Commit
•
c0da500
1
Parent(s):
8a9f3a1
Upload 8 files
Browse files- config.json +62 -0
- generation_config.json +7 -0
- model.safetensors +3 -0
- optimizer.pt +3 -0
- rng_state.pth +3 -0
- scheduler.pt +3 -0
- trainer_state.json +1764 -0
- training_args.bin +3 -0
config.json
ADDED
|
@@ -0,0 +1,62 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "google/flan-t5-base",
|
| 3 |
+
"architectures": [
|
| 4 |
+
"T5ForConditionalGeneration"
|
| 5 |
+
],
|
| 6 |
+
"classifier_dropout": 0.0,
|
| 7 |
+
"d_ff": 2048,
|
| 8 |
+
"d_kv": 64,
|
| 9 |
+
"d_model": 768,
|
| 10 |
+
"decoder_start_token_id": 0,
|
| 11 |
+
"dense_act_fn": "gelu_new",
|
| 12 |
+
"dropout_rate": 0.1,
|
| 13 |
+
"eos_token_id": 1,
|
| 14 |
+
"feed_forward_proj": "gated-gelu",
|
| 15 |
+
"initializer_factor": 1.0,
|
| 16 |
+
"is_encoder_decoder": true,
|
| 17 |
+
"is_gated_act": true,
|
| 18 |
+
"layer_norm_epsilon": 1e-06,
|
| 19 |
+
"model_type": "t5",
|
| 20 |
+
"n_positions": 512,
|
| 21 |
+
"num_decoder_layers": 12,
|
| 22 |
+
"num_heads": 12,
|
| 23 |
+
"num_layers": 12,
|
| 24 |
+
"output_past": true,
|
| 25 |
+
"pad_token_id": 0,
|
| 26 |
+
"relative_attention_max_distance": 128,
|
| 27 |
+
"relative_attention_num_buckets": 32,
|
| 28 |
+
"task_specific_params": {
|
| 29 |
+
"summarization": {
|
| 30 |
+
"early_stopping": true,
|
| 31 |
+
"length_penalty": 2.0,
|
| 32 |
+
"max_length": 200,
|
| 33 |
+
"min_length": 30,
|
| 34 |
+
"no_repeat_ngram_size": 3,
|
| 35 |
+
"num_beams": 4,
|
| 36 |
+
"prefix": "summarize: "
|
| 37 |
+
},
|
| 38 |
+
"translation_en_to_de": {
|
| 39 |
+
"early_stopping": true,
|
| 40 |
+
"max_length": 300,
|
| 41 |
+
"num_beams": 4,
|
| 42 |
+
"prefix": "translate English to German: "
|
| 43 |
+
},
|
| 44 |
+
"translation_en_to_fr": {
|
| 45 |
+
"early_stopping": true,
|
| 46 |
+
"max_length": 300,
|
| 47 |
+
"num_beams": 4,
|
| 48 |
+
"prefix": "translate English to French: "
|
| 49 |
+
},
|
| 50 |
+
"translation_en_to_ro": {
|
| 51 |
+
"early_stopping": true,
|
| 52 |
+
"max_length": 300,
|
| 53 |
+
"num_beams": 4,
|
| 54 |
+
"prefix": "translate English to Romanian: "
|
| 55 |
+
}
|
| 56 |
+
},
|
| 57 |
+
"tie_word_embeddings": false,
|
| 58 |
+
"torch_dtype": "float32",
|
| 59 |
+
"transformers_version": "4.38.2",
|
| 60 |
+
"use_cache": true,
|
| 61 |
+
"vocab_size": 32128
|
| 62 |
+
}
|
generation_config.json
ADDED
|
@@ -0,0 +1,7 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_from_model_config": true,
|
| 3 |
+
"decoder_start_token_id": 0,
|
| 4 |
+
"eos_token_id": 1,
|
| 5 |
+
"pad_token_id": 0,
|
| 6 |
+
"transformers_version": "4.38.2"
|
| 7 |
+
}
|
model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c3520b59b0dd7d97a031ecd3e65081e854bff69be00f487fb8f81f90020551e6
|
| 3 |
+
size 990345064
|
optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:95d101298f498a4f274795f5a3d6a7f4ea3705e200b3c31c973b9a7c43599c36
|
| 3 |
+
size 1980860410
|
rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:4eeef2cfc1e239441a025515ab2eb0f85d3fbcdd40a4d0c71a536cf0bc5017c2
|
| 3 |
+
size 14244
|
scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:bba85b58bc8295f09ff8afef7eba146075c94eaf5c52d2212394529c76bbb31d
|
| 3 |
+
size 1064
|
trainer_state.json
ADDED
|
@@ -0,0 +1,1764 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 2.0,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 24920,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.01,
|
| 13 |
+
"grad_norm": 2.4730186462402344,
|
| 14 |
+
"learning_rate": 2e-05,
|
| 15 |
+
"loss": 0.3314,
|
| 16 |
+
"step": 100
|
| 17 |
+
},
|
| 18 |
+
{
|
| 19 |
+
"epoch": 0.02,
|
| 20 |
+
"grad_norm": 3.50852108001709,
|
| 21 |
+
"learning_rate": 1.9999872447769624e-05,
|
| 22 |
+
"loss": 0.0313,
|
| 23 |
+
"step": 200
|
| 24 |
+
},
|
| 25 |
+
{
|
| 26 |
+
"epoch": 0.02,
|
| 27 |
+
"grad_norm": 0.40423300862312317,
|
| 28 |
+
"learning_rate": 1.9999489794332404e-05,
|
| 29 |
+
"loss": 0.0226,
|
| 30 |
+
"step": 300
|
| 31 |
+
},
|
| 32 |
+
{
|
| 33 |
+
"epoch": 0.03,
|
| 34 |
+
"grad_norm": 0.4923788011074066,
|
| 35 |
+
"learning_rate": 1.9998852049449998e-05,
|
| 36 |
+
"loss": 0.0144,
|
| 37 |
+
"step": 400
|
| 38 |
+
},
|
| 39 |
+
{
|
| 40 |
+
"epoch": 0.04,
|
| 41 |
+
"grad_norm": 0.039026230573654175,
|
| 42 |
+
"learning_rate": 1.9997959229391567e-05,
|
| 43 |
+
"loss": 0.0088,
|
| 44 |
+
"step": 500
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.05,
|
| 48 |
+
"grad_norm": 0.13339382410049438,
|
| 49 |
+
"learning_rate": 1.9996811356933346e-05,
|
| 50 |
+
"loss": 0.0097,
|
| 51 |
+
"step": 600
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.06,
|
| 55 |
+
"grad_norm": 0.03016798384487629,
|
| 56 |
+
"learning_rate": 1.9995408461358074e-05,
|
| 57 |
+
"loss": 0.0063,
|
| 58 |
+
"step": 700
|
| 59 |
+
},
|
| 60 |
+
{
|
| 61 |
+
"epoch": 0.06,
|
| 62 |
+
"grad_norm": 0.12673313915729523,
|
| 63 |
+
"learning_rate": 1.9993750578454248e-05,
|
| 64 |
+
"loss": 0.0084,
|
| 65 |
+
"step": 800
|
| 66 |
+
},
|
| 67 |
+
{
|
| 68 |
+
"epoch": 0.07,
|
| 69 |
+
"grad_norm": 0.012224284932017326,
|
| 70 |
+
"learning_rate": 1.999183775051519e-05,
|
| 71 |
+
"loss": 0.006,
|
| 72 |
+
"step": 900
|
| 73 |
+
},
|
| 74 |
+
{
|
| 75 |
+
"epoch": 0.08,
|
| 76 |
+
"grad_norm": 2.144702434539795,
|
| 77 |
+
"learning_rate": 1.9989670026338002e-05,
|
| 78 |
+
"loss": 0.0076,
|
| 79 |
+
"step": 1000
|
| 80 |
+
},
|
| 81 |
+
{
|
| 82 |
+
"epoch": 0.09,
|
| 83 |
+
"grad_norm": 0.02996153011918068,
|
| 84 |
+
"learning_rate": 1.9987247461222297e-05,
|
| 85 |
+
"loss": 0.0052,
|
| 86 |
+
"step": 1100
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.1,
|
| 90 |
+
"grad_norm": 0.008165411651134491,
|
| 91 |
+
"learning_rate": 1.9984570116968785e-05,
|
| 92 |
+
"loss": 0.0052,
|
| 93 |
+
"step": 1200
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 0.1,
|
| 97 |
+
"grad_norm": 0.07781720906496048,
|
| 98 |
+
"learning_rate": 1.9981638061877714e-05,
|
| 99 |
+
"loss": 0.0056,
|
| 100 |
+
"step": 1300
|
| 101 |
+
},
|
| 102 |
+
{
|
| 103 |
+
"epoch": 0.11,
|
| 104 |
+
"grad_norm": 0.016230745241045952,
|
| 105 |
+
"learning_rate": 1.9978451370747122e-05,
|
| 106 |
+
"loss": 0.0052,
|
| 107 |
+
"step": 1400
|
| 108 |
+
},
|
| 109 |
+
{
|
| 110 |
+
"epoch": 0.12,
|
| 111 |
+
"grad_norm": 2.58638858795166,
|
| 112 |
+
"learning_rate": 1.997501012487091e-05,
|
| 113 |
+
"loss": 0.0037,
|
| 114 |
+
"step": 1500
|
| 115 |
+
},
|
| 116 |
+
{
|
| 117 |
+
"epoch": 0.13,
|
| 118 |
+
"grad_norm": 0.029858984053134918,
|
| 119 |
+
"learning_rate": 1.9971314412036807e-05,
|
| 120 |
+
"loss": 0.0037,
|
| 121 |
+
"step": 1600
|
| 122 |
+
},
|
| 123 |
+
{
|
| 124 |
+
"epoch": 0.14,
|
| 125 |
+
"grad_norm": 0.2630612850189209,
|
| 126 |
+
"learning_rate": 1.996736432652409e-05,
|
| 127 |
+
"loss": 0.0046,
|
| 128 |
+
"step": 1700
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.14,
|
| 132 |
+
"grad_norm": 0.013704606331884861,
|
| 133 |
+
"learning_rate": 1.9963159969101207e-05,
|
| 134 |
+
"loss": 0.0056,
|
| 135 |
+
"step": 1800
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 0.15,
|
| 139 |
+
