11
Browse files- config.json +33 -0
- generation_config.json +7 -0
- logs/events.out.tfevents.1713657038.ip-10-25-205-144.147374.0 +3 -0
- model.safetensors +3 -0
- optimizer.pt +3 -0
- rng_state.pth +3 -0
- scheduler.pt +3 -0
- trainer_state.json +1219 -0
- training_args.bin +3 -0
config.json
ADDED
|
@@ -0,0 +1,33 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "google/byt5-small",
|
| 3 |
+
"architectures": [
|
| 4 |
+
"T5ForConditionalGeneration"
|
| 5 |
+
],
|
| 6 |
+
"classifier_dropout": 0.0,
|
| 7 |
+
"d_ff": 3584,
|
| 8 |
+
"d_kv": 64,
|
| 9 |
+
"d_model": 1472,
|
| 10 |
+
"decoder_start_token_id": 0,
|
| 11 |
+
"dense_act_fn": "gelu_new",
|
| 12 |
+
"dropout_rate": 0.1,
|
| 13 |
+
"eos_token_id": 1,
|
| 14 |
+
"feed_forward_proj": "gated-gelu",
|
| 15 |
+
"gradient_checkpointing": false,
|
| 16 |
+
"initializer_factor": 1.0,
|
| 17 |
+
"is_encoder_decoder": true,
|
| 18 |
+
"is_gated_act": true,
|
| 19 |
+
"layer_norm_epsilon": 1e-06,
|
| 20 |
+
"model_type": "t5",
|
| 21 |
+
"num_decoder_layers": 4,
|
| 22 |
+
"num_heads": 6,
|
| 23 |
+
"num_layers": 12,
|
| 24 |
+
"pad_token_id": 0,
|
| 25 |
+
"relative_attention_max_distance": 128,
|
| 26 |
+
"relative_attention_num_buckets": 32,
|
| 27 |
+
"tie_word_embeddings": false,
|
| 28 |
+
"tokenizer_class": "ByT5Tokenizer",
|
| 29 |
+
"torch_dtype": "float32",
|
| 30 |
+
"transformers_version": "4.35.2",
|
| 31 |
+
"use_cache": true,
|
| 32 |
+
"vocab_size": 384
|
| 33 |
+
}
|
generation_config.json
ADDED
|
@@ -0,0 +1,7 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_from_model_config": true,
|
| 3 |
+
"decoder_start_token_id": 0,
|
| 4 |
+
"eos_token_id": 1,
|
| 5 |
+
"pad_token_id": 0,
|
| 6 |
+
"transformers_version": "4.35.2"
|
| 7 |
+
}
|
logs/events.out.tfevents.1713657038.ip-10-25-205-144.147374.0
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:7828b4a1bbae6756f0f500b7952cfe497504ed6d0b3caf970c4b436c26081e6b
|
| 3 |
+
size 45338
|
model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:edaefe4b8b645cf211d87fc2a73c3f17ff02c433e6543819c4be973fb0b1d9e6
|
| 3 |
+
size 1198571496
|
optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:dd1f693a92050655102df8b9d9473753bf1039f0ae9e2581ab94f3f2077cbcf4
|
| 3 |
+
size 2397245434
|
rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a2b7f484dc2a8e5bdb68d122a616a0296870584cb3d76e740609f839f97a210d
|
| 3 |
+
size 14244
|
scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:7f2295262c8b0cd31776d8d5f370bfd4d7da47a787dc53758a73b3dbb54b80f3
|
| 3 |
+
size 1064
|
trainer_state.json
ADDED
|
@@ -0,0 +1,1219 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.5188751220703125,
|
| 3 |
+
"best_model_checkpoint": "AlexWang99/byt5_add_2k/checkpoint-450",
|
| 4 |
+
"epoch": 150.0,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 450,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 1.0,
|
| 13 |
+
"eval_loss": 7.222967147827148,
|
| 14 |
+
"eval_runtime": 10.6822,
|
| 15 |
+
"eval_samples_per_second": 936.137,
|
| 16 |
+
"eval_steps_per_second": 1.217,
|
| 17 |
+
"step": 3
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 2.0,
|
| 21 |
+
"eval_loss": 5.985061168670654,
|
| 22 |
+
"eval_runtime": 10.9203,
|
| 23 |
+
"eval_samples_per_second": 915.722,
|
| 24 |
+
"eval_steps_per_second": 1.19,
|
| 25 |
+
"step": 6
|
| 26 |
+
},
|
| 27 |
+
{
|
| 28 |
+
"epoch": 3.0,
|
| 29 |
+
"eval_loss": 5.1600236892700195,
|
| 30 |
+
"eval_runtime": 10.7228,
|
| 31 |
+
"eval_samples_per_second": 932.591,
|
| 32 |
+
"eval_steps_per_second": 1.212,
|
| 33 |
+
"step": 9
|
| 34 |
+
},
|
| 35 |
+
{
|
| 36 |
+
"epoch": 4.0,
|
| 37 |
+
"eval_loss": 4.493813514709473,
|
| 38 |
+
"eval_runtime": 11.0064,
|
| 39 |
+
"eval_samples_per_second": 908.565,
|
| 40 |
+
"eval_steps_per_second": 1.181,
|
| 41 |
+
"step": 12
|
| 42 |
+
},
|
| 43 |
+
{
|
| 44 |
+
"epoch": 5.0,
|
| 45 |
+
"eval_loss": 3.8893628120422363,
|
