Upload 8 files
Browse files- config.json +24 -0
- generation_config.json +5 -0
- model.safetensors +3 -0
- optimizer.pt +3 -0
- rng_state.pth +3 -0
- scheduler.pt +3 -0
- trainer_state.json +3339 -0
- training_args.bin +3 -0
config.json
ADDED
|
@@ -0,0 +1,24 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"architectures": [
|
| 3 |
+
"BertForMaskedLM"
|
| 4 |
+
],
|
| 5 |
+
"attention_probs_dropout_prob": 0.1,
|
| 6 |
+
"classifier_dropout": null,
|
| 7 |
+
"hidden_act": "gelu",
|
| 8 |
+
"hidden_dropout_prob": 0.1,
|
| 9 |
+
"hidden_size": 128,
|
| 10 |
+
"initializer_range": 0.02,
|
| 11 |
+
"intermediate_size": 3072,
|
| 12 |
+
"layer_norm_eps": 1e-12,
|
| 13 |
+
"max_position_embeddings": 512,
|
| 14 |
+
"model_type": "bert",
|
| 15 |
+
"num_attention_heads": 2,
|
| 16 |
+
"num_hidden_layers": 2,
|
| 17 |
+
"pad_token_id": 0,
|
| 18 |
+
"position_embedding_type": "absolute",
|
| 19 |
+
"torch_dtype": "float32",
|
| 20 |
+
"transformers_version": "4.38.2",
|
| 21 |
+
"type_vocab_size": 1,
|
| 22 |
+
"use_cache": true,
|
| 23 |
+
"vocab_size": 100000
|
| 24 |
+
}
|
generation_config.json
ADDED
|
@@ -0,0 +1,5 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_from_model_config": true,
|
| 3 |
+
"pad_token_id": 0,
|
| 4 |
+
"transformers_version": "4.38.2"
|
| 5 |
+
}
|
model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:213c06ef73e9a0e34189a9ac9ccd62585130a2b704f5fd18417bf903a9ddbe91
|
| 3 |
+
size 58785064
|
optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c173e6f2bf00016904b77a06177aea59ae615ff5e89e243ffe095b0080d828cf
|
| 3 |
+
size 117595461
|
rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:8796754fbd3d01e2dfc98f1f0acdbd72cd6e20379aeeebdebc53f5a947654187
|
| 3 |
+
size 14575
|
scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:962cb1e11198863423869361188d67ff5fef43a94fe9f1255e31d9fc95b77cd6
|
| 3 |
+
size 627
|
trainer_state.json
ADDED
|
@@ -0,0 +1,3339 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 0.13198192806300124,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 237060,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.0,
|
| 13 |
+
"grad_norm": 2.6396658420562744,
|
| 14 |
+
"learning_rate": 5e-06,
|
| 15 |
+
"loss": 11.2923,
|
| 16 |
+
"step": 500
|
| 17 |
+
},
|
| 18 |
+
{
|
| 19 |
+
"epoch": 0.0,
|
| 20 |
+
"grad_norm": 3.1433205604553223,
|
| 21 |
+
"learning_rate": 1e-05,
|
| 22 |
+
"loss": 10.5195,
|
| 23 |
+
"step": 1000
|
| 24 |
+
},
|
| 25 |
+
{
|
| 26 |
+
"epoch": 0.0,
|
| 27 |
+
"grad_norm": 2.825376510620117,
|
| 28 |
+
"learning_rate": 1.5e-05,
|
| 29 |
+
"loss": 9.4381,
|
| 30 |
+
"step": 1500
|
| 31 |
+
},
|
| 32 |
+
{
|
| 33 |
+
"epoch": 0.0,
|
| 34 |
+
"grad_norm": 6.385465621948242,
|
| 35 |
+
"learning_rate": 2e-05,
|
| 36 |
+
"loss": 7.8175,
|
| 37 |
+
"step": 2000
|
| 38 |
+
},
|
| 39 |
+
{
|
| 40 |
+
"epoch": 0.0,
|
| 41 |
+
"grad_norm": 2.1213717460632324,
|
| 42 |
+
"learning_rate": 2.5e-05,
|
| 43 |
+
"loss": 5.8245,
|
| 44 |
+
"step": 2500
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.0,
|
| 48 |
+
"grad_norm": 1.302095890045166,
|
| 49 |
+
"learning_rate": 3e-05,
|
| 50 |
+
"loss": 4.455,
|
| 51 |
+
"step": 3000
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.0,
|
| 55 |
+
"grad_norm": 3.764803886413574,
|
| 56 |
+
"learning_rate": 3.5e-05,
|
| 57 |
+
"loss": 4.1131,
|
| 58 |
+
"step": 3500
|
| 59 |
+
},
|
| 60 |
+
{
|
| 61 |
+
"epoch": 0.0,
|
| 62 |
+
"grad_norm": 1.2695655822753906,
|
| 63 |
+
"learning_rate": 4e-05,
|
| 64 |
+
"loss": 4.0351,
|
| 65 |
+
"step": 4000
|
| 66 |
+
},
|
| 67 |
+
{
|
| 68 |
+
"epoch": 0.0,
|
| 69 |
+
"grad_norm": 1.787291169166565,
|
| 70 |
+
"learning_rate": 4.5e-05,
|
| 71 |
+
"loss": 3.9797,
|
| 72 |
+
"step": 4500
|
| 73 |
+
},
|
| 74 |
+
{
|
| 75 |
+
"epoch": 0.0,
|
| 76 |
+
"grad_norm": 1.4003510475158691,
|
| 77 |
+
"learning_rate": 5e-05,
|
| 78 |
+
"loss": 3.9632,
|
| 79 |
+
"step": 5000
|
| 80 |
+
},
|
| 81 |
+
{
|
| 82 |
+
"epoch": 0.0,
|
| 83 |
+
"grad_norm": 1.7297732830047607,
|
| 84 |
+
"learning_rate": 5.500000000000001e-05,
|
| 85 |
+
"loss": 3.8884,
|
| 86 |
+
"step": 5500
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.0,
|
| 90 |
+
"grad_norm": 1.6848719120025635,
|
| 91 |
+
"learning_rate": 6e-05,
|
| 92 |
+
"loss": 3.8555,
|
| 93 |
+
"step": 6000
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 0.0,
|
| 97 |
+
"grad_norm": 1.5225064754486084,
|
| 98 |
+
"learning_rate": 6.500000000000001e-05,
|
| 99 |
+
"loss": 3.8361,
|
| 100 |
+
"step": 6500
|
| 101 |
+
},
|
| 102 |
+
{
|
| 103 |
+
"epoch": 0.0,
|
| 104 |
+
"grad_norm": 1.369327187538147,
|
| 105 |
+
"learning_rate": 7e-05,
|
| 106 |
+
"loss": 3.8418,
|
| 107 |
+
"step": 7000
|
| 108 |
+
},
|
| 109 |
+
{
|
| 110 |
+
"epoch": 0.0,
|
| 111 |
+
"grad_norm": 1.8992069959640503,
|
| 112 |
+
"learning_rate": 7.500000000000001e-05,
|
| 113 |
+
"loss": 3.7863,
|
| 114 |
+
"step": 7500
|
| 115 |
+
},
|
| 116 |
+
{
|
| 117 |
+
"epoch": 0.0,
|
| 118 |
+
"grad_norm": 0.6943246722221375,
|
| 119 |
+
"learning_rate": 8e-05,
|
| 120 |
+
"loss": 3.7756,
|
| 121 |
+
"step": 8000
|
| 122 |
+
},
|
| 123 |
+
{
|
| 124 |
+
"epoch": 0.0,
|
| 125 |
+
"grad_norm": 0.9557390213012695,
|
| 126 |
+
"learning_rate": 8.5e-05,
|
| 127 |
+
"loss": 3.7761,
|
| 128 |
+
"step": 8500
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.01,
|
| 132 |
+
"grad_norm": 5.63359260559082,
|
| 133 |
+
"learning_rate": 9e-05,
|
| 134 |
+
"loss": 3.7632,
|
| 135 |
+
"step": 9000
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 0.01,
|
| 139 |
+
"grad_norm": 2.1707723140716553,
|
| 140 |
+
"learning_rate": 9.5e-05,
|
| 141 |
+
"loss": 3.7302,
|
| 142 |
+
"step": 9500
|
| 143 |
+
},
|
| 144 |
+
{
|
| 145 |
+
"epoch": 0.01,
|
| 146 |
+
"grad_norm": 1.575547456741333,
|
| 147 |
+
"learning_rate": 0.0001,
|
| 148 |
+
"loss": 3.6666,
|
| 149 |
+
"step": 10000
|
| 150 |
+
},
|
| 151 |
+
{
|
| 152 |
+
"epoch": 0.01,
|
| 153 |
+
"grad_norm": 1.7842143774032593,
|
| 154 |
+
"learning_rate": 9.999907191972478e-05,
|
| 155 |
+
"loss": 3.6669,
|
| 156 |
+
"step": 10500
|
| 157 |
+
},
|
| 158 |
+
{
|
| 159 |
+
"epoch": 0.01,
|
| 160 |
+
"grad_norm": 3.472308874130249,
|
| 161 |
+
"learning_rate": 9.999814383944955e-05,
|
| 162 |
+
"loss": 3.6722,
|
| 163 |
+
"step": 11000
|
| 164 |
+
},
|
| 165 |
+
{
|
| 166 |
+
"epoch": 0.01,
|
| 167 |
+
"grad_norm": 4.764346122741699,
|
| 168 |
+
"learning_rate": 9.99972157591743e-05,
|
| 169 |
+
"loss": 3.6258,
|
| 170 |
+
"step": 11500
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 0.01,
|
| 174 |
+
"grad_norm": 1.4110618829727173,
|
| 175 |
+
"learning_rate": 9.999628767889908e-05,
|
| 176 |
+
"loss": 3.6232,
|
| 177 |
+
"step": 12000
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 0.01,
|
| 181 |
+
"grad_norm": 1.8532825708389282,
|
| 182 |
+
"learning_rate": 9.999535959862385e-05,
|
| 183 |
+
"loss": 3.6201,
|
| 184 |
+
"step": 12500
|
| 185 |
+
},
|
| 186 |
+
{
|
| 187 |
+
"epoch": 0.01,
|
| 188 |
+
"grad_norm": 3.7213845252990723,
|
| 189 |
+
"learning_rate": 9.999443151834862e-05,
|
| 190 |
+
"loss": 3.5771,
|
| 191 |
+
"step": 13000
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"epoch": 0.01,
|
| 195 |
+
"grad_norm": 6.306430816650391,
|
| 196 |
+
"learning_rate": 9.999350343807339e-05,
|
| 197 |
+
"loss": 3.5676,
|
| 198 |
+
"step": 13500
|
| 199 |
+
},
|
| 200 |
+
{
|
| 201 |
+
"epoch": 0.01,
|
| 202 |
+
"grad_norm": 3.5671603679656982,
|
| 203 |
+
"learning_rate": 9.999257535779816e-05,
|
| 204 |
+
"loss": 3.542,
|
| 205 |
+
"step": 14000
|
| 206 |
+
},
|
| 207 |
+
{
|
| 208 |
+
"epoch": 0.01,
|
| 209 |
+
"grad_norm": 1.9145666360855103,
|
| 210 |
+
"learning_rate": 9.999164727752292e-05,
|
| 211 |
+
"loss": 3.524,
|
| 212 |
+
"step": 14500
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 0.01,
|
| 216 |
+
"grad_norm": 2.1453213691711426,
|
| 217 |
+
"learning_rate": 9.999071919724769e-05,
|
| 218 |
+
"loss": 3.507,
|
| 219 |
+
"step": 15000
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 0.01,
|
| 223 |
+
"grad_norm": 2.5378761291503906,
|
| 224 |
+
"learning_rate": 9.998979111697245e-05,
|
| 225 |
+
"loss": 3.5117,
|
| 226 |
+
"step": 15500
|
| 227 |
+
},
|
| 228 |
+
{
|
| 229 |
+
"epoch": 0.01,
|
| 230 |
+
"grad_norm": 1.8997902870178223,
|
| 231 |
+
"learning_rate": 9.998886303669722e-05,
|
| 232 |
+
"loss": 3.4705,
|
| 233 |
+
"step": 16000
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"epoch": 0.01,
|
| 237 |
+
"grad_norm": 1.9014019966125488,
|
| 238 |
+
"learning_rate": 9.998793495642199e-05,
|
| 239 |
+
"loss": 3.4625,
|
| 240 |
+
"step": 16500
|
| 241 |
+
},
|
| 242 |
+
{
|
| 243 |
+
"epoch": 0.01,
|
| 244 |
+
"grad_norm": 4.82717227935791,
|
| 245 |
+
"learning_rate": 9.998700687614676e-05,
|
| 246 |
+
"loss": 3.462,
|
| 247 |
+
"step": 17000
|
| 248 |
+
},
|
| 249 |
+
{
|
| 250 |
+
"epoch": 0.01,
|
| 251 |
+
"grad_norm": 4.583270072937012,
|
| 252 |
+
"learning_rate": 9.998607879587154e-05,
|
| 253 |
+
"loss": 3.434,
|
| 254 |
+
"step": 17500
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 0.01,
|
| 258 |
+
"grad_norm": 2.3539793491363525,
|
| 259 |
+
"learning_rate": 9.998515071559629e-05,
|
| 260 |
+
"loss": 3.3932,
|
| 261 |
+
"step": 18000
|
| 262 |
+
},
|
| 263 |
+
{
|
| 264 |
+
"epoch": 0.01,
|
| 265 |
+
"grad_norm": 3.745349884033203,
|
| 266 |
+
"learning_rate": 9.998422263532106e-05,
|
| 267 |
+
"loss": 3.3875,
|
| 268 |
+
"step": 18500
|
| 269 |
+
},
|
| 270 |
+
{
|
| 271 |
+
"epoch": 0.01,
|
| 272 |
+
"grad_norm": 3.2298102378845215,
|
| 273 |
+
"learning_rate": 9.998329455504584e-05,
|
| 274 |
+
"loss": 3.3663,
|
| 275 |
+
"step": 19000
|
| 276 |
+
},
|
| 277 |
+
{
|
| 278 |
+
"epoch": 0.01,
|
| 279 |
+
"grad_norm": 2.3267478942871094,
|
| 280 |
+
"learning_rate": 9.998236647477061e-05,
|
| 281 |
+
"loss": 3.3471,
|
| 282 |
+
"step": 19500
|
| 283 |
+
},
|
| 284 |
+
{
|
| 285 |
+
"epoch": 0.01,
|
| 286 |
+
"grad_norm": 4.170761585235596,
|
| 287 |
+
"learning_rate": 9.998143839449538e-05,
|
| 288 |
+
"loss": 3.3523,
|
| 289 |
+
"step": 20000
|
| 290 |
+
},
|
| 291 |
+
{
|
| 292 |
+
"epoch": 0.01,
|
| 293 |
+
"grad_norm": 3.3661534786224365,
|
| 294 |
+
"learning_rate": 9.998051031422015e-05,
|
| 295 |
+
"loss": 3.3571,
|
| 296 |
+
"step": 20500
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 0.01,
|
| 300 |
+
"grad_norm": 3.1715431213378906,
|
| 301 |
+
"learning_rate": 9.997958223394491e-05,
|
| 302 |
+
"loss": 3.3627,
|
| 303 |
+
"step": 21000
|
| 304 |
+
},
|
| 305 |
+
{
|
| 306 |
+
"epoch": 0.01,
|
| 307 |
+
"grad_norm": 12.678780555725098,
|
| 308 |
+
"learning_rate": 9.997865415366968e-05,
|
| 309 |
+
"loss": 3.3433,
|
| 310 |
+
"step": 21500
|
| 311 |
+
},
|
| 312 |
+
{
|
| 313 |
+
"epoch": 0.01,
|
| 314 |
+
"grad_norm": 4.594781398773193,
|
| 315 |
+
"learning_rate": 9.997772607339445e-05,
|
| 316 |
+
"loss": 3.3309,
|
| 317 |
+
"step": 22000
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 0.01,
|
| 321 |
+
"grad_norm": 3.517892599105835,
|
| 322 |
+
"learning_rate": 9.997679799311922e-05,
|
| 323 |
+
"loss": 3.3274,
|
| 324 |
+
"step": 22500
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 0.01,
|
| 328 |
+
"grad_norm": 2.783637285232544,
|
| 329 |
+
"learning_rate": 9.9975869912844e-05,
|
| 330 |
+
"loss": 3.3317,
|
| 331 |
+
"step": 23000
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 0.01,
|
| 335 |
+
"grad_norm": 4.124772548675537,
|
| 336 |
+
"learning_rate": 9.997494183256876e-05,
|
| 337 |
+
"loss": 3.3127,
|
| 338 |
+
"step": 23500
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 0.01,
|
| 342 |
+
"grad_norm": 2.7662994861602783,
|
| 343 |
+
"learning_rate": 9.997401375229352e-05,
|
| 344 |
+
"loss": 3.3101,
|
| 345 |
+
"step": 24000
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 0.01,
|
| 349 |
+
"grad_norm": 2.652294158935547,
|
| 350 |
+
"learning_rate": 9.99730856720183e-05,
|
| 351 |
+
"loss": 3.2721,
|
| 352 |
+
"step": 24500
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 0.01,
|
| 356 |
+
"grad_norm": 2.3313260078430176,
|
| 357 |
+
"learning_rate": 9.997215759174307e-05,
|
| 358 |
+
"loss": 3.2796,
|
| 359 |
+
"step": 25000
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 0.01,
|
| 363 |
+
"grad_norm": 4.2452392578125,
|
| 364 |
+
"learning_rate": 9.997122951146782e-05,
|
| 365 |
+
"loss": 3.2846,
|
| 366 |
+
"step": 25500
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 0.01,
|
| 370 |
+
"grad_norm": 3.510793924331665,
|
| 371 |
+
"learning_rate": 9.99703014311926e-05,
|
| 372 |
+
"loss": 3.29,
|
| 373 |
+
"step": 26000
|
| 374 |
+
},
|
| 375 |
+
{
|
| 376 |
+
"epoch": 0.01,
|
| 377 |
+
"grad_norm": 2.9583752155303955,
|
| 378 |
+
"learning_rate": 9.996937335091737e-05,
|
| 379 |
+
"loss": 3.2618,
|
| 380 |
+
"step": 26500
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 0.02,
|
| 384 |
+
"grad_norm": 3.7444071769714355,