"grad_norm": 0.09422887861728668,
|
| 140 |
+
"learning_rate": 1.9958701447023188e-05,
|
| 141 |
+
"loss": 0.0041,
|
| 142 |
+
"step": 1900
|
| 143 |
+
},
|
| 144 |
+
{
|
| 145 |
+
"epoch": 0.16,
|
| 146 |
+
"grad_norm": 0.02264581061899662,
|
| 147 |
+
"learning_rate": 1.9953988874028917e-05,
|
| 148 |
+
"loss": 0.0039,
|
| 149 |
+
"step": 2000
|
| 150 |
+
},
|
| 151 |
+
{
|
| 152 |
+
"epoch": 0.17,
|
| 153 |
+
"grad_norm": 0.020697351545095444,
|
| 154 |
+
"learning_rate": 1.994902237033824e-05,
|
| 155 |
+
"loss": 0.0048,
|
| 156 |
+
"step": 2100
|
| 157 |
+
},
|
| 158 |
+
{
|
| 159 |
+
"epoch": 0.18,
|
| 160 |
+
"grad_norm": 0.1434525102376938,
|
| 161 |
+
"learning_rate": 1.9943802062648877e-05,
|
| 162 |
+
"loss": 0.0051,
|
| 163 |
+
"step": 2200
|
| 164 |
+
},
|
| 165 |
+
{
|
| 166 |
+
"epoch": 0.18,
|
| 167 |
+
"grad_norm": 0.06430936604738235,
|
| 168 |
+
"learning_rate": 1.9938328084133206e-05,
|
| 169 |
+
"loss": 0.0023,
|
| 170 |
+
"step": 2300
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 0.19,
|
| 174 |
+
"grad_norm": 0.008740647695958614,
|
| 175 |
+
"learning_rate": 1.9932600574434864e-05,
|
| 176 |
+
"loss": 0.004,
|
| 177 |
+
"step": 2400
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 0.2,
|
| 181 |
+
"grad_norm": 0.014448602683842182,
|
| 182 |
+
"learning_rate": 1.9926619679665175e-05,
|
| 183 |
+
"loss": 0.0043,
|
| 184 |
+
"step": 2500
|
| 185 |
+
},
|
| 186 |
+
{
|
| 187 |
+
"epoch": 0.21,
|
| 188 |
+
"grad_norm": 0.02901746705174446,
|
| 189 |
+
"learning_rate": 1.9920385552399434e-05,
|
| 190 |
+
"loss": 0.0019,
|
| 191 |
+
"step": 2600
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"epoch": 0.22,
|
| 195 |
+
"grad_norm": 1.6094048023223877,
|
| 196 |
+
"learning_rate": 1.9913898351673006e-05,
|
| 197 |
+
"loss": 0.0022,
|
| 198 |
+
"step": 2700
|
| 199 |
+
},
|
| 200 |
+
{
|
| 201 |
+
"epoch": 0.22,
|
| 202 |
+
"grad_norm": 0.36414438486099243,
|
| 203 |
+
"learning_rate": 1.990715824297728e-05,
|
| 204 |
+
"loss": 0.0056,
|
| 205 |
+
"step": 2800
|
| 206 |
+
},
|
| 207 |
+
{
|
| 208 |
+
"epoch": 0.23,
|
| 209 |
+
"grad_norm": 0.1421067714691162,
|
| 210 |
+
"learning_rate": 1.9900165398255434e-05,
|
| 211 |
+
"loss": 0.003,
|
| 212 |
+
"step": 2900
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 0.24,
|
| 216 |
+
"grad_norm": 0.028148509562015533,
|
| 217 |
+
"learning_rate": 1.9892919995898052e-05,
|
| 218 |
+
"loss": 0.0016,
|
| 219 |
+
"step": 3000
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 0.25,
|
| 223 |
+
"grad_norm": 0.015177663415670395,
|
| 224 |
+
"learning_rate": 1.9885422220738583e-05,
|
| 225 |
+
"loss": 0.004,
|
| 226 |
+
"step": 3100
|
| 227 |
+
},
|
| 228 |
+
{
|
| 229 |
+
"epoch": 0.26,
|
| 230 |
+
"grad_norm": 0.25167518854141235,
|
| 231 |
+
"learning_rate": 1.9877672264048618e-05,
|
| 232 |
+
"loss": 0.0027,
|
| 233 |
+
"step": 3200
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"epoch": 0.26,
|
| 237 |
+
"grad_norm": 0.03997454047203064,
|
| 238 |
+
"learning_rate": 1.9869670323533005e-05,
|
| 239 |
+
"loss": 0.003,
|
| 240 |
+
"step": 3300
|
| 241 |
+
},
|
| 242 |
+
{
|
| 243 |
+
"epoch": 0.27,
|
| 244 |
+
"grad_norm": 0.0028598604258149862,
|
| 245 |
+
"learning_rate": 1.986141660332482e-05,
|
| 246 |
+
"loss": 0.0019,
|
| 247 |
+
"step": 3400
|
| 248 |
+
},
|
| 249 |
+
{
|
| 250 |
+
"epoch": 0.28,
|
| 251 |
+
"grad_norm": 0.1343453824520111,
|
| 252 |
+
"learning_rate": 1.9852911313980146e-05,
|
| 253 |
+
"loss": 0.0034,
|
| 254 |
+
"step": 3500
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 0.29,
|
| 258 |
+
"grad_norm": 0.12534217536449432,
|
| 259 |
+
"learning_rate": 1.9844154672472707e-05,
|
| 260 |
+
"loss": 0.0026,
|
| 261 |
+
"step": 3600
|
| 262 |
+
},
|
| 263 |
+
{
|
| 264 |
+
"epoch": 0.3,
|
| 265 |
+
"grad_norm": 0.24615176022052765,
|
| 266 |
+
"learning_rate": 1.9835146902188336e-05,
|
| 267 |
+
"loss": 0.0018,
|
| 268 |
+
"step": 3700
|
| 269 |
+
},
|
| 270 |
+
{
|
| 271 |
+
"epoch": 0.3,
|
| 272 |
+
"grad_norm": 0.013632328249514103,
|
| 273 |
+
"learning_rate": 1.9825888232919268e-05,
|
| 274 |
+
"loss": 0.0019,
|
| 275 |
+
"step": 3800
|
| 276 |
+
},
|
| 277 |
+
{
|
| 278 |
+
"epoch": 0.31,
|
| 279 |
+
"grad_norm": 0.4895065724849701,
|
| 280 |
+
"learning_rate": 1.9816378900858288e-05,
|
| 281 |
+
"loss": 0.0037,
|
| 282 |
+
"step": 3900
|
| 283 |
+
},
|
| 284 |
+
{
|
| 285 |
+
"epoch": 0.32,
|
| 286 |
+
"grad_norm": 1.3317159414291382,
|
| 287 |
+
"learning_rate": 1.98066191485927e-05,
|
| 288 |
+
"loss": 0.0022,
|
| 289 |
+
"step": 4000
|
| 290 |
+
},
|
| 291 |
+
{
|
| 292 |
+
"epoch": 0.33,
|
| 293 |
+
"grad_norm": 0.008369619026780128,
|
| 294 |
+
"learning_rate": 1.9796609225098136e-05,
|
| 295 |
+
"loss": 0.0008,
|
| 296 |
+
"step": 4100
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 0.34,
|
| 300 |
+
"grad_norm": 0.007471293676644564,
|
| 301 |
+
"learning_rate": 1.9786349385732212e-05,
|
| 302 |
+
"loss": 0.0031,
|
| 303 |
+
"step": 4200
|
| 304 |
+
},
|
| 305 |
+
{
|
| 306 |
+
"epoch": 0.35,
|
| 307 |
+
"grad_norm": 0.011221354827284813,
|
| 308 |
+
"learning_rate": 1.9775839892228004e-05,
|
| 309 |
+
"loss": 0.0013,
|
| 310 |
+
"step": 4300
|
| 311 |
+
},
|
| 312 |
+
{
|
| 313 |
+
"epoch": 0.35,
|
| 314 |
+
"grad_norm": 0.006684939842671156,
|
| 315 |
+
"learning_rate": 1.976508101268738e-05,
|
| 316 |
+
"loss": 0.0024,
|
| 317 |
+
"step": 4400
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 0.36,
|
| 321 |
+
"grad_norm": 0.7115832567214966,
|
| 322 |
+
"learning_rate": 1.9754073021574153e-05,
|
| 323 |
+
"loss": 0.0027,
|
| 324 |
+
"step": 4500
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 0.37,
|
| 328 |
+
"grad_norm": 0.031222190707921982,
|
| 329 |
+
"learning_rate": 1.9742816199707096e-05,
|
| 330 |
+
"loss": 0.0013,
|
| 331 |
+
"step": 4600
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 0.38,
|
| 335 |
+
"grad_norm": 0.033377427607774734,
|
| 336 |
+
"learning_rate": 1.9731310834252747e-05,
|
| 337 |
+
"loss": 0.0046,
|
| 338 |
+
"step": 4700
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 0.39,
|
| 342 |
+
"grad_norm": 0.3707186281681061,
|
| 343 |
+
"learning_rate": 1.9719557218718116e-05,
|
| 344 |
+
"loss": 0.0025,
|
| 345 |
+
"step": 4800
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 0.39,
|
| 349 |
+
"grad_norm": 0.030659371986985207,
|
| 350 |
+
"learning_rate": 1.970755565294318e-05,
|
| 351 |
+
"loss": 0.0033,
|
| 352 |
+
"step": 4900
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 0.4,
|
| 356 |
+
"grad_norm": 0.031716570258140564,
|
| 357 |
+
"learning_rate": 1.969530644309323e-05,
|
| 358 |
+
"loss": 0.0017,
|
| 359 |
+
"step": 5000
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 0.41,
|
| 363 |
+
"grad_norm": 0.018681248649954796,
|
| 364 |
+
"learning_rate": 1.9682809901651074e-05,
|
| 365 |
+
"loss": 0.0025,
|
| 366 |
+
"step": 5100
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 0.42,
|
| 370 |
+
"grad_norm": 0.031807415187358856,
|
| 371 |
+
"learning_rate": 1.9670066347409063e-05,
|
| 372 |
+
"loss": 0.0029,
|
| 373 |
+
"step": 5200
|