| 46 |
+
"eval_runtime": 10.8904,
|
| 47 |
+
"eval_samples_per_second": 918.238,
|
| 48 |
+
"eval_steps_per_second": 1.194,
|
| 49 |
+
"step": 15
|
| 50 |
+
},
|
| 51 |
+
{
|
| 52 |
+
"epoch": 6.0,
|
| 53 |
+
"eval_loss": 3.341691493988037,
|
| 54 |
+
"eval_runtime": 10.8018,
|
| 55 |
+
"eval_samples_per_second": 925.772,
|
| 56 |
+
"eval_steps_per_second": 1.204,
|
| 57 |
+
"step": 18
|
| 58 |
+
},
|
| 59 |
+
{
|
| 60 |
+
"epoch": 7.0,
|
| 61 |
+
"eval_loss": 2.9026920795440674,
|
| 62 |
+
"eval_runtime": 10.8887,
|
| 63 |
+
"eval_samples_per_second": 918.384,
|
| 64 |
+
"eval_steps_per_second": 1.194,
|
| 65 |
+
"step": 21
|
| 66 |
+
},
|
| 67 |
+
{
|
| 68 |
+
"epoch": 8.0,
|
| 69 |
+
"eval_loss": 2.606121063232422,
|
| 70 |
+
"eval_runtime": 10.8106,
|
| 71 |
+
"eval_samples_per_second": 925.018,
|
| 72 |
+
"eval_steps_per_second": 1.203,
|
| 73 |
+
"step": 24
|
| 74 |
+
},
|
| 75 |
+
{
|
| 76 |
+
"epoch": 9.0,
|
| 77 |
+
"eval_loss": 2.402636766433716,
|
| 78 |
+
"eval_runtime": 11.0832,
|
| 79 |
+
"eval_samples_per_second": 902.265,
|
| 80 |
+
"eval_steps_per_second": 1.173,
|
| 81 |
+
"step": 27
|
| 82 |
+
},
|
| 83 |
+
{
|
| 84 |
+
"epoch": 10.0,
|
| 85 |
+
"eval_loss": 2.1915464401245117,
|
| 86 |
+
"eval_runtime": 10.8272,
|
| 87 |
+
"eval_samples_per_second": 923.604,
|
| 88 |
+
"eval_steps_per_second": 1.201,
|
| 89 |
+
"step": 30
|
| 90 |
+
},
|
| 91 |
+
{
|
| 92 |
+
"epoch": 11.0,
|
| 93 |
+
"eval_loss": 2.1069843769073486,
|
| 94 |
+
"eval_runtime": 10.8305,
|
| 95 |
+
"eval_samples_per_second": 923.32,
|
| 96 |
+
"eval_steps_per_second": 1.2,
|
| 97 |
+
"step": 33
|
| 98 |
+
},
|
| 99 |
+
{
|
| 100 |
+
"epoch": 12.0,
|
| 101 |
+
"eval_loss": 2.0630228519439697,
|
| 102 |
+
"eval_runtime": 10.9163,
|
| 103 |
+
"eval_samples_per_second": 916.063,
|
| 104 |
+
"eval_steps_per_second": 1.191,
|
| 105 |
+
"step": 36
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 13.0,
|
| 109 |
+
"eval_loss": 1.9474797248840332,
|
| 110 |
+
"eval_runtime": 10.8302,
|
| 111 |
+
"eval_samples_per_second": 923.345,
|
| 112 |
+
"eval_steps_per_second": 1.2,
|
| 113 |
+
"step": 39
|
| 114 |
+
},
|
| 115 |
+
{
|
| 116 |
+
"epoch": 14.0,
|
| 117 |
+
"eval_loss": 1.9065855741500854,
|
| 118 |
+
"eval_runtime": 11.0817,
|
| 119 |
+
"eval_samples_per_second": 902.388,
|
| 120 |
+
"eval_steps_per_second": 1.173,
|
| 121 |
+
"step": 42
|
| 122 |
+
},
|
| 123 |
+
{
|
| 124 |
+
"epoch": 15.0,
|
| 125 |
+
"eval_loss": 1.8849406242370605,
|
| 126 |
+
"eval_runtime": 10.8362,
|
| 127 |
+
"eval_samples_per_second": 922.835,
|
| 128 |
+
"eval_steps_per_second": 1.2,
|
| 129 |
+
"step": 45
|
| 130 |
+
},
|
| 131 |
+
{
|
| 132 |
+
"epoch": 16.0,
|
| 133 |
+
"eval_loss": 1.7817472219467163,
|
| 134 |
+
"eval_runtime": 10.8435,
|
| 135 |
+
"eval_samples_per_second": 922.207,
|
| 136 |
+
"eval_steps_per_second": 1.199,
|
| 137 |
+
"step": 48
|
| 138 |
+
},
|
| 139 |
+
{
|
| 140 |
+
"epoch": 17.0,
|
| 141 |
+
"eval_loss": 1.7877730131149292,
|
| 142 |
+
"eval_runtime": 10.9315,
|
| 143 |
+
"eval_samples_per_second": 914.788,
|
| 144 |
+
"eval_steps_per_second": 1.189,
|
| 145 |
+
"step": 51
|
| 146 |
+
},
|
| 147 |
+
{
|
| 148 |
+
"epoch": 18.0,
|
| 149 |
+
"eval_loss": 1.7331796884536743,
|
| 150 |
+
"eval_runtime": 10.8485,
|
| 151 |
+
"eval_samples_per_second": 921.788,
|
| 152 |
+
"eval_steps_per_second": 1.198,
|
| 153 |
+
"step": 54
|
| 154 |
+
},
|
| 155 |
+
{
|
| 156 |
+
"epoch": 19.0,
|
| 157 |
+
"eval_loss": 1.692647099494934,
|
| 158 |
+
"eval_runtime": 11.0859,
|
| 159 |
+
"eval_samples_per_second": 902.043,
|
| 160 |
+
"eval_steps_per_second": 1.173,
|
| 161 |
+
"step": 57
|
| 162 |
+
},
|
| 163 |
+
{
|
| 164 |
+
"epoch": 20.0,
|
| 165 |
+
"eval_loss": 1.6727197170257568,
|
| 166 |
+
"eval_runtime": 10.8467,
|
| 167 |
+
"eval_samples_per_second": 921.936,
|
| 168 |
+
"eval_steps_per_second": 1.199,
|
| 169 |
+
"step": 60
|
| 170 |