|
| 385 |
+
"learning_rate": 9.996844527064212e-05,
|
| 386 |
+
"loss": 3.2348,
|
| 387 |
+
"step": 27000
|
| 388 |
+
},
|
| 389 |
+
{
|
| 390 |
+
"epoch": 0.02,
|
| 391 |
+
"grad_norm": 2.852471113204956,
|
| 392 |
+
"learning_rate": 9.99675171903669e-05,
|
| 393 |
+
"loss": 3.2315,
|
| 394 |
+
"step": 27500
|
| 395 |
+
},
|
| 396 |
+
{
|
| 397 |
+
"epoch": 0.02,
|
| 398 |
+
"grad_norm": 2.542102813720703,
|
| 399 |
+
"learning_rate": 9.996658911009167e-05,
|
| 400 |
+
"loss": 3.2564,
|
| 401 |
+
"step": 28000
|
| 402 |
+
},
|
| 403 |
+
{
|
| 404 |
+
"epoch": 0.02,
|
| 405 |
+
"grad_norm": 3.195061445236206,
|
| 406 |
+
"learning_rate": 9.996566102981644e-05,
|
| 407 |
+
"loss": 3.2477,
|
| 408 |
+
"step": 28500
|
| 409 |
+
},
|
| 410 |
+
{
|
| 411 |
+
"epoch": 0.02,
|
| 412 |
+
"grad_norm": 4.133997917175293,
|
| 413 |
+
"learning_rate": 9.996473294954121e-05,
|
| 414 |
+
"loss": 3.2246,
|
| 415 |
+
"step": 29000
|
| 416 |
+
},
|
| 417 |
+
{
|
| 418 |
+
"epoch": 0.02,
|
| 419 |
+
"grad_norm": 3.147731304168701,
|
| 420 |
+
"learning_rate": 9.996380486926598e-05,
|
| 421 |
+
"loss": 3.2483,
|
| 422 |
+
"step": 29500
|
| 423 |
+
},
|
| 424 |
+
{
|
| 425 |
+
"epoch": 0.02,
|
| 426 |
+
"grad_norm": 2.417142868041992,
|
| 427 |
+
"learning_rate": 9.996287678899074e-05,
|
| 428 |
+
"loss": 3.2084,
|
| 429 |
+
"step": 30000
|
| 430 |
+
},
|
| 431 |
+
{
|
| 432 |
+
"epoch": 0.02,
|
| 433 |
+
"grad_norm": 2.3051340579986572,
|
| 434 |
+
"learning_rate": 9.996194870871551e-05,
|
| 435 |
+
"loss": 3.2305,
|
| 436 |
+
"step": 30500
|
| 437 |
+
},
|
| 438 |
+
{
|
| 439 |
+
"epoch": 0.02,
|
| 440 |
+
"grad_norm": 3.268831491470337,
|
| 441 |
+
"learning_rate": 9.996102062844028e-05,
|
| 442 |
+
"loss": 3.2198,
|
| 443 |
+
"step": 31000
|
| 444 |
+
},
|
| 445 |
+
{
|
| 446 |
+
"epoch": 0.02,
|
| 447 |
+
"grad_norm": 2.9484684467315674,
|
| 448 |
+
"learning_rate": 9.996009254816505e-05,
|
| 449 |
+
"loss": 3.1927,
|
| 450 |
+
"step": 31500
|
| 451 |
+
},
|
| 452 |
+
{
|
| 453 |
+
"epoch": 0.02,
|
| 454 |
+
"grad_norm": 5.482509613037109,
|
| 455 |
+
"learning_rate": 9.995916446788982e-05,
|
| 456 |
+
"loss": 3.2018,
|
| 457 |
+
"step": 32000
|
| 458 |
+
},
|
| 459 |
+
{
|
| 460 |
+
"epoch": 0.02,
|
| 461 |
+
"grad_norm": 3.319070339202881,
|
| 462 |
+
"learning_rate": 9.99582363876146e-05,
|
| 463 |
+
"loss": 3.2034,
|
| 464 |
+
"step": 32500
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 0.02,
|
| 468 |
+
"grad_norm": 2.9254605770111084,
|
| 469 |
+
"learning_rate": 9.995730830733935e-05,
|
| 470 |
+
"loss": 3.1935,
|
| 471 |
+
"step": 33000
|
| 472 |
+
},
|
| 473 |
+
{
|
| 474 |
+
"epoch": 0.02,
|
| 475 |
+
"grad_norm": 3.508344888687134,
|
| 476 |
+
"learning_rate": 9.995638022706412e-05,
|
| 477 |
+
"loss": 3.1913,
|
| 478 |
+
"step": 33500
|
| 479 |
+
},
|
| 480 |
+
{
|
| 481 |
+
"epoch": 0.02,
|
| 482 |
+
"grad_norm": 2.0120761394500732,
|
| 483 |
+
"learning_rate": 9.99554521467889e-05,
|
| 484 |
+
"loss": 3.2221,
|
| 485 |
+
"step": 34000
|
| 486 |
+
},
|
| 487 |
+
{
|
| 488 |
+
"epoch": 0.02,
|
| 489 |
+
"grad_norm": 2.6379246711730957,
|
| 490 |
+
"learning_rate": 9.995452406651367e-05,
|
| 491 |
+
"loss": 3.1848,
|
| 492 |
+
"step": 34500
|
| 493 |
+
},
|
| 494 |
+
{
|
| 495 |
+
"epoch": 0.02,
|
| 496 |
+
"grad_norm": 4.178890705108643,
|
| 497 |
+
"learning_rate": 9.995359598623844e-05,
|
| 498 |
+
"loss": 3.184,
|
| 499 |
+
"step": 35000
|
| 500 |
+
},
|
| 501 |
+
{
|
| 502 |
+
"epoch": 0.02,
|
| 503 |
+
"grad_norm": 3.148407459259033,
|
| 504 |
+
"learning_rate": 9.995266790596321e-05,
|
| 505 |
+
"loss": 3.2119,
|
| 506 |
+
"step": 35500
|
| 507 |
+
},
|
| 508 |
+
{
|
| 509 |
+
"epoch": 0.02,
|
| 510 |
+
"grad_norm": 2.3678390979766846,
|
| 511 |
+
"learning_rate": 9.995173982568797e-05,
|
| 512 |
+
"loss": 3.1744,
|
| 513 |
+
"step": 36000
|
| 514 |
+
},
|
| 515 |
+
{
|
| 516 |
+
"epoch": 0.02,
|
| 517 |
+
"grad_norm": 3.0567054748535156,
|
| 518 |
+
"learning_rate": 9.995081174541273e-05,
|
| 519 |
+
"loss": 3.1821,
|
| 520 |
+
"step": 36500
|
| 521 |
+
},
|
| 522 |
+
{
|
| 523 |
+
"epoch": 0.02,
|
| 524 |
+
"grad_norm": 3.4158051013946533,
|
| 525 |
+
"learning_rate": 9.99498836651375e-05,
|
| 526 |
+
"loss": 3.1615,
|
| 527 |
+
"step": 37000
|
| 528 |
+
},
|
| 529 |
+
{
|
| 530 |
+
"epoch": 0.02,
|
| 531 |
+
"grad_norm": 3.654043197631836,
|
| 532 |
+
"learning_rate": 9.994895558486227e-05,
|
| 533 |
+
"loss": 3.174,
|
| 534 |
+
"step": 37500
|
| 535 |
+
},
|
| 536 |
+
{
|
| 537 |
+
"epoch": 0.02,
|
| 538 |
+
"grad_norm": 3.634075403213501,
|
| 539 |
+
"learning_rate": 9.994802750458704e-05,
|
| 540 |
+
"loss": 3.1792,
|
| 541 |
+
"step": 38000
|
| 542 |
+
},
|
| 543 |
+
{
|
| 544 |
+
"epoch": 0.02,
|
| 545 |
+
"grad_norm": 2.8857271671295166,
|
| 546 |
+
"learning_rate": 9.994709942431181e-05,
|
| 547 |
+
"loss": 3.1441,
|
| 548 |
+
"step": 38500
|
| 549 |
+
},
|
| 550 |
+
{
|
| 551 |
+
"epoch": 0.02,
|
| 552 |
+
"grad_norm": 3.0216064453125,
|
| 553 |
+
"learning_rate": 9.994617134403658e-05,
|
| 554 |
+
"loss": 3.1464,
|
| 555 |
+
"step": 39000
|
| 556 |
+
},
|
| 557 |
+
{
|
| 558 |
+
"epoch": 0.02,
|
| 559 |
+
"grad_norm": 4.89854097366333,
|
| 560 |
+
"learning_rate": 9.994524326376134e-05,
|
| 561 |
+
"loss": 3.1514,
|
| 562 |
+
"step": 39500
|
| 563 |
+
},
|
| 564 |
+
{
|
| 565 |
+
"epoch": 0.02,
|
| 566 |
+
"grad_norm": 3.495551586151123,
|
| 567 |
+
"learning_rate": 9.994431518348611e-05,
|
| 568 |
+
"loss": 3.153,
|
| 569 |
+
"step": 40000
|
| 570 |
+
},
|
| 571 |
+
{
|
| 572 |
+
"epoch": 0.02,
|
| 573 |
+
"grad_norm": 2.9397456645965576,
|
| 574 |
+
"learning_rate": 9.994338710321088e-05,
|
| 575 |
+
"loss": 3.1329,
|
| 576 |
+
"step": 40500
|
| 577 |
+
},
|
| 578 |
+
{
|
| 579 |
+
"epoch": 0.02,
|
| 580 |
+
"grad_norm": 5.4169721603393555,
|
| 581 |
+
"learning_rate": 9.994245902293566e-05,
|
| 582 |
+
"loss": 3.1485,
|
| 583 |
+
"step": 41000
|
| 584 |
+
},
|
| 585 |
+
{
|
| 586 |
+
"epoch": 0.02,
|
| 587 |
+
"grad_norm": 2.784404754638672,
|
| 588 |
+
"learning_rate": 9.994153094266043e-05,
|
| 589 |
+
"loss": 3.1423,
|
| 590 |
+
"step": 41500
|
| 591 |
+
},
|
| 592 |
+
{
|
| 593 |
+
"epoch": 0.02,
|
| 594 |
+
"grad_norm": 2.435594320297241,
|
| 595 |
+
"learning_rate": 9.994060286238518e-05,
|
| 596 |
+
"loss": 3.133,
|
| 597 |
+
"step": 42000
|
| 598 |
+
},
|
| 599 |
+
{
|
| 600 |
+
"epoch": 0.02,
|
| 601 |
+
"grad_norm": 2.478294610977173,
|
| 602 |
+
"learning_rate": 9.993967478210996e-05,
|
| 603 |
+
"loss": 3.135,
|
| 604 |
+
"step": 42500
|
| 605 |
+
},
|
| 606 |
+
{
|
| 607 |
+
"epoch": 0.02,
|
| 608 |
+
"grad_norm": 3.871967315673828,
|
| 609 |
+
"learning_rate": 9.993874670183473e-05,
|
| 610 |
+
"loss": 3.1133,
|
| 611 |
+
"step": 43000
|
| 612 |
+
},
|
| 613 |
+
{
|
| 614 |
+
"epoch": 0.02,
|
| 615 |
+
"grad_norm": 2.735368490219116,
|
| 616 |
+
"learning_rate": 9.99378186215595e-05,
|
| 617 |
+
"loss": 3.1333,
|
| 618 |
+
"step": 43500
|
| 619 |
+
},
|
| 620 |
+
{
|
| 621 |
+
"epoch": 0.02,
|
| 622 |
+
"grad_norm": 3.2882280349731445,
|
| 623 |
+
"learning_rate": 9.993689054128427e-05,
|
| 624 |
+
"loss": 3.1231,
|
| 625 |
+
"step": 44000
|
| 626 |
+
},
|
| 627 |
+
{
|
| 628 |
+
"epoch": 0.02,
|
| 629 |
+
"grad_norm": 4.579193115234375,
|
| 630 |
+
"learning_rate": 9.993596246100904e-05,
|
| 631 |
+
"loss": 3.1051,
|
| 632 |
+
"step": 44500
|
| 633 |
+
},
|
| 634 |
+
{
|
| 635 |
+
"epoch": 0.03,
|
| 636 |
+
"grad_norm": 4.136692047119141,
|
| 637 |
+
"learning_rate": 9.99350343807338e-05,
|
| 638 |
+
"loss": 3.1295,
|
| 639 |
+
"step": 45000
|
| 640 |
+
},
|
| 641 |
+
{
|
| 642 |
+
"epoch": 0.03,
|
| 643 |
+
"grad_norm": 3.8374266624450684,
|
| 644 |
+
"learning_rate": 9.993410630045857e-05,
|
| 645 |
+
"loss": 3.1182,
|
| 646 |
+
"step": 45500
|
| 647 |
+
},
|
| 648 |
+
{
|
| 649 |
+
"epoch": 0.03,
|
| 650 |
+
"grad_norm": 3.8597323894500732,
|
| 651 |
+
"learning_rate": 9.993317822018334e-05,
|
| 652 |
+
"loss": 3.1102,
|
| 653 |
+
"step": 46000
|
| 654 |
+
},
|
| 655 |
+
{
|
| 656 |
+
"epoch": 0.03,
|
| 657 |
+
"grad_norm": 2.7364611625671387,
|
| 658 |
+
"learning_rate": 9.99322501399081e-05,
|
| 659 |
+
"loss": 3.1021,
|
| 660 |
+
"step": 46500
|
| 661 |
+
},
|
| 662 |
+
{
|
| 663 |
+
"epoch": 0.03,
|
| 664 |
+
"grad_norm": 10.505576133728027,
|
| 665 |
+
"learning_rate": 9.993132205963287e-05,
|
| 666 |
+
"loss": 3.087,
|
| 667 |
+
"step": 47000
|
| 668 |
+
},
|
| 669 |
+
{
|
| 670 |
+
"epoch": 0.03,
|
| 671 |
+
"grad_norm": 3.8307197093963623,
|
| 672 |
+
"learning_rate": 9.993039397935764e-05,
|
| 673 |
+
"loss": 3.1033,
|
| 674 |
+
"step": 47500
|
| 675 |
+
},
|
| 676 |
+
{
|
| 677 |
+
"epoch": 0.03,
|
| 678 |
+
"grad_norm": 3.2991435527801514,
|
| 679 |
+
"learning_rate": 9.992946589908241e-05,
|
| 680 |
+
"loss": 3.0994,
|
| 681 |
+
"step": 48000
|
| 682 |
+
},
|
| 683 |
+
{
|
| 684 |
+
"epoch": 0.03,
|
| 685 |
+
"grad_norm": 2.461336612701416,
|
| 686 |
+
"learning_rate": 9.992853781880717e-05,
|
| 687 |
+
"loss": 3.0768,
|
| 688 |
+
"step": 48500
|
| 689 |
+
},
|
| 690 |
+
{
|
| 691 |
+
"epoch": 0.03,
|
| 692 |
+
"grad_norm": 4.653237819671631,
|
| 693 |
+
"learning_rate": 9.992760973853194e-05,
|
| 694 |
+
"loss": 3.0751,
|
| 695 |
+
"step": 49000
|
| 696 |
+
},
|
| 697 |
+
{
|
| 698 |
+
"epoch": 0.03,
|
| 699 |
+
"grad_norm": 3.9689579010009766,
|
| 700 |
+
"learning_rate": 9.992668165825671e-05,
|
| 701 |
+
"loss": 3.0845,
|
| 702 |
+
"step": 49500
|
| 703 |
+
},
|
| 704 |
+
{
|
| 705 |
+
"epoch": 0.03,
|
| 706 |
+
"grad_norm": 3.50116229057312,
|
| 707 |
+
"learning_rate": 9.992575357798149e-05,
|
| 708 |
+
"loss": 3.0563,
|
| 709 |
+
"step": 50000
|
| 710 |
+
},
|
| 711 |
+
{
|
| 712 |
+
"epoch": 0.03,
|
| 713 |
+
"grad_norm": 7.10615873336792,
|
| 714 |
+
"learning_rate": 9.992482549770626e-05,
|
| 715 |
+
"loss": 3.0585,
|
| 716 |
+
"step": 50500
|
| 717 |
+
},
|
| 718 |
+
{
|
| 719 |
+
"epoch": 0.03,
|
| 720 |
+
"grad_norm": 3.484778642654419,
|
| 721 |
+
"learning_rate": 9.992389741743103e-05,
|
| 722 |
+
"loss": 3.0657,
|
| 723 |
+
"step": 51000
|
| 724 |
+
},
|
| 725 |
+
{
|
| 726 |
+
"epoch": 0.03,
|
| 727 |
+
"grad_norm": 4.087871074676514,
|
| 728 |
+
"learning_rate": 9.992296933715579e-05,
|
| 729 |
+
"loss": 3.0302,
|
| 730 |
+
"step": 51500
|
| 731 |
+
},
|
| 732 |
+
{
|
| 733 |
+
"epoch": 0.03,
|
| 734 |
+
"grad_norm": 6.728222846984863,
|
| 735 |
+
"learning_rate": 9.992204125688056e-05,
|
| 736 |
+
"loss": 3.0197,
|
| 737 |
+
"step": 52000
|
| 738 |
+
},
|
| 739 |
+
{
|
| 740 |
+
"epoch": 0.03,
|
| 741 |
+
"grad_norm": 4.08004093170166,
|
| 742 |
+
"learning_rate": 9.992111317660533e-05,
|
| 743 |
+
"loss": 3.0261,
|
| 744 |
+
"step": 52500
|
| 745 |
+
},
|
| 746 |
+
{
|
| 747 |
+
"epoch": 0.03,
|
| 748 |
+
"grad_norm": 4.262966156005859,
|
| 749 |
+
"learning_rate": 9.99201850963301e-05,
|
| 750 |
+
"loss": 3.0185,
|
| 751 |
+
"step": 53000
|
| 752 |
+
},
|
| 753 |
+
{
|
| 754 |
+
"epoch": 0.03,
|
| 755 |
+
"grad_norm": 3.3461170196533203,
|
| 756 |
+
"learning_rate": 9.991925701605487e-05,
|
| 757 |
+
"loss": 2.9916,
|
| 758 |
+
"step": 53500
|
| 759 |
+
},
|
| 760 |
+
{
|
| 761 |
+
"epoch": 0.03,
|
| 762 |
+
"grad_norm": 5.650246620178223,
|
| 763 |
+
"learning_rate": 9.991832893577964e-05,
|
| 764 |
+
"loss": 3.0014,
|
| 765 |
+
"step": 54000
|
| 766 |
+
},
|
| 767 |
+
{
|
| 768 |
+
"epoch": 0.03,
|
| 769 |
+
"grad_norm": 4.820484161376953,
|
| 770 |
+
"learning_rate": 9.99174008555044e-05,
|
| 771 |
+
"loss": 2.9966,
|
| 772 |
+
"step": 54500
|
| 773 |
+
},
|
| 774 |
+
{
|
| 775 |
+
"epoch": 0.03,
|
| 776 |
+
"grad_norm": 3.245210886001587,
|
| 777 |
+
"learning_rate": 9.991647277522917e-05,
|
| 778 |
+
"loss": 2.9829,
|
| 779 |
+
"step": 55000
|
| 780 |
+
},
|
| 781 |
+
{
|
| 782 |
+
"epoch": 0.03,
|
| 783 |
+
"grad_norm": 3.009565591812134,
|
| 784 |
+
"learning_rate": 9.991554469495394e-05,
|
| 785 |
+
"loss": 2.9665,
|
| 786 |
+
"step": 55500
|
| 787 |
+
},
|
| 788 |
+
{
|
| 789 |
+
"epoch": 0.03,
|
| 790 |
+
"grad_norm": 5.277273654937744,
|
| 791 |
+
"learning_rate": 9.991461661467872e-05,
|
| 792 |
+
"loss": 2.9572,
|
| 793 |
+
"step": 56000
|
| 794 |
+
},
|
| 795 |
+
{
|
| 796 |
+
"epoch": 0.03,
|
| 797 |
+
"grad_norm": 3.0885508060455322,
|
| 798 |
+
"learning_rate": 9.991368853440347e-05,
|
| 799 |
+
"loss": 2.953,
|
| 800 |
+
"step": 56500
|
| 801 |
+
},
|
| 802 |
+
{
|
| 803 |
+
"epoch": 0.03,
|
| 804 |
+
"grad_norm": 3.3525726795196533,
|
| 805 |
+
"learning_rate": 9.991276045412824e-05,
|
| 806 |
+
"loss": 2.9494,
|
| 807 |
+
"step": 57000
|
| 808 |
+
},
|
| 809 |
+
{
|
| 810 |
+
"epoch": 0.03,
|
| 811 |
+
"grad_norm": 3.699075937271118,
|
| 812 |
+
"learning_rate": 9.9911832373853e-05,
|
| 813 |
+
"loss": 2.9731,
|
| 814 |
+
"step": 57500
|
| 815 |
+
},
|
| 816 |
+
{
|
| 817 |
+
"epoch": 0.03,
|
| 818 |
+
"grad_norm": 2.363311529159546,
|
| 819 |
+