| 374 |
+
},
|
| 375 |
+
{
|
| 376 |
+
"epoch": 0.43,
|
| 377 |
+
"grad_norm": 0.00439372006803751,
|
| 378 |
+
"learning_rate": 1.9657076105460945e-05,
|
| 379 |
+
"loss": 0.0024,
|
| 380 |
+
"step": 5300
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 0.43,
|
| 384 |
+
"grad_norm": 0.2656748592853546,
|
| 385 |
+
"learning_rate": 1.964383950719359e-05,
|
| 386 |
+
"loss": 0.0019,
|
| 387 |
+
"step": 5400
|
| 388 |
+
},
|
| 389 |
+
{
|
| 390 |
+
"epoch": 0.44,
|
| 391 |
+
"grad_norm": 0.18905937671661377,
|
| 392 |
+
"learning_rate": 1.9630356890278527e-05,
|
| 393 |
+
"loss": 0.0015,
|
| 394 |
+
"step": 5500
|
| 395 |
+
},
|
| 396 |
+
{
|
| 397 |
+
"epoch": 0.45,
|
| 398 |
+
"grad_norm": 0.1290869414806366,
|
| 399 |
+
"learning_rate": 1.9616628598663322e-05,
|
| 400 |
+
"loss": 0.0034,
|
| 401 |
+
"step": 5600
|
| 402 |
+
},
|
| 403 |
+
{
|
| 404 |
+
"epoch": 0.46,
|
| 405 |
+
"grad_norm": 0.004318041726946831,
|
| 406 |
+
"learning_rate": 1.9602654982562822e-05,
|
| 407 |
+
"loss": 0.0014,
|
| 408 |
+
"step": 5700
|
| 409 |
+
},
|
| 410 |
+
{
|
| 411 |
+
"epoch": 0.47,
|
| 412 |
+
"grad_norm": 0.15103724598884583,
|
| 413 |
+
"learning_rate": 1.9588436398450206e-05,
|
| 414 |
+
"loss": 0.0013,
|
| 415 |
+
"step": 5800
|
| 416 |
+
},
|
| 417 |
+
{
|
| 418 |
+
"epoch": 0.47,
|
| 419 |
+
"grad_norm": 0.039078064262866974,
|
| 420 |
+
"learning_rate": 1.9573973209047893e-05,
|
| 421 |
+
"loss": 0.0015,
|
| 422 |
+
"step": 5900
|
| 423 |
+
},
|
| 424 |
+
{
|
| 425 |
+
"epoch": 0.48,
|
| 426 |
+
"grad_norm": 2.7625222206115723,
|
| 427 |
+
"learning_rate": 1.9559265783318304e-05,
|
| 428 |
+
"loss": 0.0018,
|
| 429 |
+
"step": 6000
|
| 430 |
+
},
|
| 431 |
+
{
|
| 432 |
+
"epoch": 0.49,
|
| 433 |
+
"grad_norm": 0.32574662566185,
|
| 434 |
+
"learning_rate": 1.9544314496454423e-05,
|
| 435 |
+
"loss": 0.0023,
|
| 436 |
+
"step": 6100
|
| 437 |
+
},
|
| 438 |
+
{
|
| 439 |
+
"epoch": 0.5,
|
| 440 |
+
"grad_norm": 0.014155671931803226,
|
| 441 |
+
"learning_rate": 1.9529119729870253e-05,
|
| 442 |
+
"loss": 0.0011,
|
| 443 |
+
"step": 6200
|
| 444 |
+
},
|
| 445 |
+
{
|
| 446 |
+
"epoch": 0.51,
|
| 447 |
+
"grad_norm": 0.007221277803182602,
|
| 448 |
+
"learning_rate": 1.9513681871191063e-05,
|
| 449 |
+
"loss": 0.0033,
|
| 450 |
+
"step": 6300
|
| 451 |
+
},
|
| 452 |
+
{
|
| 453 |
+
"epoch": 0.51,
|
| 454 |
+
"grad_norm": 0.013470535166561604,
|
| 455 |
+
"learning_rate": 1.949800131424352e-05,
|
| 456 |
+
"loss": 0.003,
|
| 457 |
+
"step": 6400
|
| 458 |
+
},
|
| 459 |
+
{
|
| 460 |
+
"epoch": 0.52,
|
| 461 |
+
"grad_norm": 0.003437698120251298,
|
| 462 |
+
"learning_rate": 1.9482078459045617e-05,
|
| 463 |
+
"loss": 0.0012,
|
| 464 |
+
"step": 6500
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 0.53,
|
| 468 |
+
"grad_norm": 0.006796371191740036,
|
| 469 |
+
"learning_rate": 1.9465913711796502e-05,
|
| 470 |
+
"loss": 0.001,
|
| 471 |
+
"step": 6600
|
| 472 |
+
},
|
| 473 |
+
{
|
| 474 |
+
"epoch": 0.54,
|
| 475 |
+
"grad_norm": 0.7563895583152771,
|
| 476 |
+
"learning_rate": 1.9449507484866084e-05,
|
| 477 |
+
"loss": 0.0018,
|
| 478 |
+
"step": 6700
|
| 479 |
+
},
|
| 480 |
+
{
|
| 481 |
+
"epoch": 0.55,
|
| 482 |
+
"grad_norm": 0.016867786645889282,
|
| 483 |
+
"learning_rate": 1.9432860196784533e-05,
|
| 484 |
+
"loss": 0.0016,
|
| 485 |
+
"step": 6800
|
| 486 |
+
},
|
| 487 |
+
{
|
| 488 |
+
"epoch": 0.55,
|
| 489 |
+
"grad_norm": 0.006416236516088247,
|
| 490 |
+
"learning_rate": 1.941597227223159e-05,
|
| 491 |
+
"loss": 0.0008,
|
| 492 |
+
"step": 6900
|
| 493 |
+
},
|
| 494 |
+
{
|
| 495 |
+
"epoch": 0.56,
|
| 496 |
+
"grad_norm": 0.0008421270758844912,
|
| 497 |
+
"learning_rate": 1.9398844142025746e-05,
|
| 498 |
+
"loss": 0.0011,
|
| 499 |
+
"step": 7000
|
| 500 |
+
},
|
| 501 |
+
{
|
| 502 |
+
"epoch": 0.57,
|
| 503 |
+
"grad_norm": 0.012384007684886456,
|
| 504 |
+
"learning_rate": 1.9381476243113243e-05,
|
| 505 |
+
"loss": 0.0017,
|
| 506 |
+
"step": 7100
|
| 507 |
+
},
|
| 508 |
+
{
|
| 509 |
+
"epoch": 0.58,
|
| 510 |
+
"grad_norm": 0.004778598435223103,
|
| 511 |
+
"learning_rate": 1.9363869018556928e-05,
|
| 512 |
+
"loss": 0.0016,
|
| 513 |
+
"step": 7200
|
| 514 |
+
},
|
| 515 |
+
{
|
| 516 |
+
"epoch": 0.59,
|
| 517 |
+
"grad_norm": 0.00394839234650135,
|
| 518 |
+
"learning_rate": 1.9346022917524958e-05,
|
| 519 |
+
"loss": 0.0017,
|
| 520 |
+
"step": 7300
|
| 521 |
+
},
|
| 522 |
+
{
|
| 523 |
+
"epoch": 0.59,
|
| 524 |
+
"grad_norm": 0.006818379741162062,
|
| 525 |
+
"learning_rate": 1.9327938395279325e-05,
|
| 526 |
+
"loss": 0.0019,
|
| 527 |
+
"step": 7400
|
| 528 |
+
},
|
| 529 |
+
{
|
| 530 |
+
"epoch": 0.6,
|
| 531 |
+
"grad_norm": 0.028212955221533775,
|
| 532 |
+
"learning_rate": 1.9309615913164262e-05,
|
| 533 |
+
"loss": 0.0013,
|
| 534 |
+
"step": 7500
|
| 535 |
+
},
|
| 536 |
+
{
|
| 537 |
+
"epoch": 0.61,
|
| 538 |
+
"grad_norm": 0.017115216702222824,
|
| 539 |
+
"learning_rate": 1.9291055938594464e-05,
|
| 540 |
+
"loss": 0.0017,
|
| 541 |
+
"step": 7600
|
| 542 |
+
},
|
| 543 |
+
{
|
| 544 |
+
"epoch": 0.62,
|
| 545 |
+
"grad_norm": 0.02033141627907753,
|
| 546 |
+
"learning_rate": 1.9272258945043154e-05,
|
| 547 |
+
"loss": 0.0006,
|
| 548 |
+
"step": 7700
|
| 549 |
+
},
|
| 550 |
+
{
|
| 551 |
+
"epoch": 0.63,
|
| 552 |
+
"grad_norm": 1.2443268299102783,
|
| 553 |
+
"learning_rate": 1.9253225412030028e-05,
|
| 554 |
+
"loss": 0.0011,
|
| 555 |
+
"step": 7800
|
| 556 |
+
},
|
| 557 |
+
{
|
| 558 |
+
"epoch": 0.63,
|
| 559 |
+
"grad_norm": 0.0025956807658076286,
|
| 560 |
+
"learning_rate": 1.9233955825109e-05,
|
| 561 |
+
"loss": 0.001,
|
| 562 |
+
"step": 7900
|
| 563 |
+
},
|
| 564 |
+
{
|
| 565 |
+
"epoch": 0.64,
|
| 566 |
+
"grad_norm": 13.584095001220703,
|
| 567 |
+
"learning_rate": 1.9214450675855832e-05,
|
| 568 |
+
"loss": 0.0029,
|
| 569 |
+
"step": 8000
|
| 570 |
+
},
|
| 571 |
+
{
|
| 572 |
+
"epoch": 0.65,
|
| 573 |
+
"grad_norm": 0.0022025350481271744,
|
| 574 |
+
"learning_rate": 1.919471046185558e-05,
|
| 575 |
+
"loss": 0.001,
|
| 576 |
+
"step": 8100
|
| 577 |
+
},
|
| 578 |
+
{
|
| 579 |
+
"epoch": 0.66,
|
| 580 |
+
"grad_norm": 0.002522263675928116,
|
| 581 |
+
"learning_rate": 1.917473568668991e-05,
|
| 582 |
+
"loss": 0.0024,
|
| 583 |
+
"step": 8200
|
| 584 |
+
},
|
| 585 |
+
{
|
| 586 |
+
"epoch": 0.67,
|
| 587 |
+
"grad_norm": 1.3481580018997192,
|
| 588 |
+
"learning_rate": 1.9154526859924242e-05,
|
| 589 |
+
"loss": 0.0017,
|
| 590 |
+
"step": 8300
|
| 591 |
+
},
|
| 592 |
+
{
|
| 593 |
+
"epoch": 0.67,
|
| 594 |
+
"grad_norm": 0.010974978096783161,
|
| 595 |
+
"learning_rate": 1.9134084497094766e-05,
|
| 596 |
+
"loss": 0.0008,
|
| 597 |
+
"step": 8400
|
| 598 |
+
},
|
| 599 |
+
{
|
| 600 |
+
"epoch": 0.68,
|
| 601 |
+
"grad_norm": 0.005469560623168945,
|
| 602 |
+
"learning_rate": 1.9113409119695276e-05,
|
| 603 |
+
"loss": 0.0012,
|
| 604 |
+
"step": 8500
|
| 605 |
+
},
|
| 606 |
+
{
|
| 607 |
+
"epoch": 0.69,
|
| 608 |
+
"grad_norm": 0.08603531867265701,
|
| 609 |
+
"learning_rate": 1.9092501255163874e-05,
|
| 610 |
+
"loss": 0.0025,
|
| 611 |
+
"step": 8600
|