+
},
|
| 171 |
+
{
|
| 172 |
+
"epoch": 21.0,
|
| 173 |
+
"eval_loss": 1.66093909740448,
|
| 174 |
+
"eval_runtime": 10.8432,
|
| 175 |
+
"eval_samples_per_second": 922.235,
|
| 176 |
+
"eval_steps_per_second": 1.199,
|
| 177 |
+
"step": 63
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 22.0,
|
| 181 |
+
"eval_loss": 1.6616569757461548,
|
| 182 |
+
"eval_runtime": 10.9265,
|
| 183 |
+
"eval_samples_per_second": 915.204,
|
| 184 |
+
"eval_steps_per_second": 1.19,
|
| 185 |
+
"step": 66
|
| 186 |
+
},
|
| 187 |
+
{
|
| 188 |
+
"epoch": 23.0,
|
| 189 |
+
"eval_loss": 1.6537153720855713,
|
| 190 |
+
"eval_runtime": 10.8532,
|
| 191 |
+
"eval_samples_per_second": 921.386,
|
| 192 |
+
"eval_steps_per_second": 1.198,
|
| 193 |
+
"step": 69
|
| 194 |
+
},
|
| 195 |
+
{
|
| 196 |
+
"epoch": 24.0,
|
| 197 |
+
"eval_loss": 1.645085334777832,
|
| 198 |
+
"eval_runtime": 11.0411,
|
| 199 |
+
"eval_samples_per_second": 905.709,
|
| 200 |
+
"eval_steps_per_second": 1.177,
|
| 201 |
+
"step": 72
|
| 202 |
+
},
|
| 203 |
+
{
|
| 204 |
+
"epoch": 25.0,
|
| 205 |
+
"eval_loss": 1.6413605213165283,
|
| 206 |
+
"eval_runtime": 10.7889,
|
| 207 |
+
"eval_samples_per_second": 926.88,
|
| 208 |
+
"eval_steps_per_second": 1.205,
|
| 209 |
+
"step": 75
|
| 210 |
+
},
|
| 211 |
+
{
|
| 212 |
+
"epoch": 26.0,
|
| 213 |
+
"eval_loss": 1.635939598083496,
|
| 214 |
+
"eval_runtime": 10.8026,
|
| 215 |
+
"eval_samples_per_second": 925.7,
|
| 216 |
+
"eval_steps_per_second": 1.203,
|
| 217 |
+
"step": 78
|
| 218 |
+
},
|
| 219 |
+
{
|
| 220 |
+
"epoch": 27.0,
|
| 221 |
+
"eval_loss": 1.6321375370025635,
|
| 222 |
+
"eval_runtime": 10.9127,
|
| 223 |
+
"eval_samples_per_second": 916.367,
|
| 224 |
+
"eval_steps_per_second": 1.191,
|
| 225 |
+
"step": 81
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"epoch": 28.0,
|
| 229 |
+
"eval_loss": 1.6276355981826782,
|
| 230 |
+
"eval_runtime": 10.8251,
|
| 231 |
+
"eval_samples_per_second": 923.783,
|
| 232 |
+
"eval_steps_per_second": 1.201,
|
| 233 |
+
"step": 84
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"epoch": 29.0,
|
| 237 |
+
"eval_loss": 1.6232202053070068,
|
| 238 |
+
"eval_runtime": 11.0643,
|
| 239 |
+
"eval_samples_per_second": 903.805,
|
| 240 |
+
"eval_steps_per_second": 1.175,
|
| 241 |
+
"step": 87
|
| 242 |
+
},
|
| 243 |
+
{
|
| 244 |
+
"epoch": 30.0,
|
| 245 |
+
"eval_loss": 1.618307113647461,
|
| 246 |
+
"eval_runtime": 10.8447,
|
| 247 |
+
"eval_samples_per_second": 922.106,
|
| 248 |
+
"eval_steps_per_second": 1.199,
|
| 249 |
+
"step": 90
|
| 250 |
+
},
|
| 251 |
+
{
|
| 252 |
+
"epoch": 31.0,
|
| 253 |
+
"eval_loss": 1.612666368484497,
|
| 254 |
+
"eval_runtime": 10.8371,
|
| 255 |
+
"eval_samples_per_second": 922.759,
|
| 256 |
+
"eval_steps_per_second": 1.2,
|
| 257 |
+
"step": 93
|
| 258 |
+
},
|
| 259 |
+
{
|
| 260 |
+
"epoch": 32.0,
|
| 261 |
+
"eval_loss": 1.6067509651184082,
|
| 262 |
+
"eval_runtime": 10.9144,
|
| 263 |
+
"eval_samples_per_second": 916.219,
|
| 264 |
+
"eval_steps_per_second": 1.191,
|
| 265 |
+
"step": 96
|
| 266 |
+
},
|
| 267 |
+
{
|
| 268 |
+
"epoch": 33.0,
|
| 269 |
+
"eval_loss": 1.5994765758514404,
|
| 270 |
+
"eval_runtime": 10.8587,
|
| 271 |
+
"eval_samples_per_second": 920.924,
|
| 272 |
+
"eval_steps_per_second": 1.197,
|
| 273 |
+
"step": 99
|
| 274 |
+
},
|
| 275 |
+
{
|
| 276 |
+
"epoch": 34.0,
|
| 277 |
+
"eval_loss": 1.590742588043213,
|
| 278 |
+
"eval_runtime": 11.0839,
|
| 279 |
+
"eval_samples_per_second": 902.211,
|
| 280 |
+
"eval_steps_per_second": 1.173,
|
| 281 |
+
"step": 102
|
| 282 |
+
},
|
| 283 |
+
{
|
| 284 |
+
"epoch": 35.0,
|
| 285 |
+
"eval_loss": 1.5813907384872437,
|
| 286 |
+
"eval_runtime": 10.8372,
|
| 287 |
+
"eval_samples_per_second": 922.744,
|
| 288 |
+
"eval_steps_per_second": 1.2,
|
| 289 |
+
"step": 105
|
| 290 |
+
},
|
| 291 |
+
{
|
| 292 |
+
"epoch": 36.0,
|
| 293 |
+
"eval_loss": 1.570920467376709,