"learning_rate": 9.991090429357777e-05,
|
| 820 |
+
"loss": 2.9257,
|
| 821 |
+
"step": 58000
|
| 822 |
+
},
|
| 823 |
+
{
|
| 824 |
+
"epoch": 0.03,
|
| 825 |
+
"grad_norm": 3.144495964050293,
|
| 826 |
+
"learning_rate": 9.990997621330255e-05,
|
| 827 |
+
"loss": 2.9135,
|
| 828 |
+
"step": 58500
|
| 829 |
+
},
|
| 830 |
+
{
|
| 831 |
+
"epoch": 0.03,
|
| 832 |
+
"grad_norm": 3.192347288131714,
|
| 833 |
+
"learning_rate": 9.990904813302732e-05,
|
| 834 |
+
"loss": 2.8925,
|
| 835 |
+
"step": 59000
|
| 836 |
+
},
|
| 837 |
+
{
|
| 838 |
+
"epoch": 0.03,
|
| 839 |
+
"grad_norm": 3.4419093132019043,
|
| 840 |
+
"learning_rate": 9.990812005275209e-05,
|
| 841 |
+
"loss": 2.9018,
|
| 842 |
+
"step": 59500
|
| 843 |
+
},
|
| 844 |
+
{
|
| 845 |
+
"epoch": 0.03,
|
| 846 |
+
"grad_norm": 3.506303071975708,
|
| 847 |
+
"learning_rate": 9.990719197247686e-05,
|
| 848 |
+
"loss": 2.8921,
|
| 849 |
+
"step": 60000
|
| 850 |
+
},
|
| 851 |
+
{
|
| 852 |
+
"epoch": 0.03,
|
| 853 |
+
"grad_norm": 2.7846992015838623,
|
| 854 |
+
"learning_rate": 9.990626389220162e-05,
|
| 855 |
+
"loss": 2.8904,
|
| 856 |
+
"step": 60500
|
| 857 |
+
},
|
| 858 |
+
{
|
| 859 |
+
"epoch": 0.03,
|
| 860 |
+
"grad_norm": 3.9778714179992676,
|
| 861 |
+
"learning_rate": 9.990533581192639e-05,
|
| 862 |
+
"loss": 2.854,
|
| 863 |
+
"step": 61000
|
| 864 |
+
},
|
| 865 |
+
{
|
| 866 |
+
"epoch": 0.03,
|
| 867 |
+
"grad_norm": 3.1654438972473145,
|
| 868 |
+
"learning_rate": 9.990440773165116e-05,
|
| 869 |
+
"loss": 2.8481,
|
| 870 |
+
"step": 61500
|
| 871 |
+
},
|
| 872 |
+
{
|
| 873 |
+
"epoch": 0.03,
|
| 874 |
+
"grad_norm": 5.007691860198975,
|
| 875 |
+
"learning_rate": 9.990347965137593e-05,
|
| 876 |
+
"loss": 2.836,
|
| 877 |
+
"step": 62000
|
| 878 |
+
},
|
| 879 |
+
{
|
| 880 |
+
"epoch": 0.03,
|
| 881 |
+
"grad_norm": 3.4846861362457275,
|
| 882 |
+
"learning_rate": 9.99025515711007e-05,
|
| 883 |
+
"loss": 2.8312,
|
| 884 |
+
"step": 62500
|
| 885 |
+
},
|
| 886 |
+
{
|
| 887 |
+
"epoch": 0.04,
|
| 888 |
+
"grad_norm": 6.449169158935547,
|
| 889 |
+
"learning_rate": 9.990162349082547e-05,
|
| 890 |
+
"loss": 2.8371,
|
| 891 |
+
"step": 63000
|
| 892 |
+
},
|
| 893 |
+
{
|
| 894 |
+
"epoch": 0.04,
|
| 895 |
+
"grad_norm": 5.483363628387451,
|
| 896 |
+
"learning_rate": 9.990069541055023e-05,
|
| 897 |
+
"loss": 2.8074,
|
| 898 |
+
"step": 63500
|
| 899 |
+
},
|
| 900 |
+
{
|
| 901 |
+
"epoch": 0.04,
|
| 902 |
+
"grad_norm": 4.194338798522949,
|
| 903 |
+
"learning_rate": 9.9899767330275e-05,
|
| 904 |
+
"loss": 2.8134,
|
| 905 |
+
"step": 64000
|
| 906 |
+
},
|
| 907 |
+
{
|
| 908 |
+
"epoch": 0.04,
|
| 909 |
+
"grad_norm": 4.445028305053711,
|
| 910 |
+
"learning_rate": 9.989883924999978e-05,
|
| 911 |
+
"loss": 2.8024,
|
| 912 |
+
"step": 64500
|
| 913 |
+
},
|
| 914 |
+
{
|
| 915 |
+
"epoch": 0.04,
|
| 916 |
+
"grad_norm": 4.259857177734375,
|
| 917 |
+
"learning_rate": 9.989791116972455e-05,
|
| 918 |
+
"loss": 2.8063,
|
| 919 |
+
"step": 65000
|
| 920 |
+
},
|
| 921 |
+
{
|
| 922 |
+
"epoch": 0.04,
|
| 923 |
+
"grad_norm": 15.38962173461914,
|
| 924 |
+
"learning_rate": 9.989698308944932e-05,
|
| 925 |
+
"loss": 2.7849,
|
| 926 |
+
"step": 65500
|
| 927 |
+
},
|
| 928 |
+
{
|
| 929 |
+
"epoch": 0.04,
|
| 930 |
+
"grad_norm": 7.087435245513916,
|
| 931 |
+
"learning_rate": 9.989605500917409e-05,
|
| 932 |
+
"loss": 2.7755,
|
| 933 |
+
"step": 66000
|
| 934 |
+
},
|
| 935 |
+
{
|
| 936 |
+
"epoch": 0.04,
|
| 937 |
+
"grad_norm": 4.6525044441223145,
|
| 938 |
+
"learning_rate": 9.989512692889885e-05,
|
| 939 |
+
"loss": 2.7692,
|
| 940 |
+
"step": 66500
|
| 941 |
+
},
|
| 942 |
+
{
|
| 943 |
+
"epoch": 0.04,
|
| 944 |
+
"grad_norm": 3.657395124435425,
|
| 945 |
+
"learning_rate": 9.98941988486236e-05,
|
| 946 |
+
"loss": 2.7754,
|
| 947 |
+
"step": 67000
|
| 948 |
+
},
|
| 949 |
+
{
|
| 950 |
+
"epoch": 0.04,
|
| 951 |
+
"grad_norm": 4.082229137420654,
|
| 952 |
+
"learning_rate": 9.989327076834838e-05,
|
| 953 |
+
"loss": 2.7462,
|
| 954 |
+
"step": 67500
|
| 955 |
+
},
|
| 956 |
+
{
|
| 957 |
+
"epoch": 0.04,
|
| 958 |
+
"grad_norm": 5.71895694732666,
|
| 959 |
+
"learning_rate": 9.989234268807315e-05,
|
| 960 |
+
"loss": 2.7489,
|
| 961 |
+
"step": 68000
|
| 962 |
+
},
|
| 963 |
+
{
|
| 964 |
+
"epoch": 0.04,
|
| 965 |
+
"grad_norm": 4.07025671005249,
|
| 966 |
+
"learning_rate": 9.989141460779792e-05,
|
| 967 |
+
"loss": 2.7429,
|
| 968 |
+
"step": 68500
|
| 969 |
+
},
|
| 970 |
+
{
|
| 971 |
+
"epoch": 0.04,
|
| 972 |
+
"grad_norm": 6.3579888343811035,
|
| 973 |
+
"learning_rate": 9.989048652752269e-05,
|
| 974 |
+
"loss": 2.7359,
|
| 975 |
+
"step": 69000
|
| 976 |
+
},
|
| 977 |
+
{
|
| 978 |
+
"epoch": 0.04,
|
| 979 |
+
"grad_norm": 5.608436107635498,
|
| 980 |
+
"learning_rate": 9.988955844724746e-05,
|
| 981 |
+
"loss": 2.7413,
|
| 982 |
+
"step": 69500
|
| 983 |
+
},
|
| 984 |
+
{
|
| 985 |
+
"epoch": 0.04,
|
| 986 |
+
"grad_norm": 3.6080968379974365,
|
| 987 |
+
"learning_rate": 9.988863036697222e-05,
|
| 988 |
+
"loss": 2.7374,
|
| 989 |
+
"step": 70000
|
| 990 |
+
},
|
| 991 |
+
{
|
| 992 |
+
"epoch": 0.04,
|
| 993 |
+
"grad_norm": 3.0696682929992676,
|
| 994 |
+
"learning_rate": 9.988770228669699e-05,
|
| 995 |
+
"loss": 2.7024,
|
| 996 |
+
"step": 70500
|
| 997 |
+
},
|
| 998 |
+
{
|
| 999 |
+
"epoch": 0.04,
|
| 1000 |
+
"grad_norm": 3.4028148651123047,
|
| 1001 |
+
"learning_rate": 9.988677420642176e-05,
|
| 1002 |
+
"loss": 2.6969,
|
| 1003 |
+
"step": 71000
|
| 1004 |
+
},
|
| 1005 |
+
{
|
| 1006 |
+
"epoch": 0.04,
|
| 1007 |
+
"grad_norm": 3.2408607006073,
|
| 1008 |
+
"learning_rate": 9.988584612614653e-05,
|
| 1009 |
+
"loss": 2.7142,
|
| 1010 |
+
"step": 71500
|
| 1011 |
+
},
|
| 1012 |
+
{
|
| 1013 |
+
"epoch": 0.04,
|
| 1014 |
+
"grad_norm": 4.686740875244141,
|
| 1015 |
+
"learning_rate": 9.98849180458713e-05,
|
| 1016 |
+
"loss": 2.6779,
|
| 1017 |
+
"step": 72000
|
| 1018 |
+
},
|
| 1019 |
+
{
|
| 1020 |
+
"epoch": 0.04,
|
| 1021 |
+
"grad_norm": 4.289364814758301,
|
| 1022 |
+
"learning_rate": 9.988398996559606e-05,
|
| 1023 |
+
"loss": 2.6924,
|
| 1024 |
+
"step": 72500
|
| 1025 |
+
},
|
| 1026 |
+
{
|
| 1027 |
+
"epoch": 0.04,
|
| 1028 |
+
"grad_norm": 5.570862293243408,
|
| 1029 |
+
"learning_rate": 9.988306188532083e-05,
|
| 1030 |
+
"loss": 2.7066,
|
| 1031 |
+
"step": 73000
|
| 1032 |
+
},
|
| 1033 |
+
{
|
| 1034 |
+
"epoch": 0.04,
|
| 1035 |
+
"grad_norm": 6.265756130218506,
|
| 1036 |
+
"learning_rate": 9.98821338050456e-05,
|
| 1037 |
+
"loss": 2.6904,
|
| 1038 |
+
"step": 73500
|
| 1039 |
+
},
|
| 1040 |
+
{
|
| 1041 |
+
"epoch": 0.04,
|
| 1042 |
+
"grad_norm": 4.149326324462891,
|
| 1043 |
+
"learning_rate": 9.988120572477038e-05,
|
| 1044 |
+
"loss": 2.675,
|
| 1045 |
+
"step": 74000
|
| 1046 |
+
},
|
| 1047 |
+
{
|
| 1048 |
+
"epoch": 0.04,
|
| 1049 |
+
"grad_norm": 3.304511547088623,
|
| 1050 |
+
"learning_rate": 9.988027764449515e-05,
|
| 1051 |
+
"loss": 2.6588,
|
| 1052 |
+
"step": 74500
|
| 1053 |
+
},
|
| 1054 |
+
{
|
| 1055 |
+
"epoch": 0.04,
|
| 1056 |
+
"grad_norm": 4.896495342254639,
|
| 1057 |
+
"learning_rate": 9.987934956421992e-05,
|
| 1058 |
+
"loss": 2.6543,
|
| 1059 |
+
"step": 75000
|
| 1060 |
+
},
|
| 1061 |
+
{
|
| 1062 |
+
"epoch": 0.04,
|
| 1063 |
+
"grad_norm": 4.827699661254883,
|
| 1064 |
+
"learning_rate": 9.987842148394468e-05,
|
| 1065 |
+
"loss": 2.6547,
|
| 1066 |
+
"step": 75500
|
| 1067 |
+
},
|
| 1068 |
+
{
|
| 1069 |
+
"epoch": 0.04,
|
| 1070 |
+
"grad_norm": 4.508848667144775,
|
| 1071 |
+
"learning_rate": 9.987749340366945e-05,
|
| 1072 |
+
"loss": 2.6737,
|
| 1073 |
+
"step": 76000
|
| 1074 |
+
},
|
| 1075 |
+
{
|
| 1076 |
+
"epoch": 0.04,
|
| 1077 |
+
"grad_norm": 2.711066722869873,
|
| 1078 |
+
"learning_rate": 9.987656532339422e-05,
|
| 1079 |
+
"loss": 2.6596,
|
| 1080 |
+
"step": 76500
|
| 1081 |
+
},
|
| 1082 |
+
{
|
| 1083 |
+
"epoch": 0.04,
|
| 1084 |
+
"grad_norm": 4.568380832672119,
|
| 1085 |
+
"learning_rate": 9.987563724311898e-05,
|
| 1086 |
+
"loss": 2.6336,
|
| 1087 |
+
"step": 77000
|
| 1088 |
+
},
|
| 1089 |
+
{
|
| 1090 |
+
"epoch": 0.04,
|
| 1091 |
+
"grad_norm": 4.230466365814209,
|
| 1092 |
+
"learning_rate": 9.987470916284375e-05,
|
| 1093 |
+
"loss": 2.6513,
|
| 1094 |
+
"step": 77500
|
| 1095 |
+
},
|
| 1096 |
+
{
|
| 1097 |
+
"epoch": 0.04,
|
| 1098 |
+
"grad_norm": 7.702442646026611,
|
| 1099 |
+
"learning_rate": 9.987378108256852e-05,
|
| 1100 |
+
"loss": 2.6448,
|
| 1101 |
+
"step": 78000
|
| 1102 |
+
},
|
| 1103 |
+
{
|
| 1104 |
+
"epoch": 0.04,
|
| 1105 |
+
"grad_norm": 5.38634729385376,
|
| 1106 |
+
"learning_rate": 9.987285300229329e-05,
|
| 1107 |
+
"loss": 2.6557,
|
| 1108 |
+
"step": 78500
|
| 1109 |
+
},
|
| 1110 |
+
{
|
| 1111 |
+
"epoch": 0.04,
|
| 1112 |
+
"grad_norm": 4.267560005187988,
|
| 1113 |
+
"learning_rate": 9.987192492201805e-05,
|
| 1114 |
+
"loss": 2.6387,
|
| 1115 |
+
"step": 79000
|
| 1116 |
+
},
|
| 1117 |
+
{
|
| 1118 |
+
"epoch": 0.04,
|
| 1119 |
+
"grad_norm": 4.3129801750183105,
|
| 1120 |
+
"learning_rate": 9.987099684174282e-05,
|
| 1121 |
+
"loss": 2.6109,
|
| 1122 |
+
"step": 79500
|
| 1123 |
+
},
|
| 1124 |
+
{
|
| 1125 |
+
"epoch": 0.04,
|
| 1126 |
+
"grad_norm": 4.636094570159912,
|
| 1127 |
+
"learning_rate": 9.98700687614676e-05,
|
| 1128 |
+
"loss": 2.6198,
|
| 1129 |
+
"step": 80000
|
| 1130 |
+
},
|
| 1131 |
+
{
|
| 1132 |
+
"epoch": 0.04,
|
| 1133 |
+
"grad_norm": 3.763615846633911,
|
| 1134 |
+
"learning_rate": 9.986914068119236e-05,
|
| 1135 |
+
"loss": 2.611,
|
| 1136 |
+
"step": 80500
|
| 1137 |
+
},
|
| 1138 |
+
{
|
| 1139 |
+
"epoch": 0.05,
|
| 1140 |
+
"grad_norm": 4.402085781097412,
|
| 1141 |
+
"learning_rate": 9.986821260091714e-05,
|
| 1142 |
+
"loss": 2.6057,
|
| 1143 |
+
"step": 81000
|
| 1144 |
+
},
|
| 1145 |
+
{
|
| 1146 |
+
"epoch": 0.05,
|
| 1147 |
+
"grad_norm": 4.437091827392578,
|
| 1148 |
+
"learning_rate": 9.986728452064191e-05,
|
| 1149 |
+
"loss": 2.6322,
|
| 1150 |
+
"step": 81500
|
| 1151 |
+
},
|
| 1152 |
+
{
|
| 1153 |
+
"epoch": 0.05,
|
| 1154 |
+
"grad_norm": 4.0002923011779785,
|
| 1155 |
+
"learning_rate": 9.986635644036667e-05,
|
| 1156 |
+
"loss": 2.5895,
|
| 1157 |
+
"step": 82000
|
| 1158 |
+
},
|
| 1159 |
+
{
|
| 1160 |
+
"epoch": 0.05,
|
| 1161 |
+
"grad_norm": 7.929385662078857,
|
| 1162 |
+
"learning_rate": 9.986542836009144e-05,
|
| 1163 |
+
"loss": 2.5913,
|
| 1164 |
+
"step": 82500
|
| 1165 |
+
},
|
| 1166 |
+
{
|
| 1167 |
+
"epoch": 0.05,
|
| 1168 |
+
"grad_norm": 3.028870105743408,
|
| 1169 |
+
"learning_rate": 9.986450027981621e-05,
|
| 1170 |
+
"loss": 2.5852,
|
| 1171 |
+
"step": 83000
|
| 1172 |
+
},
|
| 1173 |
+
{
|
| 1174 |
+
"epoch": 0.05,
|
| 1175 |
+
"grad_norm": 4.816354274749756,
|
| 1176 |
+
"learning_rate": 9.986357219954098e-05,
|
| 1177 |
+
"loss": 2.612,
|
| 1178 |
+
"step": 83500
|
| 1179 |
+
},
|
| 1180 |
+
{
|
| 1181 |
+
"epoch": 0.05,
|
| 1182 |
+
"grad_norm": 3.2722418308258057,
|
| 1183 |
+
"learning_rate": 9.986264411926575e-05,
|
| 1184 |
+
"loss": 2.5844,
|
| 1185 |
+
"step": 84000
|
| 1186 |
+
},
|
| 1187 |
+
{
|
| 1188 |
+
"epoch": 0.05,
|
| 1189 |
+
"grad_norm": 4.4086713790893555,
|
| 1190 |
+
"learning_rate": 9.986171603899052e-05,
|
| 1191 |
+
"loss": 2.5766,
|
| 1192 |
+
"step": 84500
|
| 1193 |
+
},
|
| 1194 |
+
{
|
| 1195 |
+
"epoch": 0.05,
|
| 1196 |
+
"grad_norm": 4.107178211212158,
|
| 1197 |
+
"learning_rate": 9.986078795871528e-05,
|
| 1198 |
+
"loss": 2.5861,
|
| 1199 |
+
"step": 85000
|
| 1200 |
+
},
|
| 1201 |
+
{
|
| 1202 |
+
"epoch": 0.05,
|
| 1203 |
+
"grad_norm": 4.0130133628845215,
|
| 1204 |
+
"learning_rate": 9.985985987844005e-05,
|
| 1205 |
+
"loss": 2.5946,
|
| 1206 |
+
"step": 85500
|
| 1207 |
+
},
|
| 1208 |
+
{
|
| 1209 |
+
"epoch": 0.05,
|
| 1210 |
+
"grad_norm": 4.053595066070557,
|
| 1211 |
+
"learning_rate": 9.985893179816482e-05,
|
| 1212 |
+
"loss": 2.5411,
|
| 1213 |
+
"step": 86000
|
| 1214 |
+
},
|
| 1215 |
+
{
|
| 1216 |
+
"epoch": 0.05,
|
| 1217 |
+
"grad_norm": 4.136240482330322,
|
| 1218 |
+
"learning_rate": 9.98580037178896e-05,
|
| 1219 |
+
"loss": 2.5791,
|
| 1220 |
+
"step": 86500
|
| 1221 |
+
},
|
| 1222 |
+
{
|
| 1223 |
+
"epoch": 0.05,
|
| 1224 |
+
"grad_norm": 6.301804542541504,
|
| 1225 |
+
"learning_rate": 9.985707563761437e-05,
|
| 1226 |
+
"loss": 2.5621,
|
| 1227 |
+
"step": 87000
|
| 1228 |
+
},
|
| 1229 |
+
{
|
| 1230 |
+
"epoch": 0.05,
|
| 1231 |
+
"grad_norm": 4.376827239990234,
|
| 1232 |
+
"learning_rate": 9.985614755733912e-05,
|
| 1233 |
+
"loss": 2.5693,
|
| 1234 |
+
"step": 87500
|
| 1235 |
+
},
|
| 1236 |
+
{
|
| 1237 |
+
"epoch": 0.05,
|
| 1238 |
+
"grad_norm": 3.8407466411590576,
|
| 1239 |
+
"learning_rate": 9.985521947706388e-05,
|
| 1240 |
+
"loss": 2.5354,
|
| 1241 |
+
"step": 88000
|
| 1242 |
+
},
|
| 1243 |
+
{
|
| 1244 |
+
"epoch": 0.05,
|
| 1245 |
+
"grad_norm": 6.222740650177002,
|
| 1246 |
+
