learning_rate": 2.654285714285714e-06,
|
| 772 |
+
"loss": 1.2215,
|
| 773 |
+
"step": 3075
|
| 774 |
+
},
|
| 775 |
+
{
|
| 776 |
+
"epoch": 4.01,
|
| 777 |
+
"learning_rate": 2.582857142857143e-06,
|
| 778 |
+
"loss": 1.0969,
|
| 779 |
+
"step": 3100
|
| 780 |
+
},
|
| 781 |
+
{
|
| 782 |
+
"epoch": 4.04,
|
| 783 |
+
"learning_rate": 2.5114285714285718e-06,
|
| 784 |
+
"loss": 1.0057,
|
| 785 |
+
"step": 3125
|
| 786 |
+
},
|
| 787 |
+
{
|
| 788 |
+
"epoch": 4.08,
|
| 789 |
+
"learning_rate": 2.4400000000000004e-06,
|
| 790 |
+
"loss": 1.0103,
|
| 791 |
+
"step": 3150
|
| 792 |
+
},
|
| 793 |
+
{
|
| 794 |
+
"epoch": 4.11,
|
| 795 |
+
"learning_rate": 2.3685714285714285e-06,
|
| 796 |
+
"loss": 1.0181,
|
| 797 |
+
"step": 3175
|
| 798 |
+
},
|
| 799 |
+
{
|
| 800 |
+
"epoch": 4.14,
|
| 801 |
+
"learning_rate": 2.297142857142857e-06,
|
| 802 |
+
"loss": 1.0259,
|
| 803 |
+
"step": 3200
|
| 804 |
+
},
|
| 805 |
+
{
|
| 806 |
+
"epoch": 4.17,
|
| 807 |
+
"learning_rate": 2.2257142857142857e-06,
|
| 808 |
+
"loss": 1.0801,
|
| 809 |
+
"step": 3225
|
| 810 |
+
},
|
| 811 |
+
{
|
| 812 |
+
"epoch": 4.2,
|
| 813 |
+
"learning_rate": 2.1542857142857147e-06,
|
| 814 |
+
"loss": 1.0349,
|
| 815 |
+
"step": 3250
|
| 816 |
+
},
|
| 817 |
+
{
|
| 818 |
+
"epoch": 4.24,
|
| 819 |
+
"learning_rate": 2.0828571428571433e-06,
|
| 820 |
+
"loss": 1.0428,
|
| 821 |
+
"step": 3275
|
| 822 |
+
},
|
| 823 |
+
{
|
| 824 |
+
"epoch": 4.27,
|
| 825 |
+
"learning_rate": 2.0114285714285715e-06,
|
| 826 |
+
"loss": 1.0637,
|
| 827 |
+
"step": 3300
|
| 828 |
+
},
|
| 829 |
+
{
|
| 830 |
+
"epoch": 4.3,
|
| 831 |
+
"learning_rate": 1.94e-06,
|
| 832 |
+
"loss": 0.981,
|
| 833 |
+
"step": 3325
|
| 834 |
+
},
|
| 835 |
+
{
|
| 836 |
+
"epoch": 4.33,
|
| 837 |
+
"learning_rate": 1.8685714285714289e-06,
|
| 838 |
+
"loss": 1.0443,
|
| 839 |
+
"step": 3350
|
| 840 |
+
},
|
| 841 |
+
{
|
| 842 |
+
"epoch": 4.37,
|
| 843 |
+
"learning_rate": 1.7971428571428572e-06,
|
| 844 |
+
"loss": 1.0579,
|
| 845 |
+
"step": 3375
|
| 846 |
+
},
|
| 847 |
+
{
|
| 848 |
+
"epoch": 4.4,
|
| 849 |
+
"learning_rate": 1.7257142857142858e-06,
|
| 850 |
+
"loss": 1.0075,
|
| 851 |
+
"step": 3400
|
| 852 |
+
},
|
| 853 |
+
{
|
| 854 |
+
"epoch": 4.43,