| 612 |
+
},
|
| 613 |
+
{
|
| 614 |
+
"epoch": 0.7,
|
| 615 |
+
"grad_norm": 0.0022059655748307705,
|
| 616 |
+
"learning_rate": 1.907136143686951e-05,
|
| 617 |
+
"loss": 0.0013,
|
| 618 |
+
"step": 8700
|
| 619 |
+
},
|
| 620 |
+
{
|
| 621 |
+
"epoch": 0.71,
|
| 622 |
+
"grad_norm": 0.0005606790073215961,
|
| 623 |
+
"learning_rate": 1.904999020409837e-05,
|
| 624 |
+
"loss": 0.0014,
|
| 625 |
+
"step": 8800
|
| 626 |
+
},
|
| 627 |
+
{
|
| 628 |
+
"epoch": 0.71,
|
| 629 |
+
"grad_norm": 0.012699578888714314,
|
| 630 |
+
"learning_rate": 1.902838810204015e-05,
|
| 631 |
+
"loss": 0.0012,
|
| 632 |
+
"step": 8900
|
| 633 |
+
},
|
| 634 |
+
{
|
| 635 |
+
"epoch": 0.72,
|
| 636 |
+
"grad_norm": 0.00312532065436244,
|
| 637 |
+
"learning_rate": 1.90065556817741e-05,
|
| 638 |
+
"loss": 0.0012,
|
| 639 |
+
"step": 9000
|
| 640 |
+
},
|
| 641 |
+
{
|
| 642 |
+
"epoch": 0.73,
|
| 643 |
+
"grad_norm": 0.008270219899713993,
|
| 644 |
+
"learning_rate": 1.8984493500255e-05,
|
| 645 |
+
"loss": 0.0014,
|
| 646 |
+
"step": 9100
|
| 647 |
+
},
|
| 648 |
+
{
|
| 649 |
+
"epoch": 0.74,
|
| 650 |
+
"grad_norm": 0.01274858694523573,
|
| 651 |
+
"learning_rate": 1.8962202120298948e-05,
|
| 652 |
+
"loss": 0.0013,
|
| 653 |
+
"step": 9200
|
| 654 |
+
},
|
| 655 |
+
{
|
| 656 |
+
"epoch": 0.75,
|
| 657 |
+
"grad_norm": 0.012512357905507088,
|
| 658 |
+
"learning_rate": 1.8939682110568982e-05,
|
| 659 |
+
"loss": 0.001,
|
| 660 |
+
"step": 9300
|
| 661 |
+
},
|
| 662 |
+
{
|
| 663 |
+
"epoch": 0.75,
|
| 664 |
+
"grad_norm": 0.0043404679745435715,
|
| 665 |
+
"learning_rate": 1.8916934045560603e-05,
|
| 666 |
+
"loss": 0.0023,
|
| 667 |
+
"step": 9400
|
| 668 |
+
},
|
| 669 |
+
{
|
| 670 |
+
"epoch": 0.76,
|
| 671 |
+
"grad_norm": 3.700385332107544,
|
| 672 |
+
"learning_rate": 1.8893958505587093e-05,
|
| 673 |
+
"loss": 0.0031,
|
| 674 |
+
"step": 9500
|
| 675 |
+
},
|
| 676 |
+
{
|
| 677 |
+
"epoch": 0.77,
|
| 678 |
+
"grad_norm": 0.016660619527101517,
|
| 679 |
+
"learning_rate": 1.8870756076764728e-05,
|
| 680 |
+
"loss": 0.0019,
|
| 681 |
+
"step": 9600
|
| 682 |
+
},
|
| 683 |
+
{
|
| 684 |
+
"epoch": 0.78,
|
| 685 |
+
"grad_norm": 0.047528013586997986,
|
| 686 |
+
"learning_rate": 1.8847327350997814e-05,
|
| 687 |
+
"loss": 0.0008,
|
| 688 |
+
"step": 9700
|
| 689 |
+
},
|
| 690 |
+
{
|
| 691 |
+
"epoch": 0.79,
|
| 692 |
+
"grad_norm": 0.01734699122607708,
|
| 693 |
+
"learning_rate": 1.8823672925963598e-05,
|
| 694 |
+
"loss": 0.0022,
|
| 695 |
+
"step": 9800
|
| 696 |
+
},
|
| 697 |
+
{
|
| 698 |
+
"epoch": 0.79,
|
| 699 |
+
"grad_norm": 0.019478172063827515,
|
| 700 |
+
"learning_rate": 1.879979340509701e-05,
|
| 701 |
+
"loss": 0.0009,
|
| 702 |
+
"step": 9900
|
| 703 |
+
},
|
| 704 |
+
{
|
| 705 |
+
"epoch": 0.8,
|
| 706 |
+
"grad_norm": 0.1385781317949295,
|
| 707 |
+
"learning_rate": 1.877568939757529e-05,
|
| 708 |
+
"loss": 0.0015,
|
| 709 |
+
"step": 10000
|
| 710 |
+
},
|
| 711 |
+
{
|
| 712 |
+
"epoch": 0.81,
|
| 713 |
+
"grad_norm": 0.143458753824234,
|
| 714 |
+
"learning_rate": 1.8751361518302413e-05,
|
| 715 |
+
"loss": 0.0008,
|
| 716 |
+
"step": 10100
|
| 717 |
+
},
|
| 718 |
+
{
|
| 719 |
+
"epoch": 0.82,
|
| 720 |
+
"grad_norm": 0.003875449998304248,
|
| 721 |
+
"learning_rate": 1.8726810387893438e-05,
|
| 722 |
+
"loss": 0.0029,
|
| 723 |
+
"step": 10200
|
| 724 |
+
},
|
| 725 |
+
{
|
| 726 |
+
"epoch": 0.83,
|
| 727 |
+
"grad_norm": 0.0030744324903935194,
|
| 728 |
+
"learning_rate": 1.8702036632658646e-05,
|
| 729 |
+
"loss": 0.0012,
|
| 730 |
+
"step": 10300
|
| 731 |
+
},
|
| 732 |
+
{
|
| 733 |
+
"epoch": 0.83,
|
| 734 |
+
"grad_norm": 0.012408553622663021,
|
| 735 |
+
"learning_rate": 1.867704088458759e-05,
|
| 736 |
+
"loss": 0.0017,
|
| 737 |
+
"step": 10400
|
| 738 |
+
},
|
| 739 |
+
{
|
| 740 |
+
"epoch": 0.84,
|
| 741 |
+
"grad_norm": 0.001399656874127686,
|
| 742 |
+
"learning_rate": 1.8651823781332948e-05,
|
| 743 |
+
"loss": 0.0006,
|
| 744 |
+
"step": 10500
|
| 745 |
+
},
|
| 746 |
+
{
|
| 747 |
+
"epoch": 0.85,
|
| 748 |
+
"grad_norm": 0.001192159834317863,
|
| 749 |
+
"learning_rate": 1.8626385966194275e-05,
|
| 750 |
+
"loss": 0.001,
|
| 751 |
+
"step": 10600
|
| 752 |
+
},
|
| 753 |
+
{
|
| 754 |
+
"epoch": 0.86,
|
| 755 |
+
"grad_norm": 0.6697074770927429,
|
| 756 |
+
"learning_rate": 1.8600728088101587e-05,
|
| 757 |
+
"loss": 0.0021,
|
| 758 |
+
"step": 10700
|
| 759 |
+
},
|
| 760 |
+
{
|
| 761 |
+
"epoch": 0.87,
|
| 762 |
+
"grad_norm": 0.007071709726005793,
|
| 763 |
+
"learning_rate": 1.857485080159879e-05,
|
| 764 |
+
"loss": 0.0016,
|
| 765 |
+
"step": 10800
|
| 766 |
+
},
|
| 767 |
+
{
|
| 768 |
+
"epoch": 0.87,
|
| 769 |
+
"grad_norm": 0.005609508138149977,
|
| 770 |
+
"learning_rate": 1.8548754766827016e-05,
|
| 771 |
+
"loss": 0.0022,
|
| 772 |
+
"step": 10900
|
| 773 |
+
},
|
| 774 |
+
{
|
| 775 |
+
"epoch": 0.88,
|
| 776 |
+
"grad_norm": 0.0166325643658638,
|
| 777 |
+
"learning_rate": 1.852244064950775e-05,
|
| 778 |
+
"loss": 0.0032,
|
| 779 |
+
"step": 11000
|
| 780 |
+
},
|
| 781 |
+
{
|
| 782 |
+
"epoch": 0.89,
|
| 783 |
+
"grad_norm": 0.0896034687757492,
|
| 784 |
+
"learning_rate": 1.8495909120925857e-05,
|
| 785 |
+
"loss": 0.0015,
|
| 786 |
+
"step": 11100
|
| 787 |
+
},
|
| 788 |
+
{
|
| 789 |
+
"epoch": 0.9,
|
| 790 |
+
"grad_norm": 0.24882763624191284,
|
| 791 |
+
"learning_rate": 1.846916085791247e-05,
|
| 792 |
+
"loss": 0.0017,
|
| 793 |
+
"step": 11200
|
| 794 |
+
},
|
| 795 |
+
{
|
| 796 |
+
"epoch": 0.91,
|
| 797 |
+
"grad_norm": 0.013551519252359867,
|
| 798 |
+
"learning_rate": 1.8442196542827712e-05,
|
| 799 |
+
"loss": 0.0007,
|
| 800 |
+
"step": 11300
|
| 801 |
+
},
|
| 802 |
+
{
|
| 803 |
+
"epoch": 0.91,
|
| 804 |
+
"grad_norm": 0.01764736883342266,
|
| 805 |
+
"learning_rate": 1.8415016863543286e-05,
|
| 806 |
+
"loss": 0.0005,
|
| 807 |
+
"step": 11400
|
| 808 |
+
},
|
| 809 |
+
{
|
| 810 |
+
"epoch": 0.92,
|
| 811 |
+
"grad_norm": 0.001765413791872561,
|
| 812 |
+
"learning_rate": 1.8387622513424942e-05,
|
| 813 |
+
"loss": 0.0044,
|
| 814 |
+
"step": 11500
|
| 815 |
+
},
|
| 816 |
+
{
|
| 817 |
+
"epoch": 0.93,
|
| 818 |
+
"grad_norm": 0.0014637404819950461,
|
| 819 |
+
"learning_rate": 1.836001419131476e-05,
|
| 820 |
+
"loss": 0.0012,
|
| 821 |
+
"step": 11600
|
| 822 |
+
},
|
| 823 |
+
{
|
| 824 |
+
"epoch": 0.94,
|
| 825 |
+
"grad_norm": 0.005813998635858297,
|
| 826 |
+
"learning_rate": 1.8332192601513358e-05,
|
| 827 |
+
"loss": 0.0006,
|
| 828 |
+
"step": 11700
|
| 829 |
+
},
|
| 830 |
+
{
|
| 831 |
+
"epoch": 0.95,
|
| 832 |
+
"grad_norm": 0.0008558441768400371,
|
| 833 |
+
"learning_rate": 1.8304158453761904e-05,
|
| 834 |
+
"loss": 0.002,
|
| 835 |
+
"step": 11800
|
| 836 |
+
},
|
| 837 |
+
{
|
| 838 |
+
"epoch": 0.96,
|
| 839 |
+
"grad_norm": 0.0033321972005069256,
|
| 840 |
+
"learning_rate": 1.827591246322401e-05,
|
| 841 |
+
"loss": 0.0005,
|
| 842 |
+
"step": 11900
|
| 843 |
+
},
|
| 844 |
+
{
|
| 845 |
+
"epoch": 0.96,
|
| 846 |
+
"grad_norm": 0.5678846836090088,
|
| 847 |
+
"learning_rate": 1.8247455350467496e-05,
|
| 848 |
+
"loss": 0.0014,
|
| 849 |
+