|
| 294 |
+
"eval_runtime": 10.8433,
|
| 295 |
+
"eval_samples_per_second": 922.23,
|
| 296 |
+
"eval_steps_per_second": 1.199,
|
| 297 |
+
"step": 108
|
| 298 |
+
},
|
| 299 |
+
{
|
| 300 |
+
"epoch": 37.0,
|
| 301 |
+
"eval_loss": 1.5595602989196777,
|
| 302 |
+
"eval_runtime": 10.9304,
|
| 303 |
+
"eval_samples_per_second": 914.879,
|
| 304 |
+
"eval_steps_per_second": 1.189,
|
| 305 |
+
"step": 111
|
| 306 |
+
},
|
| 307 |
+
{
|
| 308 |
+
"epoch": 38.0,
|
| 309 |
+
"eval_loss": 1.5501924753189087,
|
| 310 |
+
"eval_runtime": 10.8429,
|
| 311 |
+
"eval_samples_per_second": 922.259,
|
| 312 |
+
"eval_steps_per_second": 1.199,
|
| 313 |
+
"step": 114
|
| 314 |
+
},
|
| 315 |
+
{
|
| 316 |
+
"epoch": 39.0,
|
| 317 |
+
"eval_loss": 1.5431841611862183,
|
| 318 |
+
"eval_runtime": 11.1003,
|
| 319 |
+
"eval_samples_per_second": 900.873,
|
| 320 |
+
"eval_steps_per_second": 1.171,
|
| 321 |
+
"step": 117
|
| 322 |
+
},
|
| 323 |
+
{
|
| 324 |
+
"epoch": 40.0,
|
| 325 |
+
"eval_loss": 1.5323028564453125,
|
| 326 |
+
"eval_runtime": 10.8468,
|
| 327 |
+
"eval_samples_per_second": 921.928,
|
| 328 |
+
"eval_steps_per_second": 1.199,
|
| 329 |
+
"step": 120
|
| 330 |
+
},
|
| 331 |
+
{
|
| 332 |
+
"epoch": 41.0,
|
| 333 |
+
"eval_loss": 1.5221501588821411,
|
| 334 |
+
"eval_runtime": 10.8395,
|
| 335 |
+
"eval_samples_per_second": 922.551,
|
| 336 |
+
"eval_steps_per_second": 1.199,
|
| 337 |
+
"step": 123
|
| 338 |
+
},
|
| 339 |
+
{
|
| 340 |
+
"epoch": 42.0,
|
| 341 |
+
"eval_loss": 1.5130882263183594,
|
| 342 |
+
"eval_runtime": 10.9396,
|
| 343 |
+
"eval_samples_per_second": 914.108,
|
| 344 |
+
"eval_steps_per_second": 1.188,
|
| 345 |
+
"step": 126
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 43.0,
|
| 349 |
+
"eval_loss": 1.5008268356323242,
|
| 350 |
+
"eval_runtime": 10.8617,
|
| 351 |
+
"eval_samples_per_second": 920.668,
|
| 352 |
+
"eval_steps_per_second": 1.197,
|
| 353 |
+
"step": 129
|
| 354 |
+
},
|
| 355 |
+
{
|
| 356 |
+
"epoch": 44.0,
|
| 357 |
+
"eval_loss": 1.487107515335083,
|
| 358 |
+
"eval_runtime": 11.0741,
|
| 359 |
+
"eval_samples_per_second": 903.011,
|
| 360 |
+
"eval_steps_per_second": 1.174,
|
| 361 |
+
"step": 132
|
| 362 |
+
},
|
| 363 |
+
{
|
| 364 |
+
"epoch": 45.0,
|
| 365 |
+
"eval_loss": 1.4803069829940796,
|
| 366 |
+
"eval_runtime": 10.8489,
|
| 367 |
+
"eval_samples_per_second": 921.753,
|
| 368 |
+
"eval_steps_per_second": 1.198,
|
| 369 |
+
"step": 135
|
| 370 |
+
},
|
| 371 |
+
{
|
| 372 |
+
"epoch": 46.0,
|
| 373 |
+
"eval_loss": 1.4744497537612915,
|
| 374 |
+
"eval_runtime": 10.8329,
|
| 375 |
+
"eval_samples_per_second": 923.112,
|
| 376 |
+
"eval_steps_per_second": 1.2,
|
| 377 |
+
"step": 138
|
| 378 |
+
},
|
| 379 |
+
{
|
| 380 |
+
"epoch": 47.0,
|
| 381 |
+
"eval_loss": 1.4682186841964722,
|
| 382 |
+
"eval_runtime": 11.0865,
|
| 383 |
+
"eval_samples_per_second": 902.0,
|
| 384 |
+
"eval_steps_per_second": 1.173,
|
| 385 |
+
"step": 141
|
| 386 |
+
},
|
| 387 |
+
{
|
| 388 |
+
"epoch": 48.0,
|
| 389 |
+
"eval_loss": 1.4509010314941406,
|
| 390 |
+
"eval_runtime": 10.862,
|
| 391 |
+
"eval_samples_per_second": 920.64,
|
| 392 |
+
"eval_steps_per_second": 1.197,
|
| 393 |
+
"step": 144
|
| 394 |
+
},
|
| 395 |
+
{
|
| 396 |
+
"epoch": 49.0,
|
| 397 |
+
"eval_loss": 1.4542888402938843,
|
| 398 |
+
"eval_runtime": 11.0884,
|
| 399 |
+
"eval_samples_per_second": 901.845,
|
| 400 |
+
"eval_steps_per_second": 1.172,
|
| 401 |
+
"step": 147
|
| 402 |
+
},
|
| 403 |
+
{
|
| 404 |
+
"epoch": 50.0,
|
| 405 |
+
"eval_loss": 1.455941081047058,
|
| 406 |
+
"eval_runtime": 10.8598,
|
| 407 |
+
"eval_samples_per_second": 920.824,
|
| 408 |
+
"eval_steps_per_second": 1.197,
|
| 409 |
+
"step": 150
|
| 410 |
+
},
|
| 411 |
+
{
|
| 412 |
+
"epoch": 51.0,
|
| 413 |
+
"eval_loss": 1.4334802627563477,
|
| 414 |
+
"eval_runtime": 10.8586,
|
| 415 |
+