|
| 855 |
+
"learning_rate": 1.6542857142857144e-06,
|
| 856 |
+
"loss": 1.0172,
|
| 857 |
+
"step": 3425
|
| 858 |
+
},
|
| 859 |
+
{
|
| 860 |
+
"epoch": 4.46,
|
| 861 |
+
"learning_rate": 1.582857142857143e-06,
|
| 862 |
+
"loss": 1.0415,
|
| 863 |
+
"step": 3450
|
| 864 |
+
},
|
| 865 |
+
{
|
| 866 |
+
"epoch": 4.5,
|
| 867 |
+
"learning_rate": 1.5114285714285714e-06,
|
| 868 |
+
"loss": 1.0134,
|
| 869 |
+
"step": 3475
|
| 870 |
+
},
|
| 871 |
+
{
|
| 872 |
+
"epoch": 4.53,
|
| 873 |
+
"learning_rate": 1.44e-06,
|
| 874 |
+
"loss": 1.0163,
|
| 875 |
+
"step": 3500
|
| 876 |
+
},
|
| 877 |
+
{
|
| 878 |
+
"epoch": 4.56,
|
| 879 |
+
"learning_rate": 1.3685714285714286e-06,
|
| 880 |
+
"loss": 1.0415,
|
| 881 |
+
"step": 3525
|
| 882 |
+
},
|
| 883 |
+
{
|
| 884 |
+
"epoch": 4.59,
|
| 885 |
+
"learning_rate": 1.2971428571428574e-06,
|
| 886 |
+
"loss": 0.9805,
|
| 887 |
+
"step": 3550
|
| 888 |
+
},
|
| 889 |
+
{
|
| 890 |
+
"epoch": 4.62,
|
| 891 |
+
"learning_rate": 1.2257142857142857e-06,
|
| 892 |
+
"loss": 1.0357,
|
| 893 |
+
"step": 3575
|
| 894 |
+
},
|
| 895 |
+
{
|
| 896 |
+
"epoch": 4.66,
|
| 897 |
+
"learning_rate": 1.1542857142857143e-06,
|
| 898 |
+
"loss": 1.0072,
|
| 899 |
+
"step": 3600
|
| 900 |
+
},
|
| 901 |
+
{
|
| 902 |
+
"epoch": 4.69,
|
| 903 |
+
"learning_rate": 1.082857142857143e-06,
|
| 904 |
+
"loss": 1.0199,
|
| 905 |
+
"step": 3625
|
| 906 |
+
},
|
| 907 |
+
{
|
| 908 |
+
"epoch": 4.72,
|
| 909 |
+
"learning_rate": 1.0114285714285715e-06,
|
| 910 |
+
"loss": 1.0332,
|
| 911 |
+
"step": 3650
|
| 912 |
+
},
|
| 913 |
+
{
|
| 914 |
+
"epoch": 4.75,
|
| 915 |
+
"learning_rate": 9.400000000000001e-07,
|
| 916 |
+
"loss": 1.0361,
|
| 917 |
+
"step": 3675
|
| 918 |
+
},
|
| 919 |
+
{
|
| 920 |
+
"epoch": 4.79,
|
| 921 |
+
"learning_rate": 8.685714285714286e-07,
|
| 922 |
+
"loss": 1.0312,
|
| 923 |
+
"step": 3700
|
| 924 |
+
},
|
| 925 |
+
{
|
| 926 |
+
"epoch": 4.82,
|
| 927 |
+
"learning_rate": 7.971428571428572e-07,
|
| 928 |
+
"loss": 1.0118,
|
| 929 |
+
"step": 3725
|
| 930 |
+
},
|
| 931 |
+
{
|
| 932 |
+
"epoch": 4.85,
|
| 933 |
+
"learning_rate": 7.257142857142857e-07,
|
| 934 |
+
"loss": 1.0689,
|
| 935 |
+
"step": 3750
|
| 936 |
+
},
|
| 937 |
+
{
|
| 938 |