"step": 12000
|
| 850 |
+
},
|
| 851 |
+
{
|
| 852 |
+
"epoch": 0.97,
|
| 853 |
+
"grad_norm": 0.009241198189556599,
|
| 854 |
+
"learning_rate": 1.8218787841446003e-05,
|
| 855 |
+
"loss": 0.0004,
|
| 856 |
+
"step": 12100
|
| 857 |
+
},
|
| 858 |
+
{
|
| 859 |
+
"epoch": 0.98,
|
| 860 |
+
"grad_norm": 0.07260189205408096,
|
| 861 |
+
"learning_rate": 1.8189910667480476e-05,
|
| 862 |
+
"loss": 0.0015,
|
| 863 |
+
"step": 12200
|
| 864 |
+
},
|
| 865 |
+
{
|
| 866 |
+
"epoch": 0.99,
|
| 867 |
+
"grad_norm": 0.0022115109022706747,
|
| 868 |
+
"learning_rate": 1.8160824565240495e-05,
|
| 869 |
+
"loss": 0.0029,
|
| 870 |
+
"step": 12300
|
| 871 |
+
},
|
| 872 |
+
{
|
| 873 |
+
"epoch": 1.0,
|
| 874 |
+
"grad_norm": 0.020148636773228645,
|
| 875 |
+
"learning_rate": 1.8131530276725514e-05,
|
| 876 |
+
"loss": 0.0015,
|
| 877 |
+
"step": 12400
|
| 878 |
+
},
|
| 879 |
+
{
|
| 880 |
+
"epoch": 1.0,
|
| 881 |
+
"grad_norm": 0.2839891314506531,
|
| 882 |
+
"learning_rate": 1.8102028549245894e-05,
|
| 883 |
+
"loss": 0.0015,
|
| 884 |
+
"step": 12500
|
| 885 |
+
},
|
| 886 |
+
{
|
| 887 |
+
"epoch": 1.01,
|
| 888 |
+
"grad_norm": 0.004473233129829168,
|
| 889 |
+
"learning_rate": 1.8072320135403862e-05,
|
| 890 |
+
"loss": 0.0012,
|
| 891 |
+
"step": 12600
|
| 892 |
+
},
|
| 893 |
+
{
|
| 894 |
+
"epoch": 1.02,
|
| 895 |
+
"grad_norm": 0.0016079695196822286,
|
| 896 |
+
"learning_rate": 1.804240579307431e-05,
|
| 897 |
+
"loss": 0.0009,
|
| 898 |
+
"step": 12700
|
| 899 |
+
},
|
| 900 |
+
{
|
| 901 |
+
"epoch": 1.03,
|
| 902 |
+
"grad_norm": 0.015555166639387608,
|
| 903 |
+
"learning_rate": 1.8012286285385456e-05,
|
| 904 |
+
"loss": 0.0015,
|
| 905 |
+
"step": 12800
|
| 906 |
+
},
|
| 907 |
+
{
|
| 908 |
+
"epoch": 1.04,
|
| 909 |
+
"grad_norm": 0.014301074668765068,
|
| 910 |
+
"learning_rate": 1.7981962380699376e-05,
|
| 911 |
+
"loss": 0.0006,
|
| 912 |
+
"step": 12900
|
| 913 |
+
},
|
| 914 |
+
{
|
| 915 |
+
"epoch": 1.04,
|
| 916 |
+
"grad_norm": 0.005016946699470282,
|
| 917 |
+
"learning_rate": 1.7951434852592406e-05,
|
| 918 |
+
"loss": 0.0008,
|
| 919 |
+
"step": 13000
|
| 920 |
+
},
|
| 921 |
+
{
|
| 922 |
+
"epoch": 1.05,
|
| 923 |
+
"grad_norm": 0.006782655604183674,
|
| 924 |
+
"learning_rate": 1.79207044798354e-05,
|
| 925 |
+
"loss": 0.0004,
|
| 926 |
+
"step": 13100
|
| 927 |
+
},
|
| 928 |
+
{
|
| 929 |
+
"epoch": 1.06,
|
| 930 |
+
"grad_norm": 0.006875937804579735,
|
| 931 |
+
"learning_rate": 1.788977204637388e-05,
|
| 932 |
+
"loss": 0.0027,
|
| 933 |
+
"step": 13200
|
| 934 |
+
},
|
| 935 |
+
{
|
| 936 |
+
"epoch": 1.07,
|
| 937 |
+
"grad_norm": 0.007184536661952734,
|
| 938 |
+
"learning_rate": 1.7858638341308026e-05,
|
| 939 |
+
"loss": 0.0013,
|
| 940 |
+
"step": 13300
|
| 941 |
+
},
|
| 942 |
+
{
|
| 943 |
+
"epoch": 1.08,
|
| 944 |
+
"grad_norm": 0.0019093825249001384,
|
| 945 |
+
"learning_rate": 1.7827304158872538e-05,
|
| 946 |
+
"loss": 0.0014,
|
| 947 |
+
"step": 13400
|
| 948 |
+
},
|
| 949 |
+
{
|
| 950 |
+
"epoch": 1.08,
|
| 951 |
+
"grad_norm": 1.7438774108886719,
|
| 952 |
+
"learning_rate": 1.779577029841638e-05,
|
| 953 |
+
"loss": 0.0027,
|
| 954 |
+
"step": 13500
|
| 955 |
+
},
|
| 956 |
+
{
|
| 957 |
+
"epoch": 1.09,
|
| 958 |
+
"grad_norm": 0.0496426597237587,
|
| 959 |
+
"learning_rate": 1.776403756438241e-05,
|
| 960 |
+
"loss": 0.0016,
|
| 961 |
+
"step": 13600
|
| 962 |
+
},
|
| 963 |
+
{
|
| 964 |
+
"epoch": 1.1,
|
| 965 |
+
"grad_norm": 0.005185109097510576,
|
| 966 |
+
"learning_rate": 1.773210676628682e-05,
|
| 967 |
+
"loss": 0.0004,
|
| 968 |
+
"step": 13700
|
| 969 |
+
},
|
| 970 |
+
{
|
| 971 |
+
"epoch": 1.11,
|
| 972 |
+
"grad_norm": 0.0010710001224651933,
|
| 973 |
+
"learning_rate": 1.769997871869852e-05,
|
| 974 |
+
"loss": 0.0009,
|
| 975 |
+
"step": 13800
|
| 976 |
+
},
|
| 977 |
+
{
|
| 978 |
+
"epoch": 1.12,
|
| 979 |
+
"grad_norm": 0.013586360029876232,
|
| 980 |
+
"learning_rate": 1.7667654241218332e-05,
|
| 981 |
+
"loss": 0.0028,
|
| 982 |
+
"step": 13900
|
| 983 |
+
},
|
| 984 |
+
{
|
| 985 |
+
"epoch": 1.12,
|
| 986 |
+
"grad_norm": 0.021679196506738663,
|
| 987 |
+
"learning_rate": 1.7635134158458095e-05,
|
| 988 |
+
"loss": 0.0011,
|
| 989 |
+
"step": 14000
|
| 990 |
+
},
|
| 991 |
+
{
|
| 992 |
+
"epoch": 1.13,
|
| 993 |
+
"grad_norm": 0.2464137077331543,
|
| 994 |
+
"learning_rate": 1.7602419300019627e-05,
|
| 995 |
+
"loss": 0.0014,
|
| 996 |
+
"step": 14100
|
| 997 |
+
},
|
| 998 |
+
{
|
| 999 |
+
"epoch": 1.14,
|
| 1000 |
+
"grad_norm": 0.0015559865860268474,
|
| 1001 |
+
"learning_rate": 1.7569510500473566e-05,
|
| 1002 |
+
"loss": 0.0003,
|
| 1003 |
+
"step": 14200
|
| 1004 |
+
},
|
| 1005 |
+
{
|
| 1006 |
+
"epoch": 1.15,
|
| 1007 |
+
"grad_norm": 0.007820005528628826,
|
| 1008 |
+
"learning_rate": 1.753640859933806e-05,
|
| 1009 |
+
"loss": 0.001,
|
| 1010 |
+
"step": 14300
|
| 1011 |
+
},
|
| 1012 |
+
{
|
| 1013 |
+
"epoch": 1.16,
|
| 1014 |
+
"grad_norm": 0.0021381524857133627,
|
| 1015 |
+
"learning_rate": 1.7503114441057374e-05,
|
| 1016 |
+
"loss": 0.0019,
|
| 1017 |
+
"step": 14400
|
| 1018 |
+
},
|
| 1019 |
+
{
|
| 1020 |
+
"epoch": 1.16,
|
| 1021 |
+
"grad_norm": 0.004830517340451479,
|
| 1022 |
+
"learning_rate": 1.746962887498034e-05,
|
| 1023 |
+
"loss": 0.0006,
|
| 1024 |
+
"step": 14500
|
| 1025 |
+
},
|
| 1026 |
+
{
|
| 1027 |
+
"epoch": 1.17,
|
| 1028 |
+
"grad_norm": 0.004048046190291643,
|
| 1029 |
+
"learning_rate": 1.743595275533869e-05,
|
| 1030 |
+
"loss": 0.001,
|
| 1031 |
+
"step": 14600
|
| 1032 |
+
},
|
| 1033 |
+
{
|
| 1034 |
+
"epoch": 1.18,
|
| 1035 |
+
"grad_norm": 0.0010604397393763065,
|
| 1036 |
+
"learning_rate": 1.7402086941225246e-05,
|
| 1037 |
+
"loss": 0.0007,
|
| 1038 |
+
"step": 14700
|
| 1039 |
+
},
|
| 1040 |
+
{
|
| 1041 |
+
"epoch": 1.19,
|
| 1042 |
+
"grad_norm": 0.10798583179712296,
|
| 1043 |
+
"learning_rate": 1.736803229657204e-05,
|
| 1044 |
+
"loss": 0.0016,
|
| 1045 |
+
"step": 14800
|
| 1046 |
+
},
|
| 1047 |
+
{
|
| 1048 |
+
"epoch": 1.2,
|
| 1049 |
+
"grad_norm": 0.00545172905549407,
|
| 1050 |
+
"learning_rate": 1.7333789690128252e-05,
|
| 1051 |
+
"loss": 0.0003,
|
| 1052 |
+
"step": 14900
|
| 1053 |
+
},
|
| 1054 |
+
{
|
| 1055 |
+
"epoch": 1.2,
|
| 1056 |
+
"grad_norm": 0.051401443779468536,
|
| 1057 |
+
"learning_rate": 1.7299359995438046e-05,
|
| 1058 |
+
"loss": 0.0007,
|
| 1059 |
+
"step": 15000
|
| 1060 |
+
},
|
| 1061 |
+
{
|
| 1062 |
+
"epoch": 1.21,
|
| 1063 |
+
"grad_norm": 0.0022059613838791847,
|
| 1064 |
+
"learning_rate": 1.7264744090818284e-05,
|
| 1065 |
+
"loss": 0.0009,
|
| 1066 |
+
"step": 15100
|
| 1067 |
+
},
|
| 1068 |
+
{
|
| 1069 |
+
"epoch": 1.22,
|
| 1070 |
+
"grad_norm": 0.0031678322702646255,
|
| 1071 |
+
"learning_rate": 1.7229942859336142e-05,
|
| 1072 |
+
"loss": 0.0003,
|
| 1073 |
+
"step": 15200
|
| 1074 |
+
},
|
| 1075 |
+
{
|
| 1076 |
+
"epoch": 1.23,
|
| 1077 |
+
"grad_norm": 0.4575800597667694,
|
| 1078 |
+
"learning_rate": 1.719495718878655e-05,
|
| 1079 |
+
"loss": 0.0008,
|
| 1080 |
+
"step": 15300
|
| 1081 |
+
},
|
| 1082 |
+
{
|
| 1083 |
+
"epoch": 1.24,
|
| 1084 |
+