"eval_samples_per_second": 920.928,
|
| 416 |
+
"eval_steps_per_second": 1.197,
|
| 417 |
+
"step": 153
|
| 418 |
+
},
|
| 419 |
+
{
|
| 420 |
+
"epoch": 52.0,
|
| 421 |
+
"eval_loss": 1.4138292074203491,
|
| 422 |
+
"eval_runtime": 10.87,
|
| 423 |
+
"eval_samples_per_second": 919.964,
|
| 424 |
+
"eval_steps_per_second": 1.196,
|
| 425 |
+
"step": 156
|
| 426 |
+
},
|
| 427 |
+
{
|
| 428 |
+
"epoch": 53.0,
|
| 429 |
+
"eval_loss": 1.4225151538848877,
|
| 430 |
+
"eval_runtime": 10.8028,
|
| 431 |
+
"eval_samples_per_second": 925.687,
|
| 432 |
+
"eval_steps_per_second": 1.203,
|
| 433 |
+
"step": 159
|
| 434 |
+
},
|
| 435 |
+
{
|
| 436 |
+
"epoch": 54.0,
|
| 437 |
+
"eval_loss": 1.4350632429122925,
|
| 438 |
+
"eval_runtime": 11.0684,
|
| 439 |
+
"eval_samples_per_second": 903.472,
|
| 440 |
+
"eval_steps_per_second": 1.175,
|
| 441 |
+
"step": 162
|
| 442 |
+
},
|
| 443 |
+
{
|
| 444 |
+
"epoch": 55.0,
|
| 445 |
+
"eval_loss": 1.3809276819229126,
|
| 446 |
+
"eval_runtime": 10.7633,
|
| 447 |
+
"eval_samples_per_second": 929.082,
|
| 448 |
+
"eval_steps_per_second": 1.208,
|
| 449 |
+
"step": 165
|
| 450 |
+
},
|
| 451 |
+
{
|
| 452 |
+
"epoch": 56.0,
|
| 453 |
+
"eval_loss": 1.370450735092163,
|
| 454 |
+
"eval_runtime": 10.785,
|
| 455 |
+
"eval_samples_per_second": 927.215,
|
| 456 |
+
"eval_steps_per_second": 1.205,
|
| 457 |
+
"step": 168
|
| 458 |
+
},
|
| 459 |
+
{
|
| 460 |
+
"epoch": 57.0,
|
| 461 |
+
"eval_loss": 1.3955777883529663,
|
| 462 |
+
"eval_runtime": 10.895,
|
| 463 |
+
"eval_samples_per_second": 917.848,
|
| 464 |
+
"eval_steps_per_second": 1.193,
|
| 465 |
+
"step": 171
|
| 466 |
+
},
|
| 467 |
+
{
|
| 468 |
+
"epoch": 58.0,
|
| 469 |
+
"eval_loss": 1.368485927581787,
|
| 470 |
+
"eval_runtime": 10.819,
|
| 471 |
+
"eval_samples_per_second": 924.297,
|
| 472 |
+
"eval_steps_per_second": 1.202,
|
| 473 |
+
"step": 174
|
| 474 |
+
},
|
| 475 |
+
{
|
| 476 |
+
"epoch": 59.0,
|
| 477 |
+
"eval_loss": 1.33143949508667,
|
| 478 |
+
"eval_runtime": 10.8396,
|
| 479 |
+
"eval_samples_per_second": 922.543,
|
| 480 |
+
"eval_steps_per_second": 1.199,
|
| 481 |
+
"step": 177
|
| 482 |
+
},
|
| 483 |
+
{
|
| 484 |
+
"epoch": 60.0,
|
| 485 |
+
"eval_loss": 1.3212394714355469,
|
| 486 |
+
"eval_runtime": 10.8092,
|
| 487 |
+
"eval_samples_per_second": 925.139,
|
| 488 |
+
"eval_steps_per_second": 1.203,
|
| 489 |
+
"step": 180
|
| 490 |
+
},
|
| 491 |
+
{
|
| 492 |
+
"epoch": 61.0,
|
| 493 |
+
"eval_loss": 1.3334152698516846,
|
| 494 |
+
"eval_runtime": 10.9031,
|
| 495 |
+
"eval_samples_per_second": 917.174,
|
| 496 |
+
"eval_steps_per_second": 1.192,
|
| 497 |
+
"step": 183
|
| 498 |
+
},
|
| 499 |
+
{
|
| 500 |
+
"epoch": 62.0,
|
| 501 |
+
"eval_loss": 1.3178024291992188,
|
| 502 |
+
"eval_runtime": 10.8916,
|
| 503 |
+
"eval_samples_per_second": 918.135,
|
| 504 |
+
"eval_steps_per_second": 1.194,
|
| 505 |
+
"step": 186
|
| 506 |
+
},
|
| 507 |
+
{
|
| 508 |
+
"epoch": 63.0,
|
| 509 |
+
"eval_loss": 1.2850910425186157,
|
| 510 |
+
"eval_runtime": 10.7826,
|
| 511 |
+
"eval_samples_per_second": 927.417,
|
| 512 |
+
"eval_steps_per_second": 1.206,
|
| 513 |
+
"step": 189
|
| 514 |
+
},
|
| 515 |
+
{
|
| 516 |
+
"epoch": 64.0,
|
| 517 |
+
"eval_loss": 1.2826738357543945,
|
| 518 |
+
"eval_runtime": 11.0431,
|
| 519 |
+
"eval_samples_per_second": 905.544,
|
| 520 |
+
"eval_steps_per_second": 1.177,
|
| 521 |
+
"step": 192
|
| 522 |
+
},
|
| 523 |
+
{
|
| 524 |
+
"epoch": 65.0,
|
| 525 |
+
"eval_loss": 1.276419758796692,
|
| 526 |
+
"eval_runtime": 10.8459,
|
| 527 |
+
"eval_samples_per_second": 922.009,
|
| 528 |
+
"eval_steps_per_second": 1.199,
|
| 529 |
+
"step": 195
|
| 530 |
+
},
|
| 531 |
+
{
|
| 532 |
+
"epoch": 66.0,
|
| 533 |
+
"eval_loss": 1.2498911619186401,
|
| 534 |
+
"eval_runtime": 10.8147,
|
| 535 |
+
"eval_samples_per_second": 924.671,
|
| 536 |
+
"eval_steps_per_second": 1.202,
|
| 537 |
+
"step": 198