"learning_rate": 9.985429139678865e-05,
|
| 1247 |
+
"loss": 2.5402,
|
| 1248 |
+
"step": 88500
|
| 1249 |
+
},
|
| 1250 |
+
{
|
| 1251 |
+
"epoch": 0.05,
|
| 1252 |
+
"grad_norm": 5.93278169631958,
|
| 1253 |
+
"learning_rate": 9.985336331651342e-05,
|
| 1254 |
+
"loss": 2.5285,
|
| 1255 |
+
"step": 89000
|
| 1256 |
+
},
|
| 1257 |
+
{
|
| 1258 |
+
"epoch": 0.05,
|
| 1259 |
+
"grad_norm": 2.4788284301757812,
|
| 1260 |
+
"learning_rate": 9.98524352362382e-05,
|
| 1261 |
+
"loss": 2.511,
|
| 1262 |
+
"step": 89500
|
| 1263 |
+
},
|
| 1264 |
+
{
|
| 1265 |
+
"epoch": 0.05,
|
| 1266 |
+
"grad_norm": 3.4369866847991943,
|
| 1267 |
+
"learning_rate": 9.985150715596297e-05,
|
| 1268 |
+
"loss": 2.5223,
|
| 1269 |
+
"step": 90000
|
| 1270 |
+
},
|
| 1271 |
+
{
|
| 1272 |
+
"epoch": 0.05,
|
| 1273 |
+
"grad_norm": 5.007632732391357,
|
| 1274 |
+
"learning_rate": 9.985057907568774e-05,
|
| 1275 |
+
"loss": 2.5231,
|
| 1276 |
+
"step": 90500
|
| 1277 |
+
},
|
| 1278 |
+
{
|
| 1279 |
+
"epoch": 0.05,
|
| 1280 |
+
"grad_norm": 3.440267562866211,
|
| 1281 |
+
"learning_rate": 9.98496509954125e-05,
|
| 1282 |
+
"loss": 2.5233,
|
| 1283 |
+
"step": 91000
|
| 1284 |
+
},
|
| 1285 |
+
{
|
| 1286 |
+
"epoch": 0.05,
|
| 1287 |
+
"grad_norm": 9.757936477661133,
|
| 1288 |
+
"learning_rate": 9.984872291513727e-05,
|
| 1289 |
+
"loss": 2.5099,
|
| 1290 |
+
"step": 91500
|
| 1291 |
+
},
|
| 1292 |
+
{
|
| 1293 |
+
"epoch": 0.05,
|
| 1294 |
+
"grad_norm": 4.741192817687988,
|
| 1295 |
+
"learning_rate": 9.984779483486204e-05,
|
| 1296 |
+
"loss": 2.5212,
|
| 1297 |
+
"step": 92000
|
| 1298 |
+
},
|
| 1299 |
+
{
|
| 1300 |
+
"epoch": 0.05,
|
| 1301 |
+
"grad_norm": 4.7662811279296875,
|
| 1302 |
+
"learning_rate": 9.984686675458681e-05,
|
| 1303 |
+
"loss": 2.5023,
|
| 1304 |
+
"step": 92500
|
| 1305 |
+
},
|
| 1306 |
+
{
|
| 1307 |
+
"epoch": 0.05,
|
| 1308 |
+
"grad_norm": 4.681964874267578,
|
| 1309 |
+
"learning_rate": 9.984593867431158e-05,
|
| 1310 |
+
"loss": 2.5055,
|
| 1311 |
+
"step": 93000
|
| 1312 |
+
},
|
| 1313 |
+
{
|
| 1314 |
+
"epoch": 0.05,
|
| 1315 |
+
"grad_norm": 3.7245185375213623,
|
| 1316 |
+
"learning_rate": 9.984501059403635e-05,
|
| 1317 |
+
"loss": 2.5083,
|
| 1318 |
+
"step": 93500
|
| 1319 |
+
},
|
| 1320 |
+
{
|
| 1321 |
+
"epoch": 0.05,
|
| 1322 |
+
"grad_norm": 5.7924418449401855,
|
| 1323 |
+
"learning_rate": 9.984408251376111e-05,
|
| 1324 |
+
"loss": 2.5024,
|
| 1325 |
+
"step": 94000
|
| 1326 |
+
},
|
| 1327 |
+
{
|
| 1328 |
+
"epoch": 0.05,
|
| 1329 |
+
"grad_norm": 5.420963764190674,
|
| 1330 |
+
"learning_rate": 9.984315443348588e-05,
|
| 1331 |
+
"loss": 2.5053,
|
| 1332 |
+
"step": 94500
|
| 1333 |
+
},
|
| 1334 |
+
{
|
| 1335 |
+
"epoch": 0.05,
|
| 1336 |
+
"grad_norm": 4.608907699584961,
|
| 1337 |
+
"learning_rate": 9.984222635321065e-05,
|
| 1338 |
+
"loss": 2.4765,
|
| 1339 |
+
"step": 95000
|
| 1340 |
+
},
|
| 1341 |
+
{
|
| 1342 |
+
"epoch": 0.05,
|
| 1343 |
+
"grad_norm": 5.017517566680908,
|
| 1344 |
+
"learning_rate": 9.984129827293543e-05,
|
| 1345 |
+
"loss": 2.5004,
|
| 1346 |
+
"step": 95500
|
| 1347 |
+
},
|
| 1348 |
+
{
|
| 1349 |
+
"epoch": 0.05,
|
| 1350 |
+
"grad_norm": 6.300387859344482,
|
| 1351 |
+
"learning_rate": 9.98403701926602e-05,
|
| 1352 |
+
"loss": 2.4943,
|
| 1353 |
+
"step": 96000
|
| 1354 |
+
},
|
| 1355 |
+
{
|
| 1356 |
+
"epoch": 0.05,
|
| 1357 |
+
"grad_norm": 6.16803503036499,
|
| 1358 |
+
"learning_rate": 9.983944211238497e-05,
|
| 1359 |
+
"loss": 2.5001,
|
| 1360 |
+
"step": 96500
|
| 1361 |
+
},
|
| 1362 |
+
{
|
| 1363 |
+
"epoch": 0.05,
|
| 1364 |
+
"grad_norm": 4.007481098175049,
|
| 1365 |
+
"learning_rate": 9.983851403210973e-05,
|
| 1366 |
+
"loss": 2.4898,
|
| 1367 |
+
"step": 97000
|
| 1368 |
+
},
|
| 1369 |
+
{
|
| 1370 |
+
"epoch": 0.05,
|
| 1371 |
+
"grad_norm": 5.498426914215088,
|
| 1372 |
+
"learning_rate": 9.983758595183448e-05,
|
| 1373 |
+
"loss": 2.4874,
|
| 1374 |
+
"step": 97500
|
| 1375 |
+
},
|
| 1376 |
+
{
|
| 1377 |
+
"epoch": 0.05,
|
| 1378 |
+
"grad_norm": 4.115726470947266,
|
| 1379 |
+
"learning_rate": 9.983665787155926e-05,
|
| 1380 |
+
"loss": 2.4714,
|
| 1381 |
+
"step": 98000
|
| 1382 |
+
},
|
| 1383 |
+
{
|
| 1384 |
+
"epoch": 0.05,
|
| 1385 |
+
"grad_norm": 4.724228382110596,
|
| 1386 |
+
"learning_rate": 9.983572979128403e-05,
|
| 1387 |
+
"loss": 2.4748,
|
| 1388 |
+
"step": 98500
|
| 1389 |
+
},
|
| 1390 |
+
{
|
| 1391 |
+
"epoch": 0.06,
|
| 1392 |
+
"grad_norm": 6.842497825622559,
|
| 1393 |
+
"learning_rate": 9.98348017110088e-05,
|
| 1394 |
+
"loss": 2.4761,
|
| 1395 |
+
"step": 99000
|
| 1396 |
+
},
|
| 1397 |
+
{
|
| 1398 |
+
"epoch": 0.06,
|
| 1399 |
+
"grad_norm": 4.527170181274414,
|
| 1400 |
+
"learning_rate": 9.983387363073357e-05,
|
| 1401 |
+
"loss": 2.462,
|
| 1402 |
+
"step": 99500
|
| 1403 |
+
},
|
| 1404 |
+
{
|
| 1405 |
+
"epoch": 0.06,
|
| 1406 |
+
"grad_norm": 4.956340789794922,
|
| 1407 |
+
"learning_rate": 9.983294555045834e-05,
|
| 1408 |
+
"loss": 2.4568,
|
| 1409 |
+
"step": 100000
|
| 1410 |
+
},
|
| 1411 |
+
{
|
| 1412 |
+
"epoch": 0.06,
|
| 1413 |
+
"grad_norm": 4.177946090698242,
|
| 1414 |
+
"learning_rate": 9.98320174701831e-05,
|
| 1415 |
+
"loss": 2.4418,
|
| 1416 |
+
"step": 100500
|
| 1417 |
+
},
|
| 1418 |
+
{
|
| 1419 |
+
"epoch": 0.06,
|
| 1420 |
+
"grad_norm": 6.623388767242432,
|
| 1421 |
+
"learning_rate": 9.983108938990787e-05,
|
| 1422 |
+
"loss": 2.4467,
|
| 1423 |
+
"step": 101000
|
| 1424 |
+
},
|
| 1425 |
+
{
|
| 1426 |
+
"epoch": 0.06,
|
| 1427 |
+
"grad_norm": 6.102182865142822,
|
| 1428 |
+
"learning_rate": 9.983016130963264e-05,
|
| 1429 |
+
"loss": 2.444,
|
| 1430 |
+
"step": 101500
|
| 1431 |
+
},
|
| 1432 |
+
{
|
| 1433 |
+
"epoch": 0.06,
|
| 1434 |
+
"grad_norm": 4.186117649078369,
|
| 1435 |
+
"learning_rate": 9.982923322935741e-05,
|
| 1436 |
+
"loss": 2.4396,
|
| 1437 |
+
"step": 102000
|
| 1438 |
+
},
|
| 1439 |
+
{
|
| 1440 |
+
"epoch": 0.06,
|
| 1441 |
+
"grad_norm": 3.916994571685791,
|
| 1442 |
+
"learning_rate": 9.982830514908218e-05,
|
| 1443 |
+
"loss": 2.4612,
|
| 1444 |
+
"step": 102500
|
| 1445 |
+
},
|
| 1446 |
+
{
|
| 1447 |
+
"epoch": 0.06,
|
| 1448 |
+
"grad_norm": 4.213066577911377,
|
| 1449 |
+
"learning_rate": 9.982737706880694e-05,
|
| 1450 |
+
"loss": 2.4244,
|
| 1451 |
+
"step": 103000
|
| 1452 |
+
},
|
| 1453 |
+
{
|
| 1454 |
+
"epoch": 0.06,
|
| 1455 |
+
"grad_norm": 4.665497779846191,
|
| 1456 |
+
"learning_rate": 9.982644898853171e-05,
|
| 1457 |
+
"loss": 2.4364,
|
| 1458 |
+
"step": 103500
|
| 1459 |
+
},
|
| 1460 |
+
{
|
| 1461 |
+
"epoch": 0.06,
|
| 1462 |
+
"grad_norm": 4.808815956115723,
|
| 1463 |
+
"learning_rate": 9.982552090825648e-05,
|
| 1464 |
+
"loss": 2.4381,
|
| 1465 |
+
"step": 104000
|
| 1466 |
+
},
|
| 1467 |
+
{
|
| 1468 |
+
"epoch": 0.06,
|
| 1469 |
+
"grad_norm": 5.384184837341309,
|
| 1470 |
+
"learning_rate": 9.982459282798126e-05,
|
| 1471 |
+
"loss": 2.4219,
|
| 1472 |
+
"step": 104500
|
| 1473 |
+
},
|
| 1474 |
+
{
|
| 1475 |
+
"epoch": 0.06,
|
| 1476 |
+
"grad_norm": 8.32588005065918,
|
| 1477 |
+
"learning_rate": 9.982366474770603e-05,
|
| 1478 |
+
"loss": 2.4207,
|
| 1479 |
+
"step": 105000
|
| 1480 |
+
},
|
| 1481 |
+
{
|
| 1482 |
+
"epoch": 0.06,
|
| 1483 |
+
"grad_norm": 6.564486503601074,
|
| 1484 |
+
"learning_rate": 9.98227366674308e-05,
|
| 1485 |
+
"loss": 2.4182,
|
| 1486 |
+
"step": 105500
|
| 1487 |
+
},
|
| 1488 |
+
{
|
| 1489 |
+
"epoch": 0.06,
|
| 1490 |
+
"grad_norm": 4.123614311218262,
|
| 1491 |
+
"learning_rate": 9.982180858715556e-05,
|
| 1492 |
+
"loss": 2.4214,
|
| 1493 |
+
"step": 106000
|
| 1494 |
+
},
|
| 1495 |
+
{
|
| 1496 |
+
"epoch": 0.06,
|
| 1497 |
+
"grad_norm": 5.507079124450684,
|
| 1498 |
+
"learning_rate": 9.982088050688033e-05,
|
| 1499 |
+
"loss": 2.4096,
|
| 1500 |
+
"step": 106500
|
| 1501 |
+
},
|
| 1502 |
+
{
|
| 1503 |
+
"epoch": 0.06,
|
| 1504 |
+
"grad_norm": 5.728691577911377,
|
| 1505 |
+
"learning_rate": 9.98199524266051e-05,
|
| 1506 |
+
"loss": 2.4201,
|
| 1507 |
+
"step": 107000
|
| 1508 |
+
},
|
| 1509 |
+
{
|
| 1510 |
+
"epoch": 0.06,
|
| 1511 |
+
"grad_norm": 6.611893177032471,
|
| 1512 |
+
"learning_rate": 9.981902434632987e-05,
|
| 1513 |
+
"loss": 2.4021,
|
| 1514 |
+
"step": 107500
|
| 1515 |
+
},
|
| 1516 |
+
{
|
| 1517 |
+
"epoch": 0.06,
|
| 1518 |
+
"grad_norm": 5.187854766845703,
|
| 1519 |
+
"learning_rate": 9.981809626605463e-05,
|
| 1520 |
+
"loss": 2.3956,
|
| 1521 |
+
"step": 108000
|
| 1522 |
+
},
|
| 1523 |
+
{
|
| 1524 |
+
"epoch": 0.06,
|
| 1525 |
+
"grad_norm": 6.5400166511535645,
|
| 1526 |
+
"learning_rate": 9.98171681857794e-05,
|
| 1527 |
+
"loss": 2.4168,
|
| 1528 |
+
"step": 108500
|
| 1529 |
+
},
|
| 1530 |
+
{
|
| 1531 |
+
"epoch": 0.06,
|
| 1532 |
+
"grad_norm": 4.0128173828125,
|
| 1533 |
+
"learning_rate": 9.981624010550417e-05,
|
| 1534 |
+
"loss": 2.4017,
|
| 1535 |
+
"step": 109000
|
| 1536 |
+
},
|
| 1537 |
+
{
|
| 1538 |
+
"epoch": 0.06,
|
| 1539 |
+
"grad_norm": 3.7369205951690674,
|
| 1540 |
+
"learning_rate": 9.981531202522893e-05,
|
| 1541 |
+
"loss": 2.3895,
|
| 1542 |
+
"step": 109500
|
| 1543 |
+
},
|
| 1544 |
+
{
|
| 1545 |
+
"epoch": 0.06,
|
| 1546 |
+
"grad_norm": 3.8515870571136475,
|
| 1547 |
+
"learning_rate": 9.98143839449537e-05,
|
| 1548 |
+
"loss": 2.3728,
|
| 1549 |
+
"step": 110000
|
| 1550 |
+
},
|
| 1551 |
+
{
|
| 1552 |
+
"epoch": 0.06,
|
| 1553 |
+
"grad_norm": 4.47413969039917,
|
| 1554 |
+
"learning_rate": 9.981345586467847e-05,
|
| 1555 |
+
"loss": 2.4067,
|
| 1556 |
+
"step": 110500
|
| 1557 |
+
},
|
| 1558 |
+
{
|
| 1559 |
+
"epoch": 0.06,
|
| 1560 |
+
"grad_norm": 4.970263481140137,
|
| 1561 |
+
"learning_rate": 9.981252778440324e-05,
|
| 1562 |
+
"loss": 2.3775,
|
| 1563 |
+
"step": 111000
|
| 1564 |
+
},
|
| 1565 |
+
{
|
| 1566 |
+
"epoch": 0.06,
|
| 1567 |
+
"grad_norm": 4.086507320404053,
|
| 1568 |
+
"learning_rate": 9.981159970412802e-05,
|
| 1569 |
+
"loss": 2.3695,
|
| 1570 |
+
"step": 111500
|
| 1571 |
+
},
|
| 1572 |
+
{
|
| 1573 |
+
"epoch": 0.06,
|
| 1574 |
+
"grad_norm": 4.484976768493652,
|
| 1575 |
+
"learning_rate": 9.981067162385279e-05,
|
| 1576 |
+
"loss": 2.3774,
|
| 1577 |
+
"step": 112000
|
| 1578 |
+
},
|
| 1579 |
+
{
|
| 1580 |
+
"epoch": 0.06,
|
| 1581 |
+
"grad_norm": 4.321993350982666,
|
| 1582 |
+
"learning_rate": 9.980974354357754e-05,
|
| 1583 |
+
"loss": 2.3563,
|
| 1584 |
+
"step": 112500
|
| 1585 |
+
},
|
| 1586 |
+
{
|
| 1587 |
+
"epoch": 0.06,
|
| 1588 |
+
"grad_norm": 5.6485066413879395,
|
| 1589 |
+
"learning_rate": 9.980881546330232e-05,
|
| 1590 |
+
"loss": 2.3601,
|
| 1591 |
+
"step": 113000
|
| 1592 |
+
},
|
| 1593 |
+
{
|
| 1594 |
+
"epoch": 0.06,
|
| 1595 |
+
"grad_norm": 6.259582996368408,
|
| 1596 |
+
"learning_rate": 9.980788738302709e-05,
|
| 1597 |
+
"loss": 2.3409,
|
| 1598 |
+
"step": 113500
|
| 1599 |
+
},
|
| 1600 |
+
{
|
| 1601 |
+
"epoch": 0.06,
|
| 1602 |
+
"grad_norm": 5.368774890899658,
|
| 1603 |
+
"learning_rate": 9.980695930275186e-05,
|
| 1604 |
+
"loss": 2.3809,
|
| 1605 |
+
"step": 114000
|
| 1606 |
+
},
|
| 1607 |
+
{
|
| 1608 |
+
"epoch": 0.06,
|
| 1609 |
+
"grad_norm": 7.465442657470703,
|
| 1610 |
+
"learning_rate": 9.980603122247663e-05,
|
| 1611 |
+
"loss": 2.3439,
|
| 1612 |
+
"step": 114500
|
| 1613 |
+
},
|
| 1614 |
+
{
|
| 1615 |
+
"epoch": 0.06,
|
| 1616 |
+
"grad_norm": 10.82461929321289,
|
| 1617 |
+
"learning_rate": 9.98051031422014e-05,
|
| 1618 |
+
"loss": 2.3738,
|
| 1619 |
+
"step": 115000
|
| 1620 |
+
},
|
| 1621 |
+
{
|
| 1622 |
+
"epoch": 0.06,
|
| 1623 |
+
"grad_norm": 3.924154758453369,
|
| 1624 |
+
"learning_rate": 9.980417506192616e-05,
|
| 1625 |
+
"loss": 2.344,
|
| 1626 |
+
"step": 115500
|
| 1627 |
+
},
|
| 1628 |
+
{
|
| 1629 |
+
"epoch": 0.06,
|
| 1630 |
+
"grad_norm": 3.8393290042877197,
|
| 1631 |
+
"learning_rate": 9.980324698165093e-05,
|
| 1632 |
+
"loss": 2.3423,
|
| 1633 |
+
"step": 116000
|
| 1634 |
+
},
|
| 1635 |
+
{
|
| 1636 |
+
"epoch": 0.06,
|
| 1637 |
+
"grad_norm": 3.9444777965545654,
|
| 1638 |
+
"learning_rate": 9.98023189013757e-05,
|
| 1639 |
+
"loss": 2.3426,
|
| 1640 |
+
"step": 116500
|
| 1641 |
+
},
|
| 1642 |
+
{
|
| 1643 |
+
"epoch": 0.07,
|
| 1644 |
+
"grad_norm": 5.0410566329956055,
|
| 1645 |
+
"learning_rate": 9.980139082110047e-05,
|
| 1646 |
+
"loss": 2.3129,
|
| 1647 |
+
"step": 117000
|
| 1648 |
+
},
|
| 1649 |
+
{
|
| 1650 |
+
"epoch": 0.07,
|
| 1651 |
+
"grad_norm": 5.183213233947754,
|
| 1652 |
+
"learning_rate": 9.980046274082524e-05,
|
| 1653 |
+
"loss": 2.3187,
|
| 1654 |
+
"step": 117500
|
| 1655 |
+
},
|
| 1656 |
+
{
|
| 1657 |
+
"epoch": 0.07,
|
| 1658 |
+
"grad_norm": 6.824711322784424,
|
| 1659 |
+
"learning_rate": 9.979953466055e-05,
|
| 1660 |
+
"loss": 2.3206,
|
| 1661 |
+
"step": 118000
|
| 1662 |
+
},
|
| 1663 |
+
{
|
| 1664 |
+
"epoch": 0.07,
|
| 1665 |
+
"grad_norm": 4.512765884399414,
|
| 1666 |
+