"grad_norm": 0.0009642325458116829,
|
| 1085 |
+
"learning_rate": 1.7159787971669586e-05,
|
| 1086 |
+
"loss": 0.001,
|
| 1087 |
+
"step": 15400
|
| 1088 |
+
},
|
| 1089 |
+
{
|
| 1090 |
+
"epoch": 1.24,
|
| 1091 |
+
"grad_norm": 0.0024611325934529305,
|
| 1092 |
+
"learning_rate": 1.712443610516765e-05,
|
| 1093 |
+
"loss": 0.0007,
|
| 1094 |
+
"step": 15500
|
| 1095 |
+
},
|
| 1096 |
+
{
|
| 1097 |
+
"epoch": 1.25,
|
| 1098 |
+
"grad_norm": 0.007828389294445515,
|
| 1099 |
+
"learning_rate": 1.7088902491122636e-05,
|
| 1100 |
+
"loss": 0.0003,
|
| 1101 |
+
"step": 15600
|
| 1102 |
+
},
|
| 1103 |
+
{
|
| 1104 |
+
"epoch": 1.26,
|
| 1105 |
+
"grad_norm": 0.0010621993569657207,
|
| 1106 |
+
"learning_rate": 1.7053188036012885e-05,
|
| 1107 |
+
"loss": 0.0022,
|
| 1108 |
+
"step": 15700
|
| 1109 |
+
},
|
| 1110 |
+
{
|
| 1111 |
+
"epoch": 1.27,
|
| 1112 |
+
"grad_norm": 0.003037663409486413,
|
| 1113 |
+
"learning_rate": 1.7017293650930083e-05,
|
| 1114 |
+
"loss": 0.0011,
|
| 1115 |
+
"step": 15800
|
| 1116 |
+
},
|
| 1117 |
+
{
|
| 1118 |
+
"epoch": 1.28,
|
| 1119 |
+
"grad_norm": 0.021871333941817284,
|
| 1120 |
+
"learning_rate": 1.6981220251555996e-05,
|
| 1121 |
+
"loss": 0.0005,
|
| 1122 |
+
"step": 15900
|
| 1123 |
+
},
|
| 1124 |
+
{
|
| 1125 |
+
"epoch": 1.28,
|
| 1126 |
+
"grad_norm": 0.2676248848438263,
|
| 1127 |
+
"learning_rate": 1.6944968758139144e-05,
|
| 1128 |
+
"loss": 0.0012,
|
| 1129 |
+
"step": 16000
|
| 1130 |
+
},
|
| 1131 |
+
{
|
| 1132 |
+
"epoch": 1.29,
|
| 1133 |
+
"grad_norm": 0.002139603951945901,
|
| 1134 |
+
"learning_rate": 1.6908540095471288e-05,
|
| 1135 |
+
"loss": 0.0016,
|
| 1136 |
+
"step": 16100
|
| 1137 |
+
},
|
| 1138 |
+
{
|
| 1139 |
+
"epoch": 1.3,
|
| 1140 |
+
"grad_norm": 0.00880183931440115,
|
| 1141 |
+
"learning_rate": 1.6871935192863862e-05,
|
| 1142 |
+
"loss": 0.0007,
|
| 1143 |
+
"step": 16200
|
| 1144 |
+
},
|
| 1145 |
+
{
|
| 1146 |
+
"epoch": 1.31,
|
| 1147 |
+
"grad_norm": 0.006027602590620518,
|
| 1148 |
+
"learning_rate": 1.6835154984124266e-05,
|
| 1149 |
+
"loss": 0.0027,
|
| 1150 |
+
"step": 16300
|
| 1151 |
+
},
|
| 1152 |
+
{
|
| 1153 |
+
"epoch": 1.32,
|
| 1154 |
+
"grad_norm": 0.002160316100344062,
|
| 1155 |
+
"learning_rate": 1.6798200407532025e-05,
|
| 1156 |
+
"loss": 0.0004,
|
| 1157 |
+
"step": 16400
|
| 1158 |
+
},
|
| 1159 |
+
{
|
| 1160 |
+
"epoch": 1.32,
|
| 1161 |
+
"grad_norm": 0.004634434822946787,
|
| 1162 |
+
"learning_rate": 1.676107240581488e-05,
|
| 1163 |
+
"loss": 0.0005,
|
| 1164 |
+
"step": 16500
|
| 1165 |
+
},
|
| 1166 |
+
{
|
| 1167 |
+
"epoch": 1.33,
|
| 1168 |
+
"grad_norm": 0.013142119161784649,
|
| 1169 |
+
"learning_rate": 1.6723771926124704e-05,
|
| 1170 |
+
"loss": 0.0012,
|
| 1171 |
+
"step": 16600
|
| 1172 |
+
},
|
| 1173 |
+
{
|
| 1174 |
+
"epoch": 1.34,
|
| 1175 |
+
"grad_norm": 0.0017228337237611413,
|
| 1176 |
+
"learning_rate": 1.6686299920013388e-05,
|
| 1177 |
+
"loss": 0.001,
|
| 1178 |
+
"step": 16700
|
| 1179 |
+
},
|
| 1180 |
+
{
|
| 1181 |
+
"epoch": 1.35,
|
| 1182 |
+
"grad_norm": 0.011891527101397514,
|
| 1183 |
+
"learning_rate": 1.6648657343408517e-05,
|
| 1184 |
+
"loss": 0.0004,
|
| 1185 |
+
"step": 16800
|
| 1186 |
+
},
|
| 1187 |
+
{
|
| 1188 |
+
"epoch": 1.36,
|
| 1189 |
+
"grad_norm": 0.03485196456313133,
|
| 1190 |
+
"learning_rate": 1.661084515658901e-05,
|
| 1191 |
+
"loss": 0.0008,
|
| 1192 |
+
"step": 16900
|
| 1193 |
+
},
|
| 1194 |
+
{
|
| 1195 |
+
"epoch": 1.36,
|
| 1196 |
+
"grad_norm": 0.006045693065971136,
|
| 1197 |
+
"learning_rate": 1.6572864324160617e-05,
|
| 1198 |
+
"loss": 0.0016,
|
| 1199 |
+
"step": 17000
|
| 1200 |
+
},
|
| 1201 |
+
{
|
| 1202 |
+
"epoch": 1.37,
|
| 1203 |
+
"grad_norm": 0.0319632813334465,
|
| 1204 |
+
"learning_rate": 1.6534715815031325e-05,
|
| 1205 |
+
"loss": 0.001,
|
| 1206 |
+
"step": 17100
|
| 1207 |
+
},
|
| 1208 |
+
{
|
| 1209 |
+
"epoch": 1.38,
|
| 1210 |
+
"grad_norm": 0.000651439419016242,
|
| 1211 |
+
"learning_rate": 1.649640060238661e-05,
|
| 1212 |
+
"loss": 0.0008,
|
| 1213 |
+
"step": 17200
|
| 1214 |
+
},
|
| 1215 |
+
{
|
| 1216 |
+
"epoch": 1.39,
|
| 1217 |
+
"grad_norm": 0.013289058580994606,
|
| 1218 |
+
"learning_rate": 1.645791966366464e-05,
|
| 1219 |
+
"loss": 0.0004,
|
| 1220 |
+
"step": 17300
|
| 1221 |
+
},
|
| 1222 |
+
{
|
| 1223 |
+
"epoch": 1.4,
|
| 1224 |
+
"grad_norm": 1.1903555393218994,
|
| 1225 |
+
"learning_rate": 1.6419273980531333e-05,
|
| 1226 |
+
"loss": 0.0009,
|
| 1227 |
+
"step": 17400
|
| 1228 |
+
},
|
| 1229 |
+
{
|
| 1230 |
+
"epoch": 1.4,
|
| 1231 |
+
"grad_norm": 0.0004181715485174209,
|
| 1232 |
+
"learning_rate": 1.63804645388553e-05,
|
| 1233 |
+
"loss": 0.0008,
|
| 1234 |
+
"step": 17500
|
| 1235 |
+
},
|
| 1236 |
+
{
|
| 1237 |
+
"epoch": 1.41,
|
| 1238 |
+
"grad_norm": 0.0005597823183052242,
|
| 1239 |
+
"learning_rate": 1.6341492328682703e-05,
|
| 1240 |
+
"loss": 0.0015,
|
| 1241 |
+
"step": 17600
|
| 1242 |
+
},
|
| 1243 |
+
{
|
| 1244 |
+
"epoch": 1.42,
|
| 1245 |
+
"grad_norm": 0.003250558627769351,
|
| 1246 |
+
"learning_rate": 1.6302358344212025e-05,
|
| 1247 |
+
"loss": 0.0015,
|
| 1248 |
+
"step": 17700
|
| 1249 |
+
},
|
| 1250 |
+
{
|
| 1251 |
+
"epoch": 1.43,
|
| 1252 |
+
"grad_norm": 0.00633718678727746,
|
| 1253 |
+
"learning_rate": 1.6263063583768652e-05,
|
| 1254 |
+
"loss": 0.0007,
|
| 1255 |
+
"step": 17800
|
| 1256 |
+
},
|
| 1257 |
+
{
|
| 1258 |
+
"epoch": 1.44,
|
| 1259 |
+
"grad_norm": 0.0012848949991166592,
|
| 1260 |
+
"learning_rate": 1.622360904977946e-05,
|
| 1261 |
+
"loss": 0.001,
|
| 1262 |
+
"step": 17900
|
| 1263 |
+
},
|
| 1264 |
+
{
|
| 1265 |
+
"epoch": 1.44,
|
| 1266 |
+
"grad_norm": 0.03510229289531708,
|
| 1267 |
+
"learning_rate": 1.6183995748747204e-05,
|
| 1268 |
+
"loss": 0.001,
|
| 1269 |
+
"step": 18000
|
| 1270 |
+
},
|
| 1271 |
+
{
|
| 1272 |
+
"epoch": 1.45,
|
| 1273 |
+
"grad_norm": 0.00044461427023634315,
|
| 1274 |
+
"learning_rate": 1.6144224691224868e-05,
|
| 1275 |
+
"loss": 0.0005,
|
| 1276 |
+
"step": 18100
|
| 1277 |
+
},
|
| 1278 |
+
{
|
| 1279 |
+
"epoch": 1.46,
|
| 1280 |
+
"grad_norm": 0.056695085018873215,
|
| 1281 |
+
"learning_rate": 1.6104296891789867e-05,
|
| 1282 |
+
"loss": 0.0011,
|
| 1283 |
+
"step": 18200
|
| 1284 |
+
},
|
| 1285 |
+
{
|
| 1286 |
+
"epoch": 1.47,
|
| 1287 |
+
"grad_norm": 0.002499540336430073,
|
| 1288 |
+
"learning_rate": 1.606421336901818e-05,
|
| 1289 |
+
"loss": 0.0005,
|
| 1290 |
+
"step": 18300
|
| 1291 |
+
},
|
| 1292 |
+
{
|
| 1293 |
+
"epoch": 1.48,
|
| 1294 |
+
"grad_norm": 0.00433159526437521,
|
| 1295 |
+
"learning_rate": 1.6023975145458352e-05,
|
| 1296 |
+
"loss": 0.0007,
|
| 1297 |
+
"step": 18400
|
| 1298 |
+
},
|
| 1299 |
+
{
|
| 1300 |
+
"epoch": 1.48,
|
| 1301 |
+
"grad_norm": 0.0020198116544634104,
|
| 1302 |
+
"learning_rate": 1.5983583247605414e-05,
|
| 1303 |
+
"loss": 0.0005,
|
| 1304 |
+
"step": 18500
|
| 1305 |
+
},
|
| 1306 |
+
{
|
| 1307 |
+
"epoch": 1.49,
|
| 1308 |
+
"grad_norm": 0.004110960755497217,
|
| 1309 |
+
"learning_rate": 1.5943038705874697e-05,
|
| 1310 |
+
"loss": 0.001,
|
| 1311 |
+
"step": 18600
|
| 1312 |
+