|
| 538 |
+
},
|
| 539 |
+
{
|
| 540 |
+
"epoch": 67.0,
|
| 541 |
+
"eval_loss": 1.2304234504699707,
|
| 542 |
+
"eval_runtime": 10.9437,
|
| 543 |
+
"eval_samples_per_second": 913.768,
|
| 544 |
+
"eval_steps_per_second": 1.188,
|
| 545 |
+
"step": 201
|
| 546 |
+
},
|
| 547 |
+
{
|
| 548 |
+
"epoch": 68.0,
|
| 549 |
+
"eval_loss": 1.2450603246688843,
|
| 550 |
+
"eval_runtime": 10.8378,
|
| 551 |
+
"eval_samples_per_second": 922.698,
|
| 552 |
+
"eval_steps_per_second": 1.2,
|
| 553 |
+
"step": 204
|
| 554 |
+
},
|
| 555 |
+
{
|
| 556 |
+
"epoch": 69.0,
|
| 557 |
+
"eval_loss": 1.2278028726577759,
|
| 558 |
+
"eval_runtime": 11.123,
|
| 559 |
+
"eval_samples_per_second": 899.041,
|
| 560 |
+
"eval_steps_per_second": 1.169,
|
| 561 |
+
"step": 207
|
| 562 |
+
},
|
| 563 |
+
{
|
| 564 |
+
"epoch": 70.0,
|
| 565 |
+
"eval_loss": 1.2043875455856323,
|
| 566 |
+
"eval_runtime": 10.7963,
|
| 567 |
+
"eval_samples_per_second": 926.245,
|
| 568 |
+
"eval_steps_per_second": 1.204,
|
| 569 |
+
"step": 210
|
| 570 |
+
},
|
| 571 |
+
{
|
| 572 |
+
"epoch": 71.0,
|
| 573 |
+
"eval_loss": 1.199507236480713,
|
| 574 |
+
"eval_runtime": 10.8038,
|
| 575 |
+
"eval_samples_per_second": 925.601,
|
| 576 |
+
"eval_steps_per_second": 1.203,
|
| 577 |
+
"step": 213
|
| 578 |
+
},
|
| 579 |
+
{
|
| 580 |
+
"epoch": 72.0,
|
| 581 |
+
"eval_loss": 1.1909499168395996,
|
| 582 |
+
"eval_runtime": 10.9182,
|
| 583 |
+
"eval_samples_per_second": 915.902,
|
| 584 |
+
"eval_steps_per_second": 1.191,
|
| 585 |
+
"step": 216
|
| 586 |
+
},
|
| 587 |
+
{
|
| 588 |
+
"epoch": 73.0,
|
| 589 |
+
"eval_loss": 1.172565221786499,
|
| 590 |
+
"eval_runtime": 10.8472,
|
| 591 |
+
"eval_samples_per_second": 921.898,
|
| 592 |
+
"eval_steps_per_second": 1.198,
|
| 593 |
+
"step": 219
|
| 594 |
+
},
|
| 595 |
+
{
|
| 596 |
+
"epoch": 74.0,
|
| 597 |
+
"eval_loss": 1.1473827362060547,
|
| 598 |
+
"eval_runtime": 11.0541,
|
| 599 |
+
"eval_samples_per_second": 904.638,
|
| 600 |
+
"eval_steps_per_second": 1.176,
|
| 601 |
+
"step": 222
|
| 602 |
+
},
|
| 603 |
+
{
|
| 604 |
+
"epoch": 75.0,
|
| 605 |
+
"eval_loss": 1.1195847988128662,
|
| 606 |
+
"eval_runtime": 10.924,
|
| 607 |
+
"eval_samples_per_second": 915.419,
|
| 608 |
+
"eval_steps_per_second": 1.19,
|
| 609 |
+
"step": 225
|
| 610 |
+
},
|
| 611 |
+
{
|
| 612 |
+
"epoch": 76.0,
|
| 613 |
+
"eval_loss": 1.1185705661773682,
|
| 614 |
+
"eval_runtime": 10.8085,
|
| 615 |
+
"eval_samples_per_second": 925.199,
|
| 616 |
+
"eval_steps_per_second": 1.203,
|
| 617 |
+
"step": 228
|
| 618 |
+
},
|
| 619 |
+
{
|
| 620 |
+
"epoch": 77.0,
|
| 621 |
+
"eval_loss": 1.1256279945373535,
|
| 622 |
+
"eval_runtime": 11.0649,
|
| 623 |
+
"eval_samples_per_second": 903.755,
|
| 624 |
+
"eval_steps_per_second": 1.175,
|
| 625 |
+
"step": 231
|
| 626 |
+
},
|
| 627 |
+
{
|
| 628 |
+
"epoch": 78.0,
|
| 629 |
+
"eval_loss": 1.0834866762161255,
|
| 630 |
+
"eval_runtime": 10.8281,
|
| 631 |
+
"eval_samples_per_second": 923.523,
|
| 632 |
+
"eval_steps_per_second": 1.201,
|
| 633 |
+
"step": 234
|
| 634 |
+
},
|
| 635 |
+
{
|
| 636 |
+
"epoch": 79.0,
|
| 637 |
+
"eval_loss": 1.0588449239730835,
|
| 638 |
+
"eval_runtime": 10.7832,
|
| 639 |
+
"eval_samples_per_second": 927.372,
|
| 640 |
+
"eval_steps_per_second": 1.206,
|
| 641 |
+
"step": 237
|
| 642 |
+
},
|
| 643 |
+
{
|
| 644 |
+
"epoch": 80.0,
|
| 645 |
+
"eval_loss": 1.0453879833221436,
|
| 646 |
+
"eval_runtime": 10.8835,
|
| 647 |
+
"eval_samples_per_second": 918.819,
|
| 648 |
+
"eval_steps_per_second": 1.194,
|
| 649 |
+
"step": 240
|
| 650 |
+
},
|
| 651 |
+
{
|
| 652 |
+
"epoch": 81.0,
|
| 653 |
+
"eval_loss": 1.0334476232528687,
|
| 654 |
+
"eval_runtime": 10.8064,
|
| 655 |
+
"eval_samples_per_second": 925.376,
|
| 656 |
+
"eval_steps_per_second": 1.203,
|
| 657 |
+
"step": 243
|
| 658 |
+
},
|
| 659 |
+
{
|
| 660 |
+
"epoch": 82.0,
|
| 661 |
+