"learning_rate": 9.979860658027476e-05,
|
| 1667 |
+
"loss": 2.3176,
|
| 1668 |
+
"step": 118500
|
| 1669 |
+
},
|
| 1670 |
+
{
|
| 1671 |
+
"epoch": 0.07,
|
| 1672 |
+
"grad_norm": 6.9497480392456055,
|
| 1673 |
+
"learning_rate": 9.979767849999953e-05,
|
| 1674 |
+
"loss": 2.3418,
|
| 1675 |
+
"step": 119000
|
| 1676 |
+
},
|
| 1677 |
+
{
|
| 1678 |
+
"epoch": 0.07,
|
| 1679 |
+
"grad_norm": 4.003693580627441,
|
| 1680 |
+
"learning_rate": 9.97967504197243e-05,
|
| 1681 |
+
"loss": 2.3193,
|
| 1682 |
+
"step": 119500
|
| 1683 |
+
},
|
| 1684 |
+
{
|
| 1685 |
+
"epoch": 0.07,
|
| 1686 |
+
"grad_norm": 6.719389915466309,
|
| 1687 |
+
"learning_rate": 9.979582233944907e-05,
|
| 1688 |
+
"loss": 2.3244,
|
| 1689 |
+
"step": 120000
|
| 1690 |
+
},
|
| 1691 |
+
{
|
| 1692 |
+
"epoch": 0.07,
|
| 1693 |
+
"grad_norm": 4.46160364151001,
|
| 1694 |
+
"learning_rate": 9.979489425917385e-05,
|
| 1695 |
+
"loss": 2.3313,
|
| 1696 |
+
"step": 120500
|
| 1697 |
+
},
|
| 1698 |
+
{
|
| 1699 |
+
"epoch": 0.07,
|
| 1700 |
+
"grad_norm": 8.168338775634766,
|
| 1701 |
+
"learning_rate": 9.979396617889862e-05,
|
| 1702 |
+
"loss": 2.311,
|
| 1703 |
+
"step": 121000
|
| 1704 |
+
},
|
| 1705 |
+
{
|
| 1706 |
+
"epoch": 0.07,
|
| 1707 |
+
"grad_norm": 4.333712100982666,
|
| 1708 |
+
"learning_rate": 9.979303809862338e-05,
|
| 1709 |
+
"loss": 2.328,
|
| 1710 |
+
"step": 121500
|
| 1711 |
+
},
|
| 1712 |
+
{
|
| 1713 |
+
"epoch": 0.07,
|
| 1714 |
+
"grad_norm": 8.339402198791504,
|
| 1715 |
+
"learning_rate": 9.979211001834815e-05,
|
| 1716 |
+
"loss": 2.3118,
|
| 1717 |
+
"step": 122000
|
| 1718 |
+
},
|
| 1719 |
+
{
|
| 1720 |
+
"epoch": 0.07,
|
| 1721 |
+
"grad_norm": 4.38870906829834,
|
| 1722 |
+
"learning_rate": 9.979118193807292e-05,
|
| 1723 |
+
"loss": 2.3037,
|
| 1724 |
+
"step": 122500
|
| 1725 |
+
},
|
| 1726 |
+
{
|
| 1727 |
+
"epoch": 0.07,
|
| 1728 |
+
"grad_norm": 4.003161430358887,
|
| 1729 |
+
"learning_rate": 9.979025385779769e-05,
|
| 1730 |
+
"loss": 2.2959,
|
| 1731 |
+
"step": 123000
|
| 1732 |
+
},
|
| 1733 |
+
{
|
| 1734 |
+
"epoch": 0.07,
|
| 1735 |
+
"grad_norm": 6.8454389572143555,
|
| 1736 |
+
"learning_rate": 9.978932577752246e-05,
|
| 1737 |
+
"loss": 2.3192,
|
| 1738 |
+
"step": 123500
|
| 1739 |
+
},
|
| 1740 |
+
{
|
| 1741 |
+
"epoch": 0.07,
|
| 1742 |
+
"grad_norm": 6.631998538970947,
|
| 1743 |
+
"learning_rate": 9.978839769724723e-05,
|
| 1744 |
+
"loss": 2.2975,
|
| 1745 |
+
"step": 124000
|
| 1746 |
+
},
|
| 1747 |
+
{
|
| 1748 |
+
"epoch": 0.07,
|
| 1749 |
+
"grad_norm": 4.8582963943481445,
|
| 1750 |
+
"learning_rate": 9.978746961697199e-05,
|
| 1751 |
+
"loss": 2.2983,
|
| 1752 |
+
"step": 124500
|
| 1753 |
+
},
|
| 1754 |
+
{
|
| 1755 |
+
"epoch": 0.07,
|
| 1756 |
+
"grad_norm": 3.7582802772521973,
|
| 1757 |
+
"learning_rate": 9.978654153669676e-05,
|
| 1758 |
+
"loss": 2.3105,
|
| 1759 |
+
"step": 125000
|
| 1760 |
+
},
|
| 1761 |
+
{
|
| 1762 |
+
"epoch": 0.07,
|
| 1763 |
+
"grad_norm": 4.227505683898926,
|
| 1764 |
+
"learning_rate": 9.978561345642153e-05,
|
| 1765 |
+
"loss": 2.2866,
|
| 1766 |
+
"step": 125500
|
| 1767 |
+
},
|
| 1768 |
+
{
|
| 1769 |
+
"epoch": 0.07,
|
| 1770 |
+
"grad_norm": 5.702460765838623,
|
| 1771 |
+
"learning_rate": 9.97846853761463e-05,
|
| 1772 |
+
"loss": 2.2854,
|
| 1773 |
+
"step": 126000
|
| 1774 |
+
},
|
| 1775 |
+
{
|
| 1776 |
+
"epoch": 0.07,
|
| 1777 |
+
"grad_norm": 4.512174129486084,
|
| 1778 |
+
"learning_rate": 9.978375729587108e-05,
|
| 1779 |
+
"loss": 2.2962,
|
| 1780 |
+
"step": 126500
|
| 1781 |
+
},
|
| 1782 |
+
{
|
| 1783 |
+
"epoch": 0.07,
|
| 1784 |
+
"grad_norm": 4.127579212188721,
|
| 1785 |
+
"learning_rate": 9.978282921559585e-05,
|
| 1786 |
+
"loss": 2.3017,
|
| 1787 |
+
"step": 127000
|
| 1788 |
+
},
|
| 1789 |
+
{
|
| 1790 |
+
"epoch": 0.07,
|
| 1791 |
+
"grad_norm": 4.7636284828186035,
|
| 1792 |
+
"learning_rate": 9.97819011353206e-05,
|
| 1793 |
+
"loss": 2.26,
|
| 1794 |
+
"step": 127500
|
| 1795 |
+
},
|
| 1796 |
+
{
|
| 1797 |
+
"epoch": 0.07,
|
| 1798 |
+
"grad_norm": 5.8895697593688965,
|
| 1799 |
+
"learning_rate": 9.978097305504538e-05,
|
| 1800 |
+
"loss": 2.2851,
|
| 1801 |
+
"step": 128000
|
| 1802 |
+
},
|
| 1803 |
+
{
|
| 1804 |
+
"epoch": 0.07,
|
| 1805 |
+
"grad_norm": 5.10247278213501,
|
| 1806 |
+
"learning_rate": 9.978004497477013e-05,
|
| 1807 |
+
"loss": 2.2912,
|
| 1808 |
+
"step": 128500
|
| 1809 |
+
},
|
| 1810 |
+
{
|
| 1811 |
+
"epoch": 0.07,
|
| 1812 |
+
"grad_norm": 4.03781795501709,
|
| 1813 |
+
"learning_rate": 9.97791168944949e-05,
|
| 1814 |
+
"loss": 2.2631,
|
| 1815 |
+
"step": 129000
|
| 1816 |
+
},
|
| 1817 |
+
{
|
| 1818 |
+
"epoch": 0.07,
|
| 1819 |
+
"grad_norm": 6.225619316101074,
|
| 1820 |
+
"learning_rate": 9.977818881421968e-05,
|
| 1821 |
+
"loss": 2.2959,
|
| 1822 |
+
"step": 129500
|
| 1823 |
+
},
|
| 1824 |
+
{
|
| 1825 |
+
"epoch": 0.07,
|
| 1826 |
+
"grad_norm": 8.12752628326416,
|
| 1827 |
+
"learning_rate": 9.977726073394445e-05,
|
| 1828 |
+
"loss": 2.2574,
|
| 1829 |
+
"step": 130000
|
| 1830 |
+
},
|
| 1831 |
+
{
|
| 1832 |
+
"epoch": 0.07,
|
| 1833 |
+
"grad_norm": 8.98240852355957,
|
| 1834 |
+
"learning_rate": 9.977633265366922e-05,
|
| 1835 |
+
"loss": 2.2765,
|
| 1836 |
+
"step": 130500
|
| 1837 |
+
},
|
| 1838 |
+
{
|
| 1839 |
+
"epoch": 0.07,
|
| 1840 |
+
"grad_norm": 6.665409088134766,
|
| 1841 |
+
"learning_rate": 9.977540457339398e-05,
|
| 1842 |
+
"loss": 2.2831,
|
| 1843 |
+
"step": 131000
|
| 1844 |
+
},
|
| 1845 |
+
{
|
| 1846 |
+
"epoch": 0.07,
|
| 1847 |
+
"grad_norm": 5.666757106781006,
|
| 1848 |
+
"learning_rate": 9.977447649311875e-05,
|
| 1849 |
+
"loss": 2.2514,
|
| 1850 |
+
"step": 131500
|
| 1851 |
+
},
|
| 1852 |
+
{
|
| 1853 |
+
"epoch": 0.07,
|
| 1854 |
+
"grad_norm": 4.039821147918701,
|
| 1855 |
+
"learning_rate": 9.977354841284352e-05,
|
| 1856 |
+
"loss": 2.2553,
|
| 1857 |
+
"step": 132000
|
| 1858 |
+
},
|
| 1859 |
+
{
|
| 1860 |
+
"epoch": 0.07,
|
| 1861 |
+
"grad_norm": 4.816211223602295,
|
| 1862 |
+
"learning_rate": 9.977262033256829e-05,
|
| 1863 |
+
"loss": 2.2512,
|
| 1864 |
+
"step": 132500
|
| 1865 |
+
},
|
| 1866 |
+
{
|
| 1867 |
+
"epoch": 0.07,
|
| 1868 |
+
"grad_norm": 7.379537105560303,
|
| 1869 |
+
"learning_rate": 9.977169225229306e-05,
|
| 1870 |
+
"loss": 2.256,
|
| 1871 |
+
"step": 133000
|
| 1872 |
+
},
|
| 1873 |
+
{
|
| 1874 |
+
"epoch": 0.07,
|
| 1875 |
+
"grad_norm": 3.727262258529663,
|
| 1876 |
+
"learning_rate": 9.977076417201782e-05,
|
| 1877 |
+
"loss": 2.2447,
|
| 1878 |
+
"step": 133500
|
| 1879 |
+
},
|
| 1880 |
+
{
|
| 1881 |
+
"epoch": 0.07,
|
| 1882 |
+
"grad_norm": 4.287083625793457,
|
| 1883 |
+
"learning_rate": 9.976983609174259e-05,
|
| 1884 |
+
"loss": 2.233,
|
| 1885 |
+
"step": 134000
|
| 1886 |
+
},
|
| 1887 |
+
{
|
| 1888 |
+
"epoch": 0.07,
|
| 1889 |
+
"grad_norm": 4.225050926208496,
|
| 1890 |
+
"learning_rate": 9.976890801146736e-05,
|
| 1891 |
+
"loss": 2.2474,
|
| 1892 |
+
"step": 134500
|
| 1893 |
+
},
|
| 1894 |
+
{
|
| 1895 |
+
"epoch": 0.08,
|
| 1896 |
+
"grad_norm": 5.888236999511719,
|
| 1897 |
+
"learning_rate": 9.976797993119214e-05,
|
| 1898 |
+
"loss": 2.2187,
|
| 1899 |
+
"step": 135000
|
| 1900 |
+
},
|
| 1901 |
+
{
|
| 1902 |
+
"epoch": 0.08,
|
| 1903 |
+
"grad_norm": 5.869006156921387,
|
| 1904 |
+
"learning_rate": 9.97670518509169e-05,
|
| 1905 |
+
"loss": 2.2442,
|
| 1906 |
+
"step": 135500
|
| 1907 |
+
},
|
| 1908 |
+
{
|
| 1909 |
+
"epoch": 0.08,
|
| 1910 |
+
"grad_norm": 6.458480358123779,
|
| 1911 |
+
"learning_rate": 9.976612377064168e-05,
|
| 1912 |
+
"loss": 2.2501,
|
| 1913 |
+
"step": 136000
|
| 1914 |
+
},
|
| 1915 |
+
{
|
| 1916 |
+
"epoch": 0.08,
|
| 1917 |
+
"grad_norm": 3.7204341888427734,
|
| 1918 |
+
"learning_rate": 9.976519569036644e-05,
|
| 1919 |
+
"loss": 2.2464,
|
| 1920 |
+
"step": 136500
|
| 1921 |
+
},
|
| 1922 |
+
{
|
| 1923 |
+
"epoch": 0.08,
|
| 1924 |
+
"grad_norm": 4.325768947601318,
|
| 1925 |
+
"learning_rate": 9.976426761009121e-05,
|
| 1926 |
+
"loss": 2.2421,
|
| 1927 |
+
"step": 137000
|
| 1928 |
+
},
|
| 1929 |
+
{
|
| 1930 |
+
"epoch": 0.08,
|
| 1931 |
+
"grad_norm": 4.79429292678833,
|
| 1932 |
+
"learning_rate": 9.976333952981598e-05,
|
| 1933 |
+
"loss": 2.2466,
|
| 1934 |
+
"step": 137500
|
| 1935 |
+
},
|
| 1936 |
+
{
|
| 1937 |
+
"epoch": 0.08,
|
| 1938 |
+
"grad_norm": 8.317536354064941,
|
| 1939 |
+
"learning_rate": 9.976241144954075e-05,
|
| 1940 |
+
"loss": 2.2169,
|
| 1941 |
+
"step": 138000
|
| 1942 |
+
},
|
| 1943 |
+
{
|
| 1944 |
+
"epoch": 0.08,
|
| 1945 |
+
"grad_norm": 5.129164695739746,
|
| 1946 |
+
"learning_rate": 9.976148336926552e-05,
|
| 1947 |
+
"loss": 2.2188,
|
| 1948 |
+
"step": 138500
|
| 1949 |
+
},
|
| 1950 |
+
{
|
| 1951 |
+
"epoch": 0.08,
|
| 1952 |
+
"grad_norm": 3.7128493785858154,
|
| 1953 |
+
"learning_rate": 9.976055528899028e-05,
|
| 1954 |
+
"loss": 2.2372,
|
| 1955 |
+
"step": 139000
|
| 1956 |
+
},
|
| 1957 |
+
{
|
| 1958 |
+
"epoch": 0.08,
|
| 1959 |
+
"grad_norm": 4.394794940948486,
|
| 1960 |
+
"learning_rate": 9.975962720871505e-05,
|
| 1961 |
+
"loss": 2.2277,
|
| 1962 |
+
"step": 139500
|
| 1963 |
+
},
|
| 1964 |
+
{
|
| 1965 |
+
"epoch": 0.08,
|
| 1966 |
+
"grad_norm": 6.397000789642334,
|
| 1967 |
+
"learning_rate": 9.975869912843981e-05,
|
| 1968 |
+
"loss": 2.2509,
|
| 1969 |
+
"step": 140000
|
| 1970 |
+
},
|
| 1971 |
+
{
|
| 1972 |
+
"epoch": 0.08,
|
| 1973 |
+
"grad_norm": 5.2059526443481445,
|
| 1974 |
+
"learning_rate": 9.975777104816458e-05,
|
| 1975 |
+
"loss": 2.2082,
|
| 1976 |
+
"step": 140500
|
| 1977 |
+
},
|
| 1978 |
+
{
|
| 1979 |
+
"epoch": 0.08,
|
| 1980 |
+
"grad_norm": 3.5649585723876953,
|
| 1981 |
+
"learning_rate": 9.975684296788935e-05,
|
| 1982 |
+
"loss": 2.2383,
|
| 1983 |
+
"step": 141000
|
| 1984 |
+
},
|
| 1985 |
+
{
|
| 1986 |
+
"epoch": 0.08,
|
| 1987 |
+
"grad_norm": 4.717801094055176,
|
| 1988 |
+
"learning_rate": 9.975591488761412e-05,
|
| 1989 |
+
"loss": 2.1973,
|
| 1990 |
+
"step": 141500
|
| 1991 |
+
},
|
| 1992 |
+
{
|
| 1993 |
+
"epoch": 0.08,
|
| 1994 |
+
"grad_norm": 4.133371829986572,
|
| 1995 |
+
"learning_rate": 9.97549868073389e-05,
|
| 1996 |
+
"loss": 2.2214,
|
| 1997 |
+
"step": 142000
|
| 1998 |
+
},
|
| 1999 |
+
{
|
| 2000 |
+
"epoch": 0.08,
|
| 2001 |
+
"grad_norm": 5.321709156036377,
|
| 2002 |
+
"learning_rate": 9.975405872706367e-05,
|
| 2003 |
+
"loss": 2.2012,
|
| 2004 |
+
"step": 142500
|
| 2005 |
+
},
|
| 2006 |
+
{
|
| 2007 |
+
"epoch": 0.08,
|
| 2008 |
+
"grad_norm": 5.209156513214111,
|
| 2009 |
+
"learning_rate": 9.975313064678842e-05,
|
| 2010 |
+
"loss": 2.1892,
|
| 2011 |
+
"step": 143000
|
| 2012 |
+
},
|
| 2013 |
+
{
|
| 2014 |
+
"epoch": 0.08,
|
| 2015 |
+
"grad_norm": 5.405091285705566,
|
| 2016 |
+
"learning_rate": 9.97522025665132e-05,
|
| 2017 |
+
"loss": 2.217,
|
| 2018 |
+
"step": 143500
|
| 2019 |
+
},
|
| 2020 |
+
{
|
| 2021 |
+
"epoch": 0.08,
|
| 2022 |
+
"grad_norm": 5.35112190246582,
|
| 2023 |
+
"learning_rate": 9.975127448623797e-05,
|
| 2024 |
+
"loss": 2.2128,
|
| 2025 |
+
"step": 144000
|
| 2026 |
+
},
|
| 2027 |
+
{
|
| 2028 |
+
"epoch": 0.08,
|
| 2029 |
+
"grad_norm": 4.242053985595703,
|
| 2030 |
+
"learning_rate": 9.975034640596274e-05,
|
| 2031 |
+
"loss": 2.191,
|
| 2032 |
+
"step": 144500
|
| 2033 |
+
},
|
| 2034 |
+
{
|
| 2035 |
+
"epoch": 0.08,
|
| 2036 |
+
"grad_norm": 4.181004047393799,
|
| 2037 |
+
"learning_rate": 9.974941832568751e-05,
|
| 2038 |
+
"loss": 2.1965,
|
| 2039 |
+
"step": 145000
|
| 2040 |
+
},
|
| 2041 |
+
{
|
| 2042 |
+
"epoch": 0.08,
|
| 2043 |
+
"grad_norm": 3.5158121585845947,
|
| 2044 |
+
"learning_rate": 9.974849024541228e-05,
|
| 2045 |
+
"loss": 2.1692,
|
| 2046 |
+
"step": 145500
|
| 2047 |
+
},
|
| 2048 |
+
{
|
| 2049 |
+
"epoch": 0.08,
|
| 2050 |
+
"grad_norm": 4.8536577224731445,
|
| 2051 |
+
"learning_rate": 9.974756216513704e-05,
|
| 2052 |
+
"loss": 2.1861,
|
| 2053 |
+
"step": 146000
|
| 2054 |
+
},
|
| 2055 |
+
{
|
| 2056 |
+
"epoch": 0.08,
|
| 2057 |
+
"grad_norm": 5.284401893615723,
|
| 2058 |
+
"learning_rate": 9.974663408486181e-05,
|
| 2059 |
+
"loss": 2.1801,
|
| 2060 |
+
"step": 146500
|
| 2061 |
+
},
|
| 2062 |
+
{
|
| 2063 |
+
"epoch": 0.08,
|
| 2064 |
+
"grad_norm": 4.611184597015381,
|
| 2065 |
+
"learning_rate": 9.974570600458658e-05,
|
| 2066 |
+
"loss": 2.1879,
|
| 2067 |
+
"step": 147000
|
| 2068 |
+
},
|
| 2069 |
+
{
|
| 2070 |
+
"epoch": 0.08,
|
| 2071 |
+
"grad_norm": 4.935207843780518,
|
| 2072 |
+
"learning_rate": 9.974477792431135e-05,
|
| 2073 |
+
"loss": 2.1608,
|
| 2074 |
+
"step": 147500
|
| 2075 |
+
},
|
| 2076 |
+
{
|
| 2077 |
+
"epoch": 0.08,
|
| 2078 |
+
"grad_norm": 4.852113246917725,
|
| 2079 |
+
"learning_rate": 9.974384984403612e-05,
|
| 2080 |
+
"loss": 2.1788,
|
| 2081 |
+
"step": 148000
|
| 2082 |
+
},
|
| 2083 |
+
{
|
| 2084 |
+
"epoch": 0.08,
|
| 2085 |
+
"grad_norm": 5.22275972366333,
|
| 2086 |
+