},
|
| 1313 |
+
{
|
| 1314 |
+
"epoch": 1.5,
|
| 1315 |
+
"grad_norm": 0.0036564720794558525,
|
| 1316 |
+
"learning_rate": 1.590234255457555e-05,
|
| 1317 |
+
"loss": 0.0014,
|
| 1318 |
+
"step": 18700
|
| 1319 |
+
},
|
| 1320 |
+
{
|
| 1321 |
+
"epoch": 1.51,
|
| 1322 |
+
"grad_norm": 0.0010110210860148072,
|
| 1323 |
+
"learning_rate": 1.5861495831884942e-05,
|
| 1324 |
+
"loss": 0.0008,
|
| 1325 |
+
"step": 18800
|
| 1326 |
+
},
|
| 1327 |
+
{
|
| 1328 |
+
"epoch": 1.52,
|
| 1329 |
+
"grad_norm": 0.11147645115852356,
|
| 1330 |
+
"learning_rate": 1.582049957982099e-05,
|
| 1331 |
+
"loss": 0.0004,
|
| 1332 |
+
"step": 18900
|
| 1333 |
+
},
|
| 1334 |
+
{
|
| 1335 |
+
"epoch": 1.52,
|
| 1336 |
+
"grad_norm": 0.002563257934525609,
|
| 1337 |
+
"learning_rate": 1.5779354844216377e-05,
|
| 1338 |
+
"loss": 0.0003,
|
| 1339 |
+
"step": 19000
|
| 1340 |
+
},
|
| 1341 |
+
{
|
| 1342 |
+
"epoch": 1.53,
|
| 1343 |
+
"grad_norm": 0.002472821157425642,
|
| 1344 |
+
"learning_rate": 1.5738062674691657e-05,
|
| 1345 |
+
"loss": 0.0003,
|
| 1346 |
+
"step": 19100
|
| 1347 |
+
},
|
| 1348 |
+
{
|
| 1349 |
+
"epoch": 1.54,
|
| 1350 |
+
"grad_norm": 0.08172155171632767,
|
| 1351 |
+
"learning_rate": 1.5696624124628495e-05,
|
| 1352 |
+
"loss": 0.0005,
|
| 1353 |
+
"step": 19200
|
| 1354 |
+
},
|
| 1355 |
+
{
|
| 1356 |
+
"epoch": 1.55,
|
| 1357 |
+
"grad_norm": 0.009228968061506748,
|
| 1358 |
+
"learning_rate": 1.5655040251142787e-05,
|
| 1359 |
+
"loss": 0.0008,
|
| 1360 |
+
"step": 19300
|
| 1361 |
+
},
|
| 1362 |
+
{
|
| 1363 |
+
"epoch": 1.56,
|
| 1364 |
+
"grad_norm": 0.027170058339834213,
|
| 1365 |
+
"learning_rate": 1.5613312115057697e-05,
|
| 1366 |
+
"loss": 0.0002,
|
| 1367 |
+
"step": 19400
|
| 1368 |
+
},
|
| 1369 |
+
{
|
| 1370 |
+
"epoch": 1.57,
|
| 1371 |
+
"grad_norm": 0.003953781444579363,
|
| 1372 |
+
"learning_rate": 1.5571440780876588e-05,
|
| 1373 |
+
"loss": 0.0009,
|
| 1374 |
+
"step": 19500
|
| 1375 |
+
},
|
| 1376 |
+
{
|
| 1377 |
+
"epoch": 1.57,
|
| 1378 |
+
"grad_norm": 0.2073652297258377,
|
| 1379 |
+
"learning_rate": 1.5529427316755876e-05,
|
| 1380 |
+
"loss": 0.001,
|
| 1381 |
+
"step": 19600
|
| 1382 |
+
},
|
| 1383 |
+
{
|
| 1384 |
+
"epoch": 1.58,
|
| 1385 |
+
"grad_norm": 0.000581290340051055,
|
| 1386 |
+
"learning_rate": 1.548727279447777e-05,
|
| 1387 |
+
"loss": 0.0007,
|
| 1388 |
+
"step": 19700
|
| 1389 |
+
},
|
| 1390 |
+
{
|
| 1391 |
+
"epoch": 1.59,
|
| 1392 |
+
"grad_norm": 0.001082689268514514,
|
| 1393 |
+
"learning_rate": 1.5444978289422937e-05,
|
| 1394 |
+
"loss": 0.0011,
|
| 1395 |
+
"step": 19800
|
| 1396 |
+
},
|
| 1397 |
+
{
|
| 1398 |
+
"epoch": 1.6,
|
| 1399 |
+
"grad_norm": 0.020316652953624725,
|
| 1400 |
+
"learning_rate": 1.540254488054307e-05,
|
| 1401 |
+
"loss": 0.0007,
|
| 1402 |
+
"step": 19900
|
| 1403 |
+
},
|
| 1404 |
+
{
|
| 1405 |
+
"epoch": 1.61,
|
| 1406 |
+
"grad_norm": 0.0026674780528992414,
|
| 1407 |
+
"learning_rate": 1.5359973650333352e-05,
|
| 1408 |
+
"loss": 0.0006,
|
| 1409 |
+
"step": 20000
|
| 1410 |
+
},
|
| 1411 |
+
{
|
| 1412 |
+
"epoch": 1.61,
|
| 1413 |
+
"grad_norm": 0.007332425098866224,
|
| 1414 |
+
"learning_rate": 1.5317265684804865e-05,
|
| 1415 |
+
"loss": 0.001,
|
| 1416 |
+
"step": 20100
|
| 1417 |
+
},
|
| 1418 |
+
{
|
| 1419 |
+
"epoch": 1.62,
|
| 1420 |
+
"grad_norm": 0.0030178299639374018,
|
| 1421 |
+
"learning_rate": 1.5274422073456853e-05,
|
| 1422 |
+
"loss": 0.0002,
|
| 1423 |
+
"step": 20200
|
| 1424 |
+
},
|
| 1425 |
+
{
|
| 1426 |
+
"epoch": 1.63,
|
| 1427 |
+
"grad_norm": 0.0009229824645444751,
|
| 1428 |
+
"learning_rate": 1.5231443909248956e-05,
|
| 1429 |
+
"loss": 0.0006,
|
| 1430 |
+
"step": 20300
|
| 1431 |
+
},
|
| 1432 |
+
{
|
| 1433 |
+
"epoch": 1.64,
|
| 1434 |
+
"grad_norm": 0.000619510596152395,
|
| 1435 |
+
"learning_rate": 1.5188332288573313e-05,
|
| 1436 |
+
"loss": 0.0011,
|
| 1437 |
+
"step": 20400
|
| 1438 |
+
},
|
| 1439 |
+
{
|
| 1440 |
+
"epoch": 1.65,
|
| 1441 |
+
"grad_norm": 0.0273954626172781,
|
| 1442 |
+
"learning_rate": 1.5145088311226599e-05,
|
| 1443 |
+
"loss": 0.0004,
|
| 1444 |
+
"step": 20500
|
| 1445 |
+
},
|
| 1446 |
+
{
|
| 1447 |
+
"epoch": 1.65,
|
| 1448 |
+
"grad_norm": 0.009902927093207836,
|
| 1449 |
+
"learning_rate": 1.510171308038197e-05,
|
| 1450 |
+
"loss": 0.0006,
|
| 1451 |
+
"step": 20600
|
| 1452 |
+
},
|
| 1453 |
+
{
|
| 1454 |
+
"epoch": 1.66,
|
| 1455 |
+
"grad_norm": 0.012104487977921963,
|
| 1456 |
+
"learning_rate": 1.5058207702560907e-05,
|
| 1457 |
+
"loss": 0.0004,
|
| 1458 |
+
"step": 20700
|
| 1459 |
+
},
|
| 1460 |
+
{
|
| 1461 |
+
"epoch": 1.67,
|
| 1462 |
+
"grad_norm": 0.8639243245124817,
|
| 1463 |
+
"learning_rate": 1.501457328760501e-05,
|
| 1464 |
+
"loss": 0.0007,
|
| 1465 |
+
"step": 20800
|
| 1466 |
+
},
|
| 1467 |
+
{
|
| 1468 |
+
"epoch": 1.68,
|
| 1469 |
+
"grad_norm": 0.07374394685029984,
|
| 1470 |
+
"learning_rate": 1.4970810948647664e-05,
|
| 1471 |
+
"loss": 0.0007,
|
| 1472 |
+
"step": 20900
|
| 1473 |
+
},
|
| 1474 |
+
{
|
| 1475 |
+
"epoch": 1.69,
|
| 1476 |
+
"grad_norm": 0.0010848238598555326,
|
| 1477 |
+
"learning_rate": 1.4926921802085662e-05,
|
| 1478 |
+
"loss": 0.0008,
|
| 1479 |
+
"step": 21000
|
| 1480 |
+
},
|
| 1481 |
+
{
|
| 1482 |
+
"epoch": 1.69,
|
| 1483 |
+
"grad_norm": 0.03615230694413185,
|
| 1484 |
+
"learning_rate": 1.4882906967550708e-05,
|
| 1485 |
+
"loss": 0.0002,
|
| 1486 |
+
"step": 21100
|
| 1487 |
+
},
|
| 1488 |
+
{
|
| 1489 |
+
"epoch": 1.7,
|
| 1490 |
+
"grad_norm": 0.00893151480704546,
|
| 1491 |
+
"learning_rate": 1.4838767567880865e-05,
|
| 1492 |
+
"loss": 0.0012,
|
| 1493 |
+
"step": 21200
|
| 1494 |
+
},
|
| 1495 |
+
{
|
| 1496 |
+
"epoch": 1.71,
|
| 1497 |
+
"grad_norm": 0.0023355227895081043,
|
| 1498 |
+
"learning_rate": 1.479450472909191e-05,
|
| 1499 |
+
"loss": 0.0023,
|
| 1500 |
+
"step": 21300
|
| 1501 |
+
},
|
| 1502 |
+
{
|
| 1503 |
+
"epoch": 1.72,
|
| 1504 |
+
"grad_norm": 0.7856515645980835,
|
| 1505 |
+
"learning_rate": 1.4750119580348601e-05,
|
| 1506 |
+
"loss": 0.0008,
|
| 1507 |
+
"step": 21400
|
| 1508 |
+
},
|
| 1509 |
+
{
|
| 1510 |
+
"epoch": 1.73,
|
| 1511 |
+
"grad_norm": 0.0007943073869682848,
|
| 1512 |
+
"learning_rate": 1.4705613253935886e-05,
|
| 1513 |
+
"loss": 0.0002,
|
| 1514 |
+
"step": 21500
|
| 1515 |
+
},
|
| 1516 |
+
{
|
| 1517 |
+
"epoch": 1.73,
|
| 1518 |
+
"grad_norm": 0.000991353183053434,
|
| 1519 |
+
"learning_rate": 1.4660986885230002e-05,
|
| 1520 |
+
"loss": 0.0005,
|
| 1521 |
+
"step": 21600
|
| 1522 |
+
},
|
| 1523 |
+
{
|
| 1524 |
+
"epoch": 1.74,
|
| 1525 |
+
"grad_norm": 0.0023457545321434736,
|
| 1526 |
+
"learning_rate": 1.4616241612669523e-05,
|
| 1527 |
+
"loss": 0.0035,
|
| 1528 |
+
"step": 21700
|
| 1529 |
+
},
|
| 1530 |
+
{
|
| 1531 |
+
"epoch": 1.75,
|
| 1532 |
+
"grad_norm": 0.0007867334061302245,
|
| 1533 |
+
"learning_rate": 1.4571378577726317e-05,
|
| 1534 |
+
"loss": 0.0006,
|
| 1535 |
+
"step": 21800
|
| 1536 |
+
},
|
| 1537 |
+
{
|
| 1538 |
+
"epoch": 1.76,
|
| 1539 |
+
"grad_norm": 0.0013897059252485633,
|
| 1540 |
+
"learning_rate": 1.4526398924876407e-05,