"eval_loss": 1.028348684310913,
|
| 662 |
+
"eval_runtime": 10.9037,
|
| 663 |
+
"eval_samples_per_second": 917.121,
|
| 664 |
+
"eval_steps_per_second": 1.192,
|
| 665 |
+
"step": 246
|
| 666 |
+
},
|
| 667 |
+
{
|
| 668 |
+
"epoch": 83.0,
|
| 669 |
+
"eval_loss": 1.0091253519058228,
|
| 670 |
+
"eval_runtime": 10.8016,
|
| 671 |
+
"eval_samples_per_second": 925.788,
|
| 672 |
+
"eval_steps_per_second": 1.204,
|
| 673 |
+
"step": 249
|
| 674 |
+
},
|
| 675 |
+
{
|
| 676 |
+
"epoch": 84.0,
|
| 677 |
+
"eval_loss": 0.9820342659950256,
|
| 678 |
+
"eval_runtime": 11.0485,
|
| 679 |
+
"eval_samples_per_second": 905.101,
|
| 680 |
+
"eval_steps_per_second": 1.177,
|
| 681 |
+
"step": 252
|
| 682 |
+
},
|
| 683 |
+
{
|
| 684 |
+
"epoch": 85.0,
|
| 685 |
+
"eval_loss": 0.9488591551780701,
|
| 686 |
+
"eval_runtime": 10.8216,
|
| 687 |
+
"eval_samples_per_second": 924.074,
|
| 688 |
+
"eval_steps_per_second": 1.201,
|
| 689 |
+
"step": 255
|
| 690 |
+
},
|
| 691 |
+
{
|
| 692 |
+
"epoch": 86.0,
|
| 693 |
+
"eval_loss": 0.9406836032867432,
|
| 694 |
+
"eval_runtime": 10.8253,
|
| 695 |
+
"eval_samples_per_second": 923.765,
|
| 696 |
+
"eval_steps_per_second": 1.201,
|
| 697 |
+
"step": 258
|
| 698 |
+
},
|
| 699 |
+
{
|
| 700 |
+
"epoch": 87.0,
|
| 701 |
+
"eval_loss": 0.9392226338386536,
|
| 702 |
+
"eval_runtime": 10.9531,
|
| 703 |
+
"eval_samples_per_second": 912.984,
|
| 704 |
+
"eval_steps_per_second": 1.187,
|
| 705 |
+
"step": 261
|
| 706 |
+
},
|
| 707 |
+
{
|
| 708 |
+
"epoch": 88.0,
|
| 709 |
+
"eval_loss": 0.9232249855995178,
|
| 710 |
+
"eval_runtime": 10.835,
|
| 711 |
+
"eval_samples_per_second": 922.938,
|
| 712 |
+
"eval_steps_per_second": 1.2,
|
| 713 |
+
"step": 264
|
| 714 |
+
},
|
| 715 |
+
{
|
| 716 |
+
"epoch": 89.0,
|
| 717 |
+
"eval_loss": 0.902049720287323,
|
| 718 |
+
"eval_runtime": 10.915,
|
| 719 |
+
"eval_samples_per_second": 916.173,
|
| 720 |
+
"eval_steps_per_second": 1.191,
|
| 721 |
+
"step": 267
|
| 722 |
+
},
|
| 723 |
+
{
|
| 724 |
+
"epoch": 90.0,
|
| 725 |
+
"eval_loss": 0.8934366106987,
|
| 726 |
+
"eval_runtime": 10.833,
|
| 727 |
+
"eval_samples_per_second": 923.105,
|
| 728 |
+
"eval_steps_per_second": 1.2,
|
| 729 |
+
"step": 270
|
| 730 |
+
},
|
| 731 |
+
{
|
| 732 |
+
"epoch": 91.0,
|
| 733 |
+
"eval_loss": 0.8797369003295898,
|
| 734 |
+
"eval_runtime": 10.8275,
|
| 735 |
+
"eval_samples_per_second": 923.576,
|
| 736 |
+
"eval_steps_per_second": 1.201,
|
| 737 |
+
"step": 273
|
| 738 |
+
},
|
| 739 |
+
{
|
| 740 |
+
"epoch": 92.0,
|
| 741 |
+
"eval_loss": 0.8834591507911682,
|
| 742 |
+
"eval_runtime": 10.8227,
|
| 743 |
+
"eval_samples_per_second": 923.98,
|
| 744 |
+
"eval_steps_per_second": 1.201,
|
| 745 |
+
"step": 276
|
| 746 |
+
},
|
| 747 |
+
{
|
| 748 |
+
"epoch": 93.0,
|
| 749 |
+
"eval_loss": 0.8583576083183289,
|
| 750 |
+
"eval_runtime": 10.8262,
|
| 751 |
+
"eval_samples_per_second": 923.681,
|
| 752 |
+
"eval_steps_per_second": 1.201,
|
| 753 |
+
"step": 279
|
| 754 |
+
},
|
| 755 |
+
{
|
| 756 |
+
"epoch": 94.0,
|
| 757 |
+
"eval_loss": 0.8160658478736877,
|
| 758 |
+
"eval_runtime": 10.8597,
|
| 759 |
+
"eval_samples_per_second": 920.834,
|
| 760 |
+
"eval_steps_per_second": 1.197,
|
| 761 |
+
"step": 282
|
| 762 |
+
},
|
| 763 |
+
{
|
| 764 |
+
"epoch": 95.0,
|
| 765 |
+
"eval_loss": 0.7998712658882141,
|
| 766 |
+
"eval_runtime": 10.7874,
|
| 767 |
+
"eval_samples_per_second": 927.007,
|
| 768 |
+
"eval_steps_per_second": 1.205,
|
| 769 |
+
"step": 285
|
| 770 |
+
},
|
| 771 |
+
{
|
| 772 |
+
"epoch": 96.0,
|
| 773 |
+
"eval_loss": 0.8268041014671326,
|
| 774 |
+
"eval_runtime": 11.0529,
|
| 775 |
+
"eval_samples_per_second": 904.74,
|
| 776 |
+
"eval_steps_per_second": 1.176,
|
| 777 |
+
"step": 288
|
| 778 |
+
},
|
| 779 |
+
{
|
| 780 |
+
"epoch": 97.0,
|
| 781 |
+
"eval_loss": 0.8250266313552856,
|
| 782 |
+
"eval_runtime": 10.8021,
|
| 783 |
+
"eval_samples_per_second": 925.746,