"learning_rate": 9.974292176376088e-05,
|
| 2087 |
+
"loss": 2.1585,
|
| 2088 |
+
"step": 148500
|
| 2089 |
+
},
|
| 2090 |
+
{
|
| 2091 |
+
"epoch": 0.08,
|
| 2092 |
+
"grad_norm": 5.024623394012451,
|
| 2093 |
+
"learning_rate": 9.974199368348564e-05,
|
| 2094 |
+
"loss": 2.1757,
|
| 2095 |
+
"step": 149000
|
| 2096 |
+
},
|
| 2097 |
+
{
|
| 2098 |
+
"epoch": 0.08,
|
| 2099 |
+
"grad_norm": 5.230965614318848,
|
| 2100 |
+
"learning_rate": 9.974106560321041e-05,
|
| 2101 |
+
"loss": 2.1685,
|
| 2102 |
+
"step": 149500
|
| 2103 |
+
},
|
| 2104 |
+
{
|
| 2105 |
+
"epoch": 0.08,
|
| 2106 |
+
"grad_norm": 4.0713090896606445,
|
| 2107 |
+
"learning_rate": 9.974013752293518e-05,
|
| 2108 |
+
"loss": 2.1586,
|
| 2109 |
+
"step": 150000
|
| 2110 |
+
},
|
| 2111 |
+
{
|
| 2112 |
+
"epoch": 0.08,
|
| 2113 |
+
"grad_norm": 5.00492000579834,
|
| 2114 |
+
"learning_rate": 9.973920944265995e-05,
|
| 2115 |
+
"loss": 2.1538,
|
| 2116 |
+
"step": 150500
|
| 2117 |
+
},
|
| 2118 |
+
{
|
| 2119 |
+
"epoch": 0.08,
|
| 2120 |
+
"grad_norm": 3.9473533630371094,
|
| 2121 |
+
"learning_rate": 9.973828136238473e-05,
|
| 2122 |
+
"loss": 2.1585,
|
| 2123 |
+
"step": 151000
|
| 2124 |
+
},
|
| 2125 |
+
{
|
| 2126 |
+
"epoch": 0.08,
|
| 2127 |
+
"grad_norm": 4.452467918395996,
|
| 2128 |
+
"learning_rate": 9.97373532821095e-05,
|
| 2129 |
+
"loss": 2.1579,
|
| 2130 |
+
"step": 151500
|
| 2131 |
+
},
|
| 2132 |
+
{
|
| 2133 |
+
"epoch": 0.08,
|
| 2134 |
+
"grad_norm": 4.7615647315979,
|
| 2135 |
+
"learning_rate": 9.973642520183425e-05,
|
| 2136 |
+
"loss": 2.1609,
|
| 2137 |
+
"step": 152000
|
| 2138 |
+
},
|
| 2139 |
+
{
|
| 2140 |
+
"epoch": 0.08,
|
| 2141 |
+
"grad_norm": 4.646842956542969,
|
| 2142 |
+
"learning_rate": 9.973549712155903e-05,
|
| 2143 |
+
"loss": 2.1762,
|
| 2144 |
+
"step": 152500
|
| 2145 |
+
},
|
| 2146 |
+
{
|
| 2147 |
+
"epoch": 0.09,
|
| 2148 |
+
"grad_norm": 5.9310526847839355,
|
| 2149 |
+
"learning_rate": 9.97345690412838e-05,
|
| 2150 |
+
"loss": 2.1578,
|
| 2151 |
+
"step": 153000
|
| 2152 |
+
},
|
| 2153 |
+
{
|
| 2154 |
+
"epoch": 0.09,
|
| 2155 |
+
"grad_norm": 3.432331085205078,
|
| 2156 |
+
"learning_rate": 9.973364096100857e-05,
|
| 2157 |
+
"loss": 2.1606,
|
| 2158 |
+
"step": 153500
|
| 2159 |
+
},
|
| 2160 |
+
{
|
| 2161 |
+
"epoch": 0.09,
|
| 2162 |
+
"grad_norm": 3.7542684078216553,
|
| 2163 |
+
"learning_rate": 9.973271288073334e-05,
|
| 2164 |
+
"loss": 2.1403,
|
| 2165 |
+
"step": 154000
|
| 2166 |
+
},
|
| 2167 |
+
{
|
| 2168 |
+
"epoch": 0.09,
|
| 2169 |
+
"grad_norm": 5.2377400398254395,
|
| 2170 |
+
"learning_rate": 9.973178480045811e-05,
|
| 2171 |
+
"loss": 2.154,
|
| 2172 |
+
"step": 154500
|
| 2173 |
+
},
|
| 2174 |
+
{
|
| 2175 |
+
"epoch": 0.09,
|
| 2176 |
+
"grad_norm": 4.928728103637695,
|
| 2177 |
+
"learning_rate": 9.973085672018287e-05,
|
| 2178 |
+
"loss": 2.1478,
|
| 2179 |
+
"step": 155000
|
| 2180 |
+
},
|
| 2181 |
+
{
|
| 2182 |
+
"epoch": 0.09,
|
| 2183 |
+
"grad_norm": 7.469805717468262,
|
| 2184 |
+
"learning_rate": 9.972992863990764e-05,
|
| 2185 |
+
"loss": 2.1443,
|
| 2186 |
+
"step": 155500
|
| 2187 |
+
},
|
| 2188 |
+
{
|
| 2189 |
+
"epoch": 0.09,
|
| 2190 |
+
"grad_norm": 5.656938076019287,
|
| 2191 |
+
"learning_rate": 9.972900055963241e-05,
|
| 2192 |
+
"loss": 2.1472,
|
| 2193 |
+
"step": 156000
|
| 2194 |
+
},
|
| 2195 |
+
{
|
| 2196 |
+
"epoch": 0.09,
|
| 2197 |
+
"grad_norm": 6.054020404815674,
|
| 2198 |
+
"learning_rate": 9.972807247935718e-05,
|
| 2199 |
+
"loss": 2.1766,
|
| 2200 |
+
"step": 156500
|
| 2201 |
+
},
|
| 2202 |
+
{
|
| 2203 |
+
"epoch": 0.09,
|
| 2204 |
+
"grad_norm": 4.4340925216674805,
|
| 2205 |
+
"learning_rate": 9.972714439908195e-05,
|
| 2206 |
+
"loss": 2.1198,
|
| 2207 |
+
"step": 157000
|
| 2208 |
+
},
|
| 2209 |
+
{
|
| 2210 |
+
"epoch": 0.09,
|
| 2211 |
+
"grad_norm": 5.291192531585693,
|
| 2212 |
+
"learning_rate": 9.972621631880673e-05,
|
| 2213 |
+
"loss": 2.1446,
|
| 2214 |
+
"step": 157500
|
| 2215 |
+
},
|
| 2216 |
+
{
|
| 2217 |
+
"epoch": 0.09,
|
| 2218 |
+
"grad_norm": 5.266712188720703,
|
| 2219 |
+
"learning_rate": 9.972528823853148e-05,
|
| 2220 |
+
"loss": 2.1402,
|
| 2221 |
+
"step": 158000
|
| 2222 |
+
},
|
| 2223 |
+
{
|
| 2224 |
+
"epoch": 0.09,
|
| 2225 |
+
"grad_norm": 7.1873250007629395,
|
| 2226 |
+
"learning_rate": 9.972436015825626e-05,
|
| 2227 |
+
"loss": 2.1163,
|
| 2228 |
+
"step": 158500
|
| 2229 |
+
},
|
| 2230 |
+
{
|
| 2231 |
+
"epoch": 0.09,
|
| 2232 |
+
"grad_norm": 8.807872772216797,
|
| 2233 |
+
"learning_rate": 9.972343207798103e-05,
|
| 2234 |
+
"loss": 2.087,
|
| 2235 |
+
"step": 159000
|
| 2236 |
+
},
|
| 2237 |
+
{
|
| 2238 |
+
"epoch": 0.09,
|
| 2239 |
+
"grad_norm": 4.034912109375,
|
| 2240 |
+
"learning_rate": 9.972250399770578e-05,
|
| 2241 |
+
"loss": 2.1372,
|
| 2242 |
+
"step": 159500
|
| 2243 |
+
},
|
| 2244 |
+
{
|
| 2245 |
+
"epoch": 0.09,
|
| 2246 |
+
"grad_norm": 4.9043755531311035,
|
| 2247 |
+
"learning_rate": 9.972157591743056e-05,
|
| 2248 |
+
"loss": 2.1184,
|
| 2249 |
+
"step": 160000
|
| 2250 |
+
},
|
| 2251 |
+
{
|
| 2252 |
+
"epoch": 0.09,
|
| 2253 |
+
"grad_norm": 4.8470458984375,
|
| 2254 |
+
"learning_rate": 9.972064783715533e-05,
|
| 2255 |
+
"loss": 2.1304,
|
| 2256 |
+
"step": 160500
|
| 2257 |
+
},
|
| 2258 |
+
{
|
| 2259 |
+
"epoch": 0.09,
|
| 2260 |
+
"grad_norm": 4.68010139465332,
|
| 2261 |
+
"learning_rate": 9.97197197568801e-05,
|
| 2262 |
+
"loss": 2.1093,
|
| 2263 |
+
"step": 161000
|
| 2264 |
+
},
|
| 2265 |
+
{
|
| 2266 |
+
"epoch": 0.09,
|
| 2267 |
+
"grad_norm": 4.731649398803711,
|
| 2268 |
+
"learning_rate": 9.971879167660486e-05,
|
| 2269 |
+
"loss": 2.1178,
|
| 2270 |
+
"step": 161500
|
| 2271 |
+
},
|
| 2272 |
+
{
|
| 2273 |
+
"epoch": 0.09,
|
| 2274 |
+
"grad_norm": 4.586087226867676,
|
| 2275 |
+
"learning_rate": 9.971786359632963e-05,
|
| 2276 |
+
"loss": 2.1154,
|
| 2277 |
+
"step": 162000
|
| 2278 |
+
},
|
| 2279 |
+
{
|
| 2280 |
+
"epoch": 0.09,
|
| 2281 |
+
"grad_norm": 4.602361679077148,
|
| 2282 |
+
"learning_rate": 9.97169355160544e-05,
|
| 2283 |
+
"loss": 2.0925,
|
| 2284 |
+
"step": 162500
|
| 2285 |
+
},
|
| 2286 |
+
{
|
| 2287 |
+
"epoch": 0.09,
|
| 2288 |
+
"grad_norm": 5.405307292938232,
|
| 2289 |
+
"learning_rate": 9.971600743577917e-05,
|
| 2290 |
+
"loss": 2.1131,
|
| 2291 |
+
"step": 163000
|
| 2292 |
+
},
|
| 2293 |
+
{
|
| 2294 |
+
"epoch": 0.09,
|
| 2295 |
+
"grad_norm": 5.402783393859863,
|
| 2296 |
+
"learning_rate": 9.971507935550394e-05,
|
| 2297 |
+
"loss": 2.1112,
|
| 2298 |
+
"step": 163500
|
| 2299 |
+
},
|
| 2300 |
+
{
|
| 2301 |
+
"epoch": 0.09,
|
| 2302 |
+
"grad_norm": 3.7548913955688477,
|
| 2303 |
+
"learning_rate": 9.97141512752287e-05,
|
| 2304 |
+
"loss": 2.0885,
|
| 2305 |
+
"step": 164000
|
| 2306 |
+
},
|
| 2307 |
+
{
|
| 2308 |
+
"epoch": 0.09,
|
| 2309 |
+
"grad_norm": 5.215375900268555,
|
| 2310 |
+
"learning_rate": 9.971322319495347e-05,
|
| 2311 |
+
"loss": 2.1016,
|
| 2312 |
+
"step": 164500
|
| 2313 |
+
},
|
| 2314 |
+
{
|
| 2315 |
+
"epoch": 0.09,
|
| 2316 |
+
"grad_norm": 5.195767402648926,
|
| 2317 |
+
"learning_rate": 9.971229511467824e-05,
|
| 2318 |
+
"loss": 2.0821,
|
| 2319 |
+
"step": 165000
|
| 2320 |
+
},
|
| 2321 |
+
{
|
| 2322 |
+
"epoch": 0.09,
|
| 2323 |
+
"grad_norm": 5.254913806915283,
|
| 2324 |
+
"learning_rate": 9.971136703440301e-05,
|
| 2325 |
+
"loss": 2.0876,
|
| 2326 |
+
"step": 165500
|
| 2327 |
+
},
|
| 2328 |
+
{
|
| 2329 |
+
"epoch": 0.09,
|
| 2330 |
+
"grad_norm": 4.161681652069092,
|
| 2331 |
+
"learning_rate": 9.971043895412779e-05,
|
| 2332 |
+
"loss": 2.0771,
|
| 2333 |
+
"step": 166000
|
| 2334 |
+
},
|
| 2335 |
+
{
|
| 2336 |
+
"epoch": 0.09,
|
| 2337 |
+
"grad_norm": 7.586195468902588,
|
| 2338 |
+
"learning_rate": 9.970951087385256e-05,
|
| 2339 |
+
"loss": 2.0893,
|
| 2340 |
+
"step": 166500
|
| 2341 |
+
},
|
| 2342 |
+
{
|
| 2343 |
+
"epoch": 0.09,
|
| 2344 |
+
"grad_norm": 4.742598533630371,
|
| 2345 |
+
"learning_rate": 9.970858279357731e-05,
|
| 2346 |
+
"loss": 2.1043,
|
| 2347 |
+
"step": 167000
|
| 2348 |
+
},
|
| 2349 |
+
{
|
| 2350 |
+
"epoch": 0.09,
|
| 2351 |
+
"grad_norm": 7.583818435668945,
|
| 2352 |
+
"learning_rate": 9.970765471330209e-05,
|
| 2353 |
+
"loss": 2.0643,
|
| 2354 |
+
"step": 167500
|
| 2355 |
+
},
|
| 2356 |
+
{
|
| 2357 |
+
"epoch": 0.09,
|
| 2358 |
+
"grad_norm": 15.07123851776123,
|
| 2359 |
+
"learning_rate": 9.970672663302686e-05,
|
| 2360 |
+
"loss": 2.0719,
|
| 2361 |
+
"step": 168000
|
| 2362 |
+
},
|
| 2363 |
+
{
|
| 2364 |
+
"epoch": 0.09,
|
| 2365 |
+
"grad_norm": 4.9100565910339355,
|
| 2366 |
+
"learning_rate": 9.970579855275163e-05,
|
| 2367 |
+
"loss": 2.0645,
|
| 2368 |
+
"step": 168500
|
| 2369 |
+
},
|
| 2370 |
+
{
|
| 2371 |
+
"epoch": 0.09,
|
| 2372 |
+
"grad_norm": 4.305502891540527,
|
| 2373 |
+
"learning_rate": 9.97048704724764e-05,
|
| 2374 |
+
"loss": 2.0661,
|
| 2375 |
+
"step": 169000
|
| 2376 |
+
},
|
| 2377 |
+
{
|
| 2378 |
+
"epoch": 0.09,
|
| 2379 |
+
"grad_norm": 4.631031513214111,
|
| 2380 |
+
"learning_rate": 9.970394239220116e-05,
|
| 2381 |
+
"loss": 2.0729,
|
| 2382 |
+
"step": 169500
|
| 2383 |
+
},
|
| 2384 |
+
{
|
| 2385 |
+
"epoch": 0.09,
|
| 2386 |
+
"grad_norm": 4.102370738983154,
|
| 2387 |
+
"learning_rate": 9.970301431192593e-05,
|
| 2388 |
+
"loss": 2.0621,
|
| 2389 |
+
"step": 170000
|
| 2390 |
+
},
|
| 2391 |
+
{
|
| 2392 |
+
"epoch": 0.09,
|
| 2393 |
+
"grad_norm": 9.088913917541504,
|
| 2394 |
+
"learning_rate": 9.970208623165069e-05,
|
| 2395 |
+
"loss": 2.0566,
|
| 2396 |
+
"step": 170500
|
| 2397 |
+
},
|
| 2398 |
+
{
|
| 2399 |
+
"epoch": 0.1,
|
| 2400 |
+
"grad_norm": 3.1607444286346436,
|
| 2401 |
+
"learning_rate": 9.970115815137546e-05,
|
| 2402 |
+
"loss": 2.053,
|
| 2403 |
+
"step": 171000
|
| 2404 |
+
},
|
| 2405 |
+
{
|
| 2406 |
+
"epoch": 0.1,
|
| 2407 |
+
"grad_norm": 5.192753791809082,
|
| 2408 |
+
"learning_rate": 9.970023007110023e-05,
|
| 2409 |
+
"loss": 2.0443,
|
| 2410 |
+
"step": 171500
|
| 2411 |
+
},
|
| 2412 |
+
{
|
| 2413 |
+
"epoch": 0.1,
|
| 2414 |
+
"grad_norm": 5.527139186859131,
|
| 2415 |
+
"learning_rate": 9.9699301990825e-05,
|
| 2416 |
+
"loss": 2.0356,
|
| 2417 |
+
"step": 172000
|
| 2418 |
+
},
|
| 2419 |
+
{
|
| 2420 |
+
"epoch": 0.1,
|
| 2421 |
+
"grad_norm": 4.856675148010254,
|
| 2422 |
+
"learning_rate": 9.969837391054977e-05,
|
| 2423 |
+
"loss": 2.0405,
|
| 2424 |
+
"step": 172500
|
| 2425 |
+
},
|
| 2426 |
+
{
|
| 2427 |
+
"epoch": 0.1,
|
| 2428 |
+
"grad_norm": 6.9087724685668945,
|
| 2429 |
+
"learning_rate": 9.969744583027454e-05,
|
| 2430 |
+
"loss": 2.0544,
|
| 2431 |
+
"step": 173000
|
| 2432 |
+
},
|
| 2433 |
+
{
|
| 2434 |
+
"epoch": 0.1,
|
| 2435 |
+
"grad_norm": 5.108808517456055,
|
| 2436 |
+
"learning_rate": 9.96965177499993e-05,
|
| 2437 |
+
"loss": 2.0277,
|
| 2438 |
+
"step": 173500
|
| 2439 |
+
},
|
| 2440 |
+
{
|
| 2441 |
+
"epoch": 0.1,
|
| 2442 |
+
"grad_norm": 3.581963300704956,
|
| 2443 |
+
"learning_rate": 9.969558966972407e-05,
|
| 2444 |
+
"loss": 2.0569,
|
| 2445 |
+
"step": 174000
|
| 2446 |
+
},
|
| 2447 |
+
{
|
| 2448 |
+
"epoch": 0.1,
|
| 2449 |
+
"grad_norm": 2.8868958950042725,
|
| 2450 |
+
"learning_rate": 9.969466158944885e-05,
|
| 2451 |
+
"loss": 2.0569,
|
| 2452 |
+
"step": 174500
|
| 2453 |
+
},
|
| 2454 |
+
{
|
| 2455 |
+
"epoch": 0.1,
|
| 2456 |
+
"grad_norm": 4.1303510665893555,
|
| 2457 |
+
"learning_rate": 9.969373350917362e-05,
|
| 2458 |
+
"loss": 2.0378,
|
| 2459 |
+
"step": 175000
|
| 2460 |
+
},
|
| 2461 |
+
{
|
| 2462 |
+
"epoch": 0.1,
|
| 2463 |
+
"grad_norm": 3.9498367309570312,
|
| 2464 |
+
"learning_rate": 9.969280542889839e-05,
|
| 2465 |
+
"loss": 2.0487,
|
| 2466 |
+
"step": 175500
|
| 2467 |
+
},
|
| 2468 |
+
{
|
| 2469 |
+
"epoch": 0.1,
|
| 2470 |
+
"grad_norm": 4.58161735534668,
|
| 2471 |
+
"learning_rate": 9.969187734862316e-05,
|
| 2472 |
+
"loss": 2.0113,
|
| 2473 |
+
"step": 176000
|
| 2474 |
+
},
|
| 2475 |
+
{
|
| 2476 |
+
"epoch": 0.1,
|
| 2477 |
+
"grad_norm": 4.350900650024414,
|
| 2478 |
+
"learning_rate": 9.969094926834792e-05,
|
| 2479 |
+
"loss": 2.0371,
|
| 2480 |
+
"step": 176500
|
| 2481 |
+
},
|
| 2482 |
+
{
|
| 2483 |
+
"epoch": 0.1,
|
| 2484 |
+
"grad_norm": 4.318021297454834,
|
| 2485 |
+
"learning_rate": 9.969002118807269e-05,
|
| 2486 |
+
"loss": 2.0379,
|
| 2487 |
+
"step": 177000
|
| 2488 |
+
},
|
| 2489 |
+
{
|
| 2490 |
+
"epoch": 0.1,
|
| 2491 |
+
"grad_norm": 5.780760288238525,
|
| 2492 |
+
"learning_rate": 9.968909310779746e-05,
|
| 2493 |
+
"loss": 2.0117,
|
| 2494 |
+
"step": 177500
|
| 2495 |
+
},
|
| 2496 |
+
{
|
| 2497 |
+
"epoch": 0.1,
|
| 2498 |
+
"grad_norm": 4.759991645812988,
|
| 2499 |
+
"learning_rate": 9.968816502752223e-05,
|
| 2500 |
+
"loss": 2.034,
|
| 2501 |
+
"step": 178000
|
| 2502 |
+
},
|
| 2503 |
+
{
|
| 2504 |
+
"epoch": 0.1,
|
| 2505 |
+
"grad_norm": 3.5731894969940186,
|
| 2506 |