|
| 1541 |
+
"loss": 0.0005,
|
| 1542 |
+
"step": 21900
|
| 1543 |
+
},
|
| 1544 |
+
{
|
| 1545 |
+
"epoch": 1.77,
|
| 1546 |
+
"grad_norm": 0.0020729138050228357,
|
| 1547 |
+
"learning_rate": 1.4481303801570805e-05,
|
| 1548 |
+
"loss": 0.0009,
|
| 1549 |
+
"step": 22000
|
| 1550 |
+
},
|
| 1551 |
+
{
|
| 1552 |
+
"epoch": 1.77,
|
| 1553 |
+
"grad_norm": 0.002158819232136011,
|
| 1554 |
+
"learning_rate": 1.4436094358206224e-05,
|
| 1555 |
+
"loss": 0.0005,
|
| 1556 |
+
"step": 22100
|
| 1557 |
+
},
|
| 1558 |
+
{
|
| 1559 |
+
"epoch": 1.78,
|
| 1560 |
+
"grad_norm": 0.0038599702529609203,
|
| 1561 |
+
"learning_rate": 1.4390771748095735e-05,
|
| 1562 |
+
"loss": 0.0007,
|
| 1563 |
+
"step": 22200
|
| 1564 |
+
},
|
| 1565 |
+
{
|
| 1566 |
+
"epoch": 1.79,
|
| 1567 |
+
"grad_norm": 0.0022886607330292463,
|
| 1568 |
+
"learning_rate": 1.4345337127439333e-05,
|
| 1569 |
+
"loss": 0.0002,
|
| 1570 |
+
"step": 22300
|
| 1571 |
+
},
|
| 1572 |
+
{
|
| 1573 |
+
"epoch": 1.8,
|
| 1574 |
+
"grad_norm": 0.011882874183356762,
|
| 1575 |
+
"learning_rate": 1.4299791655294461e-05,
|
| 1576 |
+
"loss": 0.0021,
|
| 1577 |
+
"step": 22400
|
| 1578 |
+
},
|
| 1579 |
+
{
|
| 1580 |
+
"epoch": 1.81,
|
| 1581 |
+
"grad_norm": 0.011625411920249462,
|
| 1582 |
+
"learning_rate": 1.4254136493546432e-05,
|
| 1583 |
+
"loss": 0.0003,
|
| 1584 |
+
"step": 22500
|
| 1585 |
+
},
|
| 1586 |
+
{
|
| 1587 |
+
"epoch": 1.81,
|
| 1588 |
+
"grad_norm": 0.004132369067519903,
|
| 1589 |
+
"learning_rate": 1.4208372806878782e-05,
|
| 1590 |
+
"loss": 0.0004,
|
| 1591 |
+
"step": 22600
|
| 1592 |
+
},
|
| 1593 |
+
{
|
| 1594 |
+
"epoch": 1.82,
|
| 1595 |
+
"grad_norm": 0.0008625888731330633,
|
| 1596 |
+
"learning_rate": 1.4162501762743579e-05,
|
| 1597 |
+
"loss": 0.0013,
|
| 1598 |
+
"step": 22700
|
| 1599 |
+
},
|
| 1600 |
+
{
|
| 1601 |
+
"epoch": 1.83,
|
| 1602 |
+
"grad_norm": 0.02038603462278843,
|
| 1603 |
+
"learning_rate": 1.4116524531331616e-05,
|
| 1604 |
+
"loss": 0.001,
|
| 1605 |
+
"step": 22800
|
| 1606 |
+
},
|
| 1607 |
+
{
|
| 1608 |
+
"epoch": 1.84,
|
| 1609 |
+
"grad_norm": 0.021469444036483765,
|
| 1610 |
+
"learning_rate": 1.4070442285542579e-05,
|
| 1611 |
+
"loss": 0.0004,
|
| 1612 |
+
"step": 22900
|
| 1613 |
+
},
|
| 1614 |
+
{
|
| 1615 |
+
"epoch": 1.85,
|
| 1616 |
+
"grad_norm": 0.7221536040306091,
|
| 1617 |
+
"learning_rate": 1.402425620095511e-05,
|
| 1618 |
+
"loss": 0.0007,
|
| 1619 |
+
"step": 23000
|
| 1620 |
+
},
|
| 1621 |
+
{
|
| 1622 |
+
"epoch": 1.85,
|
| 1623 |
+
"grad_norm": 0.0006885859766043723,
|
| 1624 |
+
"learning_rate": 1.3977967455796828e-05,
|
| 1625 |
+
"loss": 0.0009,
|
| 1626 |
+
"step": 23100
|
| 1627 |
+
},
|
| 1628 |
+
{
|
| 1629 |
+
"epoch": 1.86,
|
| 1630 |
+
"grad_norm": 0.0010358589934185147,
|
| 1631 |
+
"learning_rate": 1.393157723091428e-05,
|
| 1632 |
+
"loss": 0.0003,
|
| 1633 |
+
"step": 23200
|
| 1634 |
+
},
|
| 1635 |
+
{
|
| 1636 |
+
"epoch": 1.87,
|
| 1637 |
+
"grad_norm": 0.0008139715064316988,
|
| 1638 |
+
"learning_rate": 1.3885086709742788e-05,
|
| 1639 |
+
"loss": 0.0005,
|
| 1640 |
+
"step": 23300
|
| 1641 |
+
},
|
| 1642 |
+
{
|
| 1643 |
+
"epoch": 1.88,
|
| 1644 |
+
"grad_norm": 0.6740529537200928,
|
| 1645 |
+
"learning_rate": 1.3838497078276288e-05,
|
| 1646 |
+
"loss": 0.0018,
|
| 1647 |
+
"step": 23400
|
| 1648 |
+
},
|
| 1649 |
+
{
|
| 1650 |
+
"epoch": 1.89,
|
| 1651 |
+
"grad_norm": 0.0022279045078903437,
|
| 1652 |
+
"learning_rate": 1.3791809525037057e-05,
|
| 1653 |
+
"loss": 0.0005,
|
| 1654 |
+
"step": 23500
|
| 1655 |
+
},
|
| 1656 |
+
{
|
| 1657 |
+
"epoch": 1.89,
|
| 1658 |
+
"grad_norm": 0.0020128132309764624,
|
| 1659 |
+
"learning_rate": 1.3745025241045414e-05,
|
| 1660 |
+
"loss": 0.0002,
|
| 1661 |
+
"step": 23600
|
| 1662 |
+
},
|
| 1663 |
+
{
|
| 1664 |
+
"epoch": 1.9,
|
| 1665 |
+
"grad_norm": 0.000531968311406672,
|
| 1666 |
+
"learning_rate": 1.3698145419789302e-05,
|
| 1667 |
+
"loss": 0.0007,
|
| 1668 |
+
"step": 23700
|
| 1669 |
+
},
|
| 1670 |
+
{
|
| 1671 |
+
"epoch": 1.91,
|
| 1672 |
+
"grad_norm": 0.00039000410470180213,
|
| 1673 |
+
"learning_rate": 1.3651171257193883e-05,
|
| 1674 |
+
"loss": 0.0006,
|
| 1675 |
+
"step": 23800
|
| 1676 |
+
},
|
| 1677 |
+
{
|
| 1678 |
+
"epoch": 1.92,
|
| 1679 |
+
"grad_norm": 0.041364409029483795,
|
| 1680 |
+
"learning_rate": 1.3604103951590993e-05,
|
| 1681 |
+
"loss": 0.0003,
|
| 1682 |
+
"step": 23900
|
| 1683 |
+
},
|
| 1684 |
+
{
|
| 1685 |
+
"epoch": 1.93,
|
| 1686 |
+
"grad_norm": 0.0011203879257664084,
|
| 1687 |
+
"learning_rate": 1.3556944703688592e-05,
|
| 1688 |
+
"loss": 0.0003,
|
| 1689 |
+
"step": 24000
|
| 1690 |
+
},
|
| 1691 |
+
{
|
| 1692 |
+
"epoch": 1.93,
|
| 1693 |
+
"grad_norm": 0.001194770447909832,
|
| 1694 |
+
"learning_rate": 1.3509694716540135e-05,
|
| 1695 |
+
"loss": 0.0002,
|
| 1696 |
+
"step": 24100
|
| 1697 |
+
},
|
| 1698 |
+
{
|
| 1699 |
+
"epoch": 1.94,
|
| 1700 |
+
"grad_norm": 0.002378718461841345,
|
| 1701 |
+
"learning_rate": 1.3462355195513868e-05,
|
| 1702 |
+
"loss": 0.0006,
|
| 1703 |
+
"step": 24200
|
| 1704 |
+
},
|
| 1705 |
+
{
|
| 1706 |
+
"epoch": 1.95,
|
| 1707 |
+
"grad_norm": 0.32328903675079346,
|
| 1708 |
+
"learning_rate": 1.341492734826209e-05,
|
| 1709 |
+
"loss": 0.001,
|
| 1710 |
+
"step": 24300
|
| 1711 |
+
},
|
| 1712 |
+
{
|
| 1713 |
+
"epoch": 1.96,
|
| 1714 |
+
"grad_norm": 0.002817463595420122,
|
| 1715 |
+
"learning_rate": 1.3367412384690346e-05,
|
| 1716 |
+
"loss": 0.0017,
|
| 1717 |
+
"step": 24400
|
| 1718 |
+
},
|
| 1719 |
+
{
|
| 1720 |
+
"epoch": 1.97,
|
| 1721 |
+
"grad_norm": 0.0024801292456686497,
|
| 1722 |
+
"learning_rate": 1.3319811516926541e-05,
|
| 1723 |
+
"loss": 0.0005,
|
| 1724 |
+
"step": 24500
|
| 1725 |
+
},
|
| 1726 |
+
{
|
| 1727 |
+
"epoch": 1.97,
|
| 1728 |
+
"grad_norm": 0.004569421522319317,
|
| 1729 |
+
"learning_rate": 1.3272125959290059e-05,
|
| 1730 |
+
"loss": 0.0008,
|
| 1731 |
+
"step": 24600
|
| 1732 |
+
},
|
| 1733 |
+
{
|
| 1734 |
+
"epoch": 1.98,
|
| 1735 |
+
"grad_norm": 0.1658785045146942,
|
| 1736 |
+
"learning_rate": 1.3224356928260735e-05,
|
| 1737 |
+
"loss": 0.0005,
|
| 1738 |
+
"step": 24700
|
| 1739 |
+
},
|
| 1740 |
+
{
|
| 1741 |
+
"epoch": 1.99,
|
| 1742 |
+
"grad_norm": 0.001219902653247118,
|
| 1743 |
+
"learning_rate": 1.317650564244787e-05,
|
| 1744 |
+
"loss": 0.0005,
|
| 1745 |
+
"step": 24800
|
| 1746 |
+
},
|
| 1747 |
+
{
|
| 1748 |
+
"epoch": 2.0,
|
| 1749 |
+
"grad_norm": 0.002142696175724268,
|
| 1750 |
+
"learning_rate": 1.3128573322559097e-05,
|
| 1751 |
+
"loss": 0.0022,
|
| 1752 |
+
"step": 24900
|
| 1753 |
+
}
|
| 1754 |
+
],
|
| 1755 |
+
"logging_steps": 100,
|
| 1756 |
+
"max_steps": 62300,
|
| 1757 |
+
"num_input_tokens_seen": 0,
|
| 1758 |
+
"num_train_epochs": 5,
|
| 1759 |
+
"save_steps": 500,
|
| 1760 |
+
"total_flos": 1.706415322300416e+16,
|
| 1761 |
+
"train_batch_size": 1,
|
| 1762 |
+
"trial_name": null,
|
| 1763 |
+
"trial_params": null
|
| 1764 |
+
}
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f238f91742ac8f2503b69da046cf2a89c46a16fd4e8d337a4c3ce135ac08f2ea
|
| 3 |
+
size 4984
|