+
"epoch": 4.88,
|
| 939 |
+
"learning_rate": 6.542857142857144e-07,
|
| 940 |
+
"loss": 1.0719,
|
| 941 |
+
"step": 3775
|
| 942 |
+
},
|
| 943 |
+
{
|
| 944 |
+
"epoch": 4.92,
|
| 945 |
+
"learning_rate": 5.82857142857143e-07,
|
| 946 |
+
"loss": 1.0092,
|
| 947 |
+
"step": 3800
|
| 948 |
+
},
|
| 949 |
+
{
|
| 950 |
+
"epoch": 4.95,
|
| 951 |
+
"learning_rate": 5.114285714285714e-07,
|
| 952 |
+
"loss": 1.0135,
|
| 953 |
+
"step": 3825
|
| 954 |
+
},
|
| 955 |
+
{
|
| 956 |
+
"epoch": 4.98,
|
| 957 |
+
"learning_rate": 4.4e-07,
|
| 958 |
+
"loss": 1.0456,
|
| 959 |
+
"step": 3850
|
| 960 |
+
},
|
| 961 |
+
{
|
| 962 |
+
"epoch": 5.01,
|
| 963 |
+
"learning_rate": 3.685714285714286e-07,
|
| 964 |
+
"loss": 0.9733,
|
| 965 |
+
"step": 3875
|
| 966 |
+
},
|
| 967 |
+
{
|
| 968 |
+
"epoch": 5.05,
|
| 969 |
+
"learning_rate": 2.9714285714285715e-07,
|
| 970 |
+
"loss": 0.9442,
|
| 971 |
+
"step": 3900
|
| 972 |
+
},
|
| 973 |
+
{
|
| 974 |
+
"epoch": 5.08,
|
| 975 |
+
"learning_rate": 2.2571428571428574e-07,
|
| 976 |
+
"loss": 0.9531,
|
| 977 |
+
"step": 3925
|
| 978 |
+
},
|
| 979 |
+
{
|
| 980 |
+
"epoch": 5.11,
|
| 981 |
+
"learning_rate": 1.542857142857143e-07,
|
| 982 |
+
"loss": 0.9083,
|
| 983 |
+
"step": 3950
|
| 984 |
+
},
|
| 985 |
+
{
|
| 986 |
+
"epoch": 5.14,
|
| 987 |
+
"learning_rate": 8.285714285714285e-08,
|
| 988 |
+
"loss": 0.9279,
|
| 989 |
+
"step": 3975
|
| 990 |
+
},
|
| 991 |
+
{
|
| 992 |
+
"epoch": 5.17,
|
| 993 |
+
"learning_rate": 1.142857142857143e-08,
|
| 994 |
+
"loss": 0.8998,
|
| 995 |
+
"step": 4000
|
| 996 |
+
},
|
| 997 |
+
{
|
| 998 |
+
"epoch": 5.17,
|
| 999 |
+
"eval_loss": 2.1974828243255615,
|
| 1000 |
+
"eval_runtime": 1430.422,
|
| 1001 |
+
"eval_samples_per_second": 2.162,
|
| 1002 |
+
"eval_steps_per_second": 0.271,
|
| 1003 |
+
"eval_wer": 104.12321600170837,
|
| 1004 |
+
"step": 4000
|
| 1005 |
+
}
|
| 1006 |
+
],
|
| 1007 |
+
"max_steps": 4000,
|
| 1008 |
+
"num_train_epochs": 6,
|
| 1009 |
+
"total_flos": 1.846946562048e+19,
|
| 1010 |
+
"trial_name": null,
|
| 1011 |
+
"trial_params": null
|
| 1012 |
+
}
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:97e8f93fdbac12be15ed9b546ccb3c7f34754c61c4598f9d6840f4a5505b7096
|
| 3 |
+
size 3631
|