+
"learning_rate": 9.9687236947247e-05,
|
| 2507 |
+
"loss": 2.0386,
|
| 2508 |
+
"step": 178500
|
| 2509 |
+
},
|
| 2510 |
+
{
|
| 2511 |
+
"epoch": 0.1,
|
| 2512 |
+
"grad_norm": 5.845817565917969,
|
| 2513 |
+
"learning_rate": 9.968630886697176e-05,
|
| 2514 |
+
"loss": 2.0078,
|
| 2515 |
+
"step": 179000
|
| 2516 |
+
},
|
| 2517 |
+
{
|
| 2518 |
+
"epoch": 0.1,
|
| 2519 |
+
"grad_norm": 4.691517353057861,
|
| 2520 |
+
"learning_rate": 9.968538078669653e-05,
|
| 2521 |
+
"loss": 2.0223,
|
| 2522 |
+
"step": 179500
|
| 2523 |
+
},
|
| 2524 |
+
{
|
| 2525 |
+
"epoch": 0.1,
|
| 2526 |
+
"grad_norm": 3.678426742553711,
|
| 2527 |
+
"learning_rate": 9.968445270642129e-05,
|
| 2528 |
+
"loss": 2.0235,
|
| 2529 |
+
"step": 180000
|
| 2530 |
+
},
|
| 2531 |
+
{
|
| 2532 |
+
"epoch": 0.1,
|
| 2533 |
+
"grad_norm": 4.184782981872559,
|
| 2534 |
+
"learning_rate": 9.968352462614606e-05,
|
| 2535 |
+
"loss": 2.0036,
|
| 2536 |
+
"step": 180500
|
| 2537 |
+
},
|
| 2538 |
+
{
|
| 2539 |
+
"epoch": 0.1,
|
| 2540 |
+
"grad_norm": 5.912731647491455,
|
| 2541 |
+
"learning_rate": 9.968259654587083e-05,
|
| 2542 |
+
"loss": 2.016,
|
| 2543 |
+
"step": 181000
|
| 2544 |
+
},
|
| 2545 |
+
{
|
| 2546 |
+
"epoch": 0.1,
|
| 2547 |
+
"grad_norm": 5.004282474517822,
|
| 2548 |
+
"learning_rate": 9.96816684655956e-05,
|
| 2549 |
+
"loss": 1.9935,
|
| 2550 |
+
"step": 181500
|
| 2551 |
+
},
|
| 2552 |
+
{
|
| 2553 |
+
"epoch": 0.1,
|
| 2554 |
+
"grad_norm": 4.136642932891846,
|
| 2555 |
+
"learning_rate": 9.968074038532038e-05,
|
| 2556 |
+
"loss": 2.0201,
|
| 2557 |
+
"step": 182000
|
| 2558 |
+
},
|
| 2559 |
+
{
|
| 2560 |
+
"epoch": 0.1,
|
| 2561 |
+
"grad_norm": 4.304795742034912,
|
| 2562 |
+
"learning_rate": 9.967981230504513e-05,
|
| 2563 |
+
"loss": 2.0064,
|
| 2564 |
+
"step": 182500
|
| 2565 |
+
},
|
| 2566 |
+
{
|
| 2567 |
+
"epoch": 0.1,
|
| 2568 |
+
"grad_norm": 4.177530765533447,
|
| 2569 |
+
"learning_rate": 9.96788842247699e-05,
|
| 2570 |
+
"loss": 2.0087,
|
| 2571 |
+
"step": 183000
|
| 2572 |
+
},
|
| 2573 |
+
{
|
| 2574 |
+
"epoch": 0.1,
|
| 2575 |
+
"grad_norm": 3.730872869491577,
|
| 2576 |
+
"learning_rate": 9.967795614449468e-05,
|
| 2577 |
+
"loss": 2.0008,
|
| 2578 |
+
"step": 183500
|
| 2579 |
+
},
|
| 2580 |
+
{
|
| 2581 |
+
"epoch": 0.1,
|
| 2582 |
+
"grad_norm": 4.738286972045898,
|
| 2583 |
+
"learning_rate": 9.967702806421945e-05,
|
| 2584 |
+
"loss": 1.9961,
|
| 2585 |
+
"step": 184000
|
| 2586 |
+
},
|
| 2587 |
+
{
|
| 2588 |
+
"epoch": 0.1,
|
| 2589 |
+
"grad_norm": 4.57652473449707,
|
| 2590 |
+
"learning_rate": 9.967609998394422e-05,
|
| 2591 |
+
"loss": 1.9861,
|
| 2592 |
+
"step": 184500
|
| 2593 |
+
},
|
| 2594 |
+
{
|
| 2595 |
+
"epoch": 0.1,
|
| 2596 |
+
"grad_norm": 3.8996493816375732,
|
| 2597 |
+
"learning_rate": 9.967517190366899e-05,
|
| 2598 |
+
"loss": 1.9821,
|
| 2599 |
+
"step": 185000
|
| 2600 |
+
},
|
| 2601 |
+
{
|
| 2602 |
+
"epoch": 0.1,
|
| 2603 |
+
"grad_norm": 5.269779205322266,
|
| 2604 |
+
"learning_rate": 9.967424382339375e-05,
|
| 2605 |
+
"loss": 1.9761,
|
| 2606 |
+
"step": 185500
|
| 2607 |
+
},
|
| 2608 |
+
{
|
| 2609 |
+
"epoch": 0.1,
|
| 2610 |
+
"grad_norm": 3.3818140029907227,
|
| 2611 |
+
"learning_rate": 9.967331574311852e-05,
|
| 2612 |
+
"loss": 1.9679,
|
| 2613 |
+
"step": 186000
|
| 2614 |
+
},
|
| 2615 |
+
{
|
| 2616 |
+
"epoch": 0.1,
|
| 2617 |
+
"grad_norm": 3.7173426151275635,
|
| 2618 |
+
"learning_rate": 9.967238766284329e-05,
|
| 2619 |
+
"loss": 1.9992,
|
| 2620 |
+
"step": 186500
|
| 2621 |
+
},
|
| 2622 |
+
{
|
| 2623 |
+
"epoch": 0.1,
|
| 2624 |
+
"grad_norm": 3.1090087890625,
|
| 2625 |
+
"learning_rate": 9.967145958256806e-05,
|
| 2626 |
+
"loss": 1.9782,
|
| 2627 |
+
"step": 187000
|
| 2628 |
+
},
|
| 2629 |
+
{
|
| 2630 |
+
"epoch": 0.1,
|
| 2631 |
+
"grad_norm": 4.101827621459961,
|
| 2632 |
+
"learning_rate": 9.967053150229283e-05,
|
| 2633 |
+
"loss": 1.9704,
|
| 2634 |
+
"step": 187500
|
| 2635 |
+
},
|
| 2636 |
+
{
|
| 2637 |
+
"epoch": 0.1,
|
| 2638 |
+
"grad_norm": 3.682985782623291,
|
| 2639 |
+
"learning_rate": 9.96696034220176e-05,
|
| 2640 |
+
"loss": 2.0032,
|
| 2641 |
+
"step": 188000
|
| 2642 |
+
},
|
| 2643 |
+
{
|
| 2644 |
+
"epoch": 0.1,
|
| 2645 |
+
"grad_norm": 19.02726173400879,
|
| 2646 |
+
"learning_rate": 9.966867534174236e-05,
|
| 2647 |
+
"loss": 1.9946,
|
| 2648 |
+
"step": 188500
|
| 2649 |
+
},
|
| 2650 |
+
{
|
| 2651 |
+
"epoch": 0.11,
|
| 2652 |
+
"grad_norm": 4.69189453125,
|
| 2653 |
+
"learning_rate": 9.966774726146713e-05,
|
| 2654 |
+
"loss": 1.9245,
|
| 2655 |
+
"step": 189000
|
| 2656 |
+
},
|
| 2657 |
+
{
|
| 2658 |
+
"epoch": 0.11,
|
| 2659 |
+
"grad_norm": 12.00914192199707,
|
| 2660 |
+
"learning_rate": 9.96668191811919e-05,
|
| 2661 |
+
"loss": 1.9725,
|
| 2662 |
+
"step": 189500
|
| 2663 |
+
},
|
| 2664 |
+
{
|
| 2665 |
+
"epoch": 0.11,
|
| 2666 |
+
"grad_norm": 6.565895080566406,
|
| 2667 |
+
"learning_rate": 9.966589110091668e-05,
|
| 2668 |
+
"loss": 1.982,
|
| 2669 |
+
"step": 190000
|
| 2670 |
+
},
|
| 2671 |
+
{
|
| 2672 |
+
"epoch": 0.11,
|
| 2673 |
+
"grad_norm": 4.286581993103027,
|
| 2674 |
+
"learning_rate": 9.966496302064143e-05,
|
| 2675 |
+
"loss": 1.9978,
|
| 2676 |
+
"step": 190500
|
| 2677 |
+
},
|
| 2678 |
+
{
|
| 2679 |
+
"epoch": 0.11,
|
| 2680 |
+
"grad_norm": 5.0877509117126465,
|
| 2681 |
+
"learning_rate": 9.96640349403662e-05,
|
| 2682 |
+
"loss": 1.9735,
|
| 2683 |
+
"step": 191000
|
| 2684 |
+
},
|
| 2685 |
+
{
|
| 2686 |
+
"epoch": 0.11,
|
| 2687 |
+
"grad_norm": 4.9631242752075195,
|
| 2688 |
+
"learning_rate": 9.966310686009098e-05,
|
| 2689 |
+
"loss": 1.9508,
|
| 2690 |
+
"step": 191500
|
| 2691 |
+
},
|
| 2692 |
+
{
|
| 2693 |
+
"epoch": 0.11,
|
| 2694 |
+
"grad_norm": 4.4933552742004395,
|
| 2695 |
+
"learning_rate": 9.966217877981574e-05,
|
| 2696 |
+
"loss": 1.962,
|
| 2697 |
+
"step": 192000
|
| 2698 |
+
},
|
| 2699 |
+
{
|
| 2700 |
+
"epoch": 0.11,
|
| 2701 |
+
"grad_norm": 4.3112874031066895,
|
| 2702 |
+
"learning_rate": 9.966125069954051e-05,
|
| 2703 |
+
"loss": 1.9566,
|
| 2704 |
+
"step": 192500
|
| 2705 |
+
},
|
| 2706 |
+
{
|
| 2707 |
+
"epoch": 0.11,
|
| 2708 |
+
"grad_norm": 4.248539924621582,
|
| 2709 |
+
"learning_rate": 9.966032261926528e-05,
|
| 2710 |
+
"loss": 1.9379,
|
| 2711 |
+
"step": 193000
|
| 2712 |
+
},
|
| 2713 |
+
{
|
| 2714 |
+
"epoch": 0.11,
|
| 2715 |
+
"grad_norm": 3.26794171333313,
|
| 2716 |
+
"learning_rate": 9.965939453899005e-05,
|
| 2717 |
+
"loss": 1.9297,
|
| 2718 |
+
"step": 193500
|
| 2719 |
+
},
|
| 2720 |
+
{
|
| 2721 |
+
"epoch": 0.11,
|
| 2722 |
+
"grad_norm": 3.2197368144989014,
|
| 2723 |
+
"learning_rate": 9.965846645871482e-05,
|
| 2724 |
+
"loss": 1.9494,
|
| 2725 |
+
"step": 194000
|
| 2726 |
+
},
|
| 2727 |
+
{
|
| 2728 |
+
"epoch": 0.11,
|
| 2729 |
+
"grad_norm": 4.919404983520508,
|
| 2730 |
+
"learning_rate": 9.965753837843958e-05,
|
| 2731 |
+
"loss": 1.9529,
|
| 2732 |
+
"step": 194500
|
| 2733 |
+
},
|
| 2734 |
+
{
|
| 2735 |
+
"epoch": 0.11,
|
| 2736 |
+
"grad_norm": 4.171731948852539,
|
| 2737 |
+
"learning_rate": 9.965661029816435e-05,
|
| 2738 |
+
"loss": 1.9323,
|
| 2739 |
+
"step": 195000
|
| 2740 |
+
},
|
| 2741 |
+
{
|
| 2742 |
+
"epoch": 0.11,
|
| 2743 |
+
"grad_norm": 4.448825836181641,
|
| 2744 |
+
"learning_rate": 9.965568221788912e-05,
|
| 2745 |
+
"loss": 1.9484,
|
| 2746 |
+
"step": 195500
|
| 2747 |
+
},
|
| 2748 |
+
{
|
| 2749 |
+
"epoch": 0.11,
|
| 2750 |
+
"grad_norm": 5.62270975112915,
|
| 2751 |
+
"learning_rate": 9.965475413761389e-05,
|
| 2752 |
+
"loss": 1.9269,
|
| 2753 |
+
"step": 196000
|
| 2754 |
+
},
|
| 2755 |
+
{
|
| 2756 |
+
"epoch": 0.11,
|
| 2757 |
+
"grad_norm": 6.439905166625977,
|
| 2758 |
+
"learning_rate": 9.965382605733866e-05,
|
| 2759 |
+
"loss": 1.9193,
|
| 2760 |
+
"step": 196500
|
| 2761 |
+
},
|
| 2762 |
+
{
|
| 2763 |
+
"epoch": 0.11,
|
| 2764 |
+
"grad_norm": 4.133171081542969,
|
| 2765 |
+
"learning_rate": 9.965289797706344e-05,
|
| 2766 |
+
"loss": 1.9307,
|
| 2767 |
+
"step": 197000
|
| 2768 |
+
},
|
| 2769 |
+
{
|
| 2770 |
+
"epoch": 0.11,
|
| 2771 |
+
"grad_norm": 5.984546184539795,
|
| 2772 |
+
"learning_rate": 9.96519698967882e-05,
|
| 2773 |
+
"loss": 1.9042,
|
| 2774 |
+
"step": 197500
|
| 2775 |
+
},
|
| 2776 |
+
{
|
| 2777 |
+
"epoch": 0.11,
|
| 2778 |
+
"grad_norm": 4.515778064727783,
|
| 2779 |
+
"learning_rate": 9.965104181651297e-05,
|
| 2780 |
+
"loss": 1.9541,
|
| 2781 |
+
"step": 198000
|
| 2782 |
+
},
|
| 2783 |
+
{
|
| 2784 |
+
"epoch": 0.11,
|
| 2785 |
+
"grad_norm": 3.409749984741211,
|
| 2786 |
+
"learning_rate": 9.965011373623774e-05,
|
| 2787 |
+
"loss": 1.9203,
|
| 2788 |
+
"step": 198500
|
| 2789 |
+
},
|
| 2790 |
+
{
|
| 2791 |
+
"epoch": 0.11,
|
| 2792 |
+
"grad_norm": 3.724917411804199,
|
| 2793 |
+
"learning_rate": 9.964918565596251e-05,
|
| 2794 |
+
"loss": 1.9262,
|
| 2795 |
+
"step": 199000
|
| 2796 |
+
},
|
| 2797 |
+
{
|
| 2798 |
+
"epoch": 0.11,
|
| 2799 |
+
"grad_norm": 3.5863378047943115,
|
| 2800 |
+
"learning_rate": 9.964825757568728e-05,
|
| 2801 |
+
"loss": 1.9383,
|
| 2802 |
+
"step": 199500
|
| 2803 |
+
},
|
| 2804 |
+
{
|
| 2805 |
+
"epoch": 0.11,
|
| 2806 |
+
"grad_norm": 6.107095718383789,
|
| 2807 |
+
"learning_rate": 9.964732949541205e-05,
|
| 2808 |
+
"loss": 1.9263,
|
| 2809 |
+
"step": 200000
|
| 2810 |
+
},
|
| 2811 |
+
{
|
| 2812 |
+
"epoch": 0.11,
|
| 2813 |
+
"grad_norm": 3.91813588142395,
|
| 2814 |
+
"learning_rate": 9.964640141513681e-05,
|
| 2815 |
+
"loss": 1.9217,
|
| 2816 |
+
"step": 200500
|
| 2817 |
+
},
|
| 2818 |
+
{
|
| 2819 |
+
"epoch": 0.11,
|
| 2820 |
+
"grad_norm": 4.080438137054443,
|
| 2821 |
+
"learning_rate": 9.964547333486157e-05,
|
| 2822 |
+
"loss": 1.9089,
|
| 2823 |
+
"step": 201000
|
| 2824 |
+
},
|
| 2825 |
+
{
|
| 2826 |
+
"epoch": 0.11,
|
| 2827 |
+
"grad_norm": 3.550732374191284,
|
| 2828 |
+
"learning_rate": 9.964454525458634e-05,
|
| 2829 |
+
"loss": 1.9029,
|
| 2830 |
+
"step": 201500
|
| 2831 |
+
},
|
| 2832 |
+
{
|
| 2833 |
+
"epoch": 0.11,
|
| 2834 |
+
"grad_norm": 6.719958305358887,
|
| 2835 |
+
"learning_rate": 9.964361717431111e-05,
|
| 2836 |
+
"loss": 1.9201,
|
| 2837 |
+
"step": 202000
|
| 2838 |
+
},
|
| 2839 |
+
{
|
| 2840 |
+
"epoch": 0.11,
|
| 2841 |
+
"grad_norm": 4.790652751922607,
|
| 2842 |
+
"learning_rate": 9.964268909403588e-05,
|
| 2843 |
+
"loss": 1.8942,
|
| 2844 |
+
"step": 202500
|
| 2845 |
+
},
|
| 2846 |
+
{
|
| 2847 |
+
"epoch": 0.11,
|
| 2848 |
+
"grad_norm": 5.303153991699219,
|
| 2849 |
+
"learning_rate": 9.964176101376065e-05,
|
| 2850 |
+
"loss": 1.8945,
|
| 2851 |
+
"step": 203000
|
| 2852 |
+
},
|
| 2853 |
+
{
|
| 2854 |
+
"epoch": 0.11,
|
| 2855 |
+
"grad_norm": 6.912900924682617,
|
| 2856 |
+
"learning_rate": 9.964083293348542e-05,
|
| 2857 |
+
"loss": 1.9146,
|
| 2858 |
+
"step": 203500
|
| 2859 |
+
},
|
| 2860 |
+
{
|
| 2861 |
+
"epoch": 0.11,
|
| 2862 |
+
"grad_norm": 4.400740623474121,
|
| 2863 |
+
"learning_rate": 9.963990485321018e-05,
|
| 2864 |
+
"loss": 1.8796,
|
| 2865 |
+
"step": 204000
|
| 2866 |
+
},
|
| 2867 |
+
{
|
| 2868 |
+
"epoch": 0.11,
|
| 2869 |
+
"grad_norm": 5.205254554748535,
|
| 2870 |
+
"learning_rate": 9.963897677293495e-05,
|
| 2871 |
+
"loss": 1.8949,
|
| 2872 |
+
"step": 204500
|
| 2873 |
+
},
|
| 2874 |
+
{
|
| 2875 |
+
"epoch": 0.11,
|
| 2876 |
+
"grad_norm": 5.543479919433594,
|
| 2877 |
+
"learning_rate": 9.963804869265972e-05,
|
| 2878 |
+
"loss": 1.8846,
|
| 2879 |
+
"step": 205000
|
| 2880 |
+
},
|
| 2881 |
+
{
|
| 2882 |
+
"epoch": 0.11,
|
| 2883 |
+
"grad_norm": 8.72082233428955,
|
| 2884 |
+
"learning_rate": 9.96371206123845e-05,
|
| 2885 |
+
"loss": 1.8709,
|
| 2886 |
+
"step": 205500
|
| 2887 |
+
},
|
| 2888 |
+
{
|
| 2889 |
+
"epoch": 0.11,
|
| 2890 |
+
"grad_norm": 6.174784183502197,
|
| 2891 |
+
"learning_rate": 9.963619253210927e-05,
|
| 2892 |
+
"loss": 1.9181,
|
| 2893 |
+
"step": 206000
|
| 2894 |
+
},
|
| 2895 |
+
{
|
| 2896 |
+
"epoch": 0.11,
|
| 2897 |
+
"grad_norm": 5.813675880432129,
|
| 2898 |
+
"learning_rate": 9.963526445183404e-05,
|
| 2899 |
+
"loss": 1.8797,
|
| 2900 |
+
"step": 206500
|
| 2901 |
+
},
|
| 2902 |
+
{
|
| 2903 |
+
"epoch": 0.12,
|
| 2904 |
+
"grad_norm": 5.238151550292969,
|
| 2905 |
+
"learning_rate": 9.96343363715588e-05,
|
| 2906 |
+
"loss": 1.8747,
|
| 2907 |
+
"step": 207000
|
| 2908 |
+
},
|
| 2909 |
+
{
|
| 2910 |
+
"epoch": 0.12,
|
| 2911 |
+
"grad_norm": 3.6273601055145264,
|
| 2912 |
+
"learning_rate": 9.963340829128357e-05,
|
| 2913 |
+
"loss": 1.8942,
|
| 2914 |
+
"step": 207500
|
| 2915 |
+
},
|
| 2916 |
+
{
|
| 2917 |
+
"epoch": 0.12,
|
| 2918 |
+
"grad_norm": 6.024681091308594,
|
| 2919 |
+
"learning_rate": 9.963248021100834e-05,
|
| 2920 |
+
"loss": 1.8974,
|
| 2921 |
+
"step": 208000
|
| 2922 |
+
},
|
| 2923 |
+
{
|
| 2924 |
+
"epoch": 0.12,
|
| 2925 |
+
"grad_norm": 6.6017746925354,
|
| 2926 |
+
"learning_rate": 9.963155213073311e-05,
|
| 2927 |
+
"loss": 1.8878,
|
| 2928 |
+
"step": 208500
|
| 2929 |
+
},
|
| 2930 |
+
{
|
| 2931 |
+
"epoch": 0.12,
|
| 2932 |
+
"grad_norm": 4.99990177154541,
|
| 2933 |
+
"learning_rate": 9.963062405045788e-05,
|
| 2934 |
+
"loss": 1.8741,
|
| 2935 |
+
"step": 209000
|
| 2936 |
+
},
|
| 2937 |
+
{
|
| 2938 |
+
"epoch": 0.12,
|
| 2939 |
+
"grad_norm": 4.064300060272217,
|
| 2940 |
+
"learning_rate": 9.962969597018264e-05,
|
| 2941 |
+
"loss": 1.8638,
|
| 2942 |
+
"step": 209500
|
| 2943 |
+
},
|
| 2944 |
+
{
|
| 2945 |
+
"epoch": 0.12,
|
| 2946 |
+
"grad_norm": 4.6583685874938965,
|
| 2947 |
+
"learning_rate": 9.962876788990741e-05,
|
| 2948 |
+
"loss": 1.8723,
|
| 2949 |
+
"step": 210000
|
| 2950 |
+
},
|
| 2951 |
+
{
|
| 2952 |
+
"epoch": 0.12,
|
| 2953 |
+
"grad_norm": 4.108057498931885,
|
| 2954 |
+
"learning_rate": 9.962783980963218e-05,
|
| 2955 |
+
"loss": 1.8643,
|
| 2956 |
+
"step": 210500
|
| 2957 |
+
},
|
| 2958 |
+
{
|
| 2959 |
+
"epoch": 0.12,
|
| 2960 |
+
"grad_norm": 4.108830451965332,
|
| 2961 |
+
"learning_rate": 9.962691172935694e-05,
|
| 2962 |
+
"loss": 1.866,
|
| 2963 |
+
"step": 211000
|
| 2964 |
+
},
|
| 2965 |
+
{
|
| 2966 |
+
"epoch": 0.12,
|
| 2967 |
+
"grad_norm": 4.949222564697266,
|
| 2968 |
+
"learning_rate": 9.962598364908171e-05,
|
| 2969 |
+
"loss": 1.8723,
|
| 2970 |
+
"step": 211500
|
| 2971 |
+
},
|
| 2972 |
+
{
|
| 2973 |
+
"epoch": 0.12,
|
| 2974 |
+
"grad_norm": 3.929126501083374,
|
| 2975 |
+
"learning_rate": 9.962505556880648e-05,
|
| 2976 |
+
"loss": 1.85,
|
| 2977 |
+
"step": 212000
|
| 2978 |
+
},
|
| 2979 |
+
{
|
| 2980 |
+
"epoch": 0.12,
|
| 2981 |
+
"grad_norm": 4.211127281188965,
|
| 2982 |
+
"learning_rate": 9.962412748853125e-05,
|
| 2983 |
+
"loss": 1.8463,
|
| 2984 |
+
"step": 212500
|
| 2985 |
+
},
|
| 2986 |
+
{
|
| 2987 |
+
"epoch": 0.12,
|
| 2988 |
+
"grad_norm": 4.776076316833496,
|
| 2989 |
+
"learning_rate": 9.962319940825601e-05,
|
| 2990 |
+
"loss": 1.869,
|
| 2991 |
+
"step": 213000
|
| 2992 |
+
},
|
| 2993 |
+
{
|
| 2994 |
+
"epoch": 0.12,
|
| 2995 |
+
"grad_norm": 4.429445743560791,
|
| 2996 |
+
"learning_rate": 9.962227132798078e-05,
|
| 2997 |
+
"loss": 1.8667,
|
| 2998 |
+
"step": 213500
|
| 2999 |
+
},
|
| 3000 |
+
{
|
| 3001 |
+
"epoch": 0.12,
|
| 3002 |
+
"grad_norm": 3.9247748851776123,
|
| 3003 |
+
"learning_rate": 9.962134324770555e-05,
|
| 3004 |
+
"loss": 1.8632,
|
| 3005 |
+
"step": 214000
|
| 3006 |
+
},
|
| 3007 |
+
{
|
| 3008 |
+
"epoch": 0.12,
|
| 3009 |
+
"grad_norm": 4.082817554473877,
|
| 3010 |
+
"learning_rate": 9.962041516743033e-05,
|
| 3011 |
+
"loss": 1.8438,
|
| 3012 |
+
"step": 214500
|
| 3013 |
+
},
|
| 3014 |
+
{
|
| 3015 |
+
"epoch": 0.12,
|
| 3016 |
+
"grad_norm": 5.592115879058838,
|
| 3017 |
+
"learning_rate": 9.96194870871551e-05,
|
| 3018 |
+
"loss": 1.8395,
|
| 3019 |
+
"step": 215000
|
| 3020 |
+
},
|
| 3021 |
+
{
|
| 3022 |
+
"epoch": 0.12,
|
| 3023 |
+
"grad_norm": 3.4248578548431396,
|
| 3024 |
+
"learning_rate": 9.961855900687987e-05,
|
| 3025 |
+
"loss": 1.8423,
|
| 3026 |
+
"step": 215500
|
| 3027 |
+
},
|
| 3028 |
+
{
|
| 3029 |
+
"epoch": 0.12,
|
| 3030 |
+
"grad_norm": 3.3221466541290283,
|
| 3031 |
+
"learning_rate": 9.961763092660463e-05,
|
| 3032 |
+
"loss": 1.8258,
|
| 3033 |
+
"step": 216000
|
| 3034 |
+
},
|
| 3035 |
+
{
|
| 3036 |
+
"epoch": 0.12,
|
| 3037 |
+
"grad_norm": 3.8490660190582275,
|
| 3038 |
+
"learning_rate": 9.96167028463294e-05,
|
| 3039 |
+
"loss": 1.8574,
|
| 3040 |
+
"step": 216500
|
| 3041 |
+
},
|
| 3042 |
+
{
|
| 3043 |
+
"epoch": 0.12,
|
| 3044 |
+
"grad_norm": 5.5256547927856445,
|
| 3045 |
+
"learning_rate": 9.961577476605417e-05,
|
| 3046 |
+
"loss": 1.8312,
|
| 3047 |
+
"step": 217000
|
| 3048 |
+
},
|
| 3049 |
+
{
|
| 3050 |
+
"epoch": 0.12,
|
| 3051 |
+
"grad_norm": 3.6203320026397705,
|
| 3052 |
+
"learning_rate": 9.961484668577894e-05,
|
| 3053 |
+
"loss": 1.8483,
|
| 3054 |
+
"step": 217500
|
| 3055 |
+
},
|
| 3056 |
+
{
|
| 3057 |
+
"epoch": 0.12,
|
| 3058 |
+
"grad_norm": 4.3675856590271,
|
| 3059 |
+
"learning_rate": 9.961391860550371e-05,
|
| 3060 |
+
"loss": 1.8526,
|
| 3061 |
+
"step": 218000
|
| 3062 |
+
},
|
| 3063 |
+
{
|
| 3064 |
+
"epoch": 0.12,
|
| 3065 |
+
"grad_norm": 4.40755558013916,
|
| 3066 |
+
"learning_rate": 9.961299052522848e-05,
|
| 3067 |
+
"loss": 1.8617,
|
| 3068 |
+
"step": 218500
|
| 3069 |
+
},
|
| 3070 |
+
{
|
| 3071 |
+
"epoch": 0.12,
|
| 3072 |
+
"grad_norm": 12.9111967086792,
|
| 3073 |
+
"learning_rate": 9.961206244495324e-05,
|
| 3074 |
+
"loss": 1.8488,
|
| 3075 |
+
"step": 219000
|
| 3076 |
+
},
|
| 3077 |
+
{
|
| 3078 |
+
"epoch": 0.12,
|
| 3079 |
+
"grad_norm": 4.0077009201049805,
|
| 3080 |
+
"learning_rate": 9.961113436467801e-05,
|
| 3081 |
+
"loss": 1.8336,
|
| 3082 |
+
"step": 219500
|
| 3083 |
+
},
|
| 3084 |
+
{
|
| 3085 |
+
"epoch": 0.12,
|
| 3086 |
+
"grad_norm": 16.747453689575195,
|
| 3087 |
+
"learning_rate": 9.961020628440278e-05,
|
| 3088 |
+
"loss": 1.8535,
|
| 3089 |
+
"step": 220000
|
| 3090 |
+
},
|
| 3091 |
+
{
|
| 3092 |
+
"epoch": 0.12,
|
| 3093 |
+
"grad_norm": 5.349842071533203,
|
| 3094 |
+
"learning_rate": 9.960927820412756e-05,
|
| 3095 |
+
"loss": 1.8272,
|
| 3096 |
+
"step": 220500
|
| 3097 |
+
},
|
| 3098 |
+
{
|
| 3099 |
+
"epoch": 0.12,
|
| 3100 |
+
"grad_norm": 4.0585808753967285,
|
| 3101 |
+
"learning_rate": 9.960835012385231e-05,
|
| 3102 |
+
"loss": 1.8361,
|
| 3103 |
+
"step": 221000
|
| 3104 |
+
},
|
| 3105 |
+
{
|
| 3106 |
+
"epoch": 0.12,
|
| 3107 |
+
"grad_norm": 10.322525978088379,
|
| 3108 |
+
"learning_rate": 9.960742204357709e-05,
|
| 3109 |
+
"loss": 1.8171,
|
| 3110 |
+
"step": 221500
|
| 3111 |
+
},
|
| 3112 |
+
{
|
| 3113 |
+
"epoch": 0.12,
|
| 3114 |
+
"grad_norm": 3.9094197750091553,
|
| 3115 |
+
"learning_rate": 9.960649396330186e-05,
|
| 3116 |
+
"loss": 1.8434,
|
| 3117 |
+
"step": 222000
|
| 3118 |
+
},
|
| 3119 |
+
{
|
| 3120 |
+
"epoch": 0.12,
|
| 3121 |
+
"grad_norm": 4.250068187713623,
|
| 3122 |
+
"learning_rate": 9.960556588302661e-05,
|
| 3123 |
+
"loss": 1.8312,
|
| 3124 |
+
"step": 222500
|
| 3125 |
+
},
|
| 3126 |
+
{
|
| 3127 |
+
"epoch": 0.12,
|
| 3128 |
+
"grad_norm": 3.535179853439331,
|
| 3129 |
+
"learning_rate": 9.960463780275139e-05,
|
| 3130 |
+
"loss": 1.8263,
|
| 3131 |
+
"step": 223000
|
| 3132 |
+
},
|
| 3133 |
+
{
|
| 3134 |
+
"epoch": 0.12,
|
| 3135 |
+
"grad_norm": 3.5814437866210938,
|
| 3136 |
+
"learning_rate": 9.960370972247616e-05,
|
| 3137 |
+
"loss": 1.833,
|
| 3138 |
+
"step": 223500
|
| 3139 |
+
},
|
| 3140 |
+
{
|
| 3141 |
+
"epoch": 0.12,
|
| 3142 |
+
"grad_norm": 5.914824962615967,
|
| 3143 |
+
"learning_rate": 9.960278164220093e-05,
|
| 3144 |
+
"loss": 1.8196,
|
| 3145 |
+
"step": 224000
|
| 3146 |
+
},
|
| 3147 |
+
{
|
| 3148 |
+
"epoch": 0.12,
|
| 3149 |
+
"grad_norm": 3.966111660003662,
|
| 3150 |
+
"learning_rate": 9.96018535619257e-05,
|
| 3151 |
+
"loss": 1.8219,
|
| 3152 |
+
"step": 224500
|
| 3153 |
+
},
|
| 3154 |
+
{
|
| 3155 |
+
"epoch": 0.13,
|
| 3156 |
+
"grad_norm": 3.899972915649414,
|
| 3157 |
+
"learning_rate": 9.960092548165046e-05,
|
| 3158 |
+
"loss": 1.8173,
|
| 3159 |
+
"step": 225000
|
| 3160 |
+
},
|
| 3161 |
+
{
|
| 3162 |
+
"epoch": 0.13,
|
| 3163 |
+
"grad_norm": 3.0840604305267334,
|
| 3164 |
+
"learning_rate": 9.959999740137523e-05,
|
| 3165 |
+
"loss": 1.8036,
|
| 3166 |
+
"step": 225500
|
| 3167 |
+
},
|
| 3168 |
+
{
|
| 3169 |
+
"epoch": 0.13,
|
| 3170 |
+
"grad_norm": 4.453856468200684,
|
| 3171 |
+
"learning_rate": 9.95990693211e-05,
|
| 3172 |
+
"loss": 1.839,
|
| 3173 |
+
"step": 226000
|
| 3174 |
+
},
|
| 3175 |
+
{
|
| 3176 |
+
"epoch": 0.13,
|
| 3177 |
+
"grad_norm": 5.8350443840026855,
|
| 3178 |
+
"learning_rate": 9.959814124082477e-05,
|
| 3179 |
+
"loss": 1.7956,
|
| 3180 |
+
"step": 226500
|
| 3181 |
+
},
|
| 3182 |
+
{
|
| 3183 |
+
"epoch": 0.13,
|
| 3184 |
+
"grad_norm": 3.109984874725342,
|
| 3185 |
+
"learning_rate": 9.959721316054954e-05,
|
| 3186 |
+
"loss": 1.8198,
|
| 3187 |
+
"step": 227000
|
| 3188 |
+
},
|
| 3189 |
+
{
|
| 3190 |
+
"epoch": 0.13,
|
| 3191 |
+
"grad_norm": 5.034292221069336,
|
| 3192 |
+
"learning_rate": 9.959628508027431e-05,
|
| 3193 |
+
"loss": 1.7978,
|
| 3194 |
+
"step": 227500
|
| 3195 |
+
},
|
| 3196 |
+
{
|
| 3197 |
+
"epoch": 0.13,
|
| 3198 |
+
"grad_norm": 5.237682819366455,
|
| 3199 |
+
"learning_rate": 9.959535699999907e-05,
|
| 3200 |
+
"loss": 1.8309,
|
| 3201 |
+
"step": 228000
|
| 3202 |
+
},
|
| 3203 |
+
{
|
| 3204 |
+
"epoch": 0.13,
|
| 3205 |
+
"grad_norm": 3.0368969440460205,
|
| 3206 |
+
"learning_rate": 9.959442891972384e-05,
|
| 3207 |
+
"loss": 1.8012,
|
| 3208 |
+
"step": 228500
|
| 3209 |
+
},
|
| 3210 |
+
{
|
| 3211 |
+
"epoch": 0.13,
|
| 3212 |
+
"grad_norm": 3.3586440086364746,
|
| 3213 |
+
"learning_rate": 9.959350083944862e-05,
|
| 3214 |
+
"loss": 1.7965,
|
| 3215 |
+
"step": 229000
|
| 3216 |
+
},
|
| 3217 |
+
{
|
| 3218 |
+
"epoch": 0.13,
|
| 3219 |
+
"grad_norm": 4.194300651550293,
|
| 3220 |
+
"learning_rate": 9.959257275917339e-05,
|
| 3221 |
+
"loss": 1.8158,
|
| 3222 |
+
"step": 229500
|
| 3223 |
+
},
|
| 3224 |
+
{
|
| 3225 |
+
"epoch": 0.13,
|
| 3226 |
+
"grad_norm": 5.295820713043213,
|
| 3227 |
+
"learning_rate": 9.959164467889816e-05,
|
| 3228 |
+
"loss": 1.8035,
|
| 3229 |
+
"step": 230000
|
| 3230 |
+
},
|
| 3231 |
+
{
|
| 3232 |
+
"epoch": 0.13,
|
| 3233 |
+
"grad_norm": 3.1233255863189697,
|
| 3234 |
+
"learning_rate": 9.959071659862293e-05,
|
| 3235 |
+
"loss": 1.8076,
|
| 3236 |
+
"step": 230500
|
| 3237 |
+
},
|
| 3238 |
+
{
|
| 3239 |
+
"epoch": 0.13,
|
| 3240 |
+
"grad_norm": 4.108482837677002,
|
| 3241 |
+
"learning_rate": 9.958978851834769e-05,
|
| 3242 |
+
"loss": 1.8169,
|
| 3243 |
+
"step": 231000
|
| 3244 |
+
},
|
| 3245 |
+
{
|
| 3246 |
+
"epoch": 0.13,
|
| 3247 |
+
"grad_norm": 3.3223822116851807,
|
| 3248 |
+
"learning_rate": 9.958886043807245e-05,
|
| 3249 |
+
"loss": 1.7914,
|
| 3250 |
+
"step": 231500
|
| 3251 |
+
},
|
| 3252 |
+
{
|
| 3253 |
+
"epoch": 0.13,
|
| 3254 |
+
"grad_norm": 3.5875344276428223,
|
| 3255 |
+
"learning_rate": 9.958793235779722e-05,
|
| 3256 |
+
"loss": 1.8124,
|
| 3257 |
+
"step": 232000
|
| 3258 |
+
},
|
| 3259 |
+
{
|
| 3260 |
+
"epoch": 0.13,
|
| 3261 |
+
"grad_norm": 3.7725822925567627,
|
| 3262 |
+
"learning_rate": 9.958700427752199e-05,
|
| 3263 |
+
"loss": 1.7907,
|
| 3264 |
+
"step": 232500
|
| 3265 |
+
},
|
| 3266 |
+
{
|
| 3267 |
+
"epoch": 0.13,
|
| 3268 |
+
"grad_norm": 3.5657663345336914,
|
| 3269 |
+
"learning_rate": 9.958607619724676e-05,
|
| 3270 |
+
"loss": 1.795,
|
| 3271 |
+
"step": 233000
|
| 3272 |
+
},
|
| 3273 |
+
{
|
| 3274 |
+
"epoch": 0.13,
|
| 3275 |
+
"grad_norm": 4.3107404708862305,
|
| 3276 |
+
"learning_rate": 9.958514811697153e-05,
|
| 3277 |
+
"loss": 1.7687,
|
| 3278 |
+
"step": 233500
|
| 3279 |
+
},
|
| 3280 |
+
{
|
| 3281 |
+
"epoch": 0.13,
|
| 3282 |
+
"grad_norm": 4.887343883514404,
|
| 3283 |
+
"learning_rate": 9.95842200366963e-05,
|
| 3284 |
+
"loss": 1.7831,
|
| 3285 |
+
"step": 234000
|
| 3286 |
+
},
|
| 3287 |
+
{
|
| 3288 |
+
"epoch": 0.13,
|
| 3289 |
+
"grad_norm": 3.7076539993286133,
|
| 3290 |
+
"learning_rate": 9.958329195642106e-05,
|
| 3291 |
+
"loss": 1.7842,
|
| 3292 |
+
"step": 234500
|
| 3293 |
+
},
|
| 3294 |
+
{
|
| 3295 |
+
"epoch": 0.13,
|
| 3296 |
+
"grad_norm": 3.3431081771850586,
|
| 3297 |
+
"learning_rate": 9.958236387614583e-05,
|
| 3298 |
+
"loss": 1.797,
|
| 3299 |
+
"step": 235000
|
| 3300 |
+
},
|
| 3301 |
+
{
|
| 3302 |
+
"epoch": 0.13,
|
| 3303 |
+
"grad_norm": 2.8526740074157715,
|
| 3304 |
+
"learning_rate": 9.95814357958706e-05,
|
| 3305 |
+
"loss": 1.7879,
|
| 3306 |
+
"step": 235500
|
| 3307 |
+
},
|
| 3308 |
+
{
|
| 3309 |
+
"epoch": 0.13,
|
| 3310 |
+
"grad_norm": 4.57257604598999,
|
| 3311 |
+
"learning_rate": 9.958050771559537e-05,
|
| 3312 |
+
"loss": 1.7924,
|
| 3313 |
+
"step": 236000
|
| 3314 |
+
},
|
| 3315 |
+
{
|
| 3316 |
+
"epoch": 0.13,
|
| 3317 |
+
"grad_norm": 8.872370719909668,
|
| 3318 |
+
"learning_rate": 9.957957963532015e-05,
|
| 3319 |
+
"loss": 1.796,
|
| 3320 |
+
"step": 236500
|
| 3321 |
+
},
|
| 3322 |
+
{
|
| 3323 |
+
"epoch": 0.13,
|
| 3324 |
+
"grad_norm": 3.647901773452759,
|
| 3325 |
+
"learning_rate": 9.957865155504492e-05,
|
| 3326 |
+
"loss": 1.7801,
|
| 3327 |
+
"step": 237000
|
| 3328 |
+
}
|
| 3329 |
+
],
|
| 3330 |
+
"logging_steps": 500,
|
| 3331 |
+
"max_steps": 53884650,
|
| 3332 |
+
"num_input_tokens_seen": 0,
|
| 3333 |
+
"num_train_epochs": 30,
|
| 3334 |
+
"save_steps": 10,
|
| 3335 |
+
"total_flos": 1.6575934896869376e+16,
|
| 3336 |
+
"train_batch_size": 16,
|
| 3337 |
+
"trial_name": null,
|
| 3338 |
+
"trial_params": null
|
| 3339 |
+
}
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ef8acd3dba41f228f6668f97aa2a01d59dd83344f59c67d2177448d590f0f7e5
|
| 3 |
+
size 4411
|