|
| 784 |
+
"eval_steps_per_second": 1.203,
|
| 785 |
+
"step": 291
|
| 786 |
+
},
|
| 787 |
+
{
|
| 788 |
+
"epoch": 98.0,
|
| 789 |
+
"eval_loss": 0.7785258293151855,
|
| 790 |
+
"eval_runtime": 10.7519,
|
| 791 |
+
"eval_samples_per_second": 930.071,
|
| 792 |
+
"eval_steps_per_second": 1.209,
|
| 793 |
+
"step": 294
|
| 794 |
+
},
|
| 795 |
+
{
|
| 796 |
+
"epoch": 99.0,
|
| 797 |
+
"eval_loss": 0.7797490358352661,
|
| 798 |
+
"eval_runtime": 10.8842,
|
| 799 |
+
"eval_samples_per_second": 918.761,
|
| 800 |
+
"eval_steps_per_second": 1.194,
|
| 801 |
+
"step": 297
|
| 802 |
+
},
|
| 803 |
+
{
|
| 804 |
+
"epoch": 100.0,
|
| 805 |
+
"eval_loss": 0.7988857626914978,
|
| 806 |
+
"eval_runtime": 10.8165,
|
| 807 |
+
"eval_samples_per_second": 924.512,
|
| 808 |
+
"eval_steps_per_second": 1.202,
|
| 809 |
+
"step": 300
|
| 810 |
+
},
|
| 811 |
+
{
|
| 812 |
+
"epoch": 101.0,
|
| 813 |
+
"eval_loss": 0.7674239277839661,
|
| 814 |
+
"eval_runtime": 10.8984,
|
| 815 |
+
"eval_samples_per_second": 917.569,
|
| 816 |
+
"eval_steps_per_second": 1.193,
|
| 817 |
+
"step": 303
|
| 818 |
+
},
|
| 819 |
+
{
|
| 820 |
+
"epoch": 102.0,
|
| 821 |
+
"eval_loss": 0.7283704280853271,
|
| 822 |
+
"eval_runtime": 10.7882,
|
| 823 |
+
"eval_samples_per_second": 926.941,
|
| 824 |
+
"eval_steps_per_second": 1.205,
|
| 825 |
+
"step": 306
|
| 826 |
+
},
|
| 827 |
+
{
|
| 828 |
+
"epoch": 103.0,
|
| 829 |
+
"eval_loss": 0.7265847325325012,
|
| 830 |
+
"eval_runtime": 10.7846,
|
| 831 |
+
"eval_samples_per_second": 927.248,
|
| 832 |
+
"eval_steps_per_second": 1.205,
|
| 833 |
+
"step": 309
|
| 834 |
+
},
|
| 835 |
+
{
|
| 836 |
+
"epoch": 104.0,
|
| 837 |
+
"eval_loss": 0.7218329906463623,
|
| 838 |
+
"eval_runtime": 10.8432,
|
| 839 |
+
"eval_samples_per_second": 922.237,
|
| 840 |
+
"eval_steps_per_second": 1.199,
|
| 841 |
+
"step": 312
|
| 842 |
+
},
|
| 843 |
+
{
|
| 844 |
+
"epoch": 105.0,
|
| 845 |
+
"eval_loss": 0.7195408344268799,
|
| 846 |
+
"eval_runtime": 10.775,
|
| 847 |
+
"eval_samples_per_second": 928.073,
|
| 848 |
+
"eval_steps_per_second": 1.206,
|
| 849 |
+
"step": 315
|
| 850 |
+
},
|
| 851 |
+
{
|
| 852 |
+
"epoch": 106.0,
|
| 853 |
+
"eval_loss": 0.720020592212677,
|
| 854 |
+
"eval_runtime": 11.033,
|
| 855 |
+
"eval_samples_per_second": 906.372,
|
| 856 |
+
"eval_steps_per_second": 1.178,
|
| 857 |
+
"step": 318
|
| 858 |
+
},
|
| 859 |
+
{
|
| 860 |
+
"epoch": 107.0,
|
| 861 |
+
"eval_loss": 0.7001694440841675,
|
| 862 |
+
"eval_runtime": 10.7937,
|
| 863 |
+
"eval_samples_per_second": 926.47,
|
| 864 |
+
"eval_steps_per_second": 1.204,
|
| 865 |
+
"step": 321
|
| 866 |
+
},
|
| 867 |
+
{
|
| 868 |
+
"epoch": 108.0,
|
| 869 |
+
"eval_loss": 0.6704011559486389,
|
| 870 |
+
"eval_runtime": 10.9903,
|
| 871 |
+
"eval_samples_per_second": 909.893,
|
| 872 |
+
"eval_steps_per_second": 1.183,
|
| 873 |
+
"step": 324
|
| 874 |
+
},
|
| 875 |
+
{
|
| 876 |
+
"epoch": 109.0,
|
| 877 |
+
"eval_loss": 0.6617050766944885,
|
| 878 |
+
"eval_runtime": 11.0129,
|
| 879 |
+
"eval_samples_per_second": 908.026,
|
| 880 |
+
"eval_steps_per_second": 1.18,
|
| 881 |
+
"step": 327
|
| 882 |
+
},
|
| 883 |
+
{
|
| 884 |
+
"epoch": 110.0,
|
| 885 |
+
"eval_loss": 0.6687906384468079,
|
| 886 |
+
"eval_runtime": 10.7417,
|
| 887 |
+
"eval_samples_per_second": 930.948,
|
| 888 |
+
"eval_steps_per_second": 1.21,
|
| 889 |
+
"step": 330
|
| 890 |
+
},
|
| 891 |
+
{
|
| 892 |
+
"epoch": 111.0,
|
| 893 |
+
"eval_loss": 0.6637664437294006,
|
| 894 |
+
"eval_runtime": 10.8682,
|
| 895 |
+
"eval_samples_per_second": 920.118,
|
| 896 |
+
"eval_steps_per_second": 1.196,
|
| 897 |
+
"step": 333
|
| 898 |
+
},
|
| 899 |
+
{
|
| 900 |
+
"epoch": 112.0,
|
| 901 |
+
"eval_loss": 0.6454914212226868,
|
| 902 |
+
"eval_runtime": 10.7981,
|
| 903 |
+
"eval_samples_per_second": 926.087,
|
| 904 |
+
"eval_steps_per_second": 1.204,
|
| 905 |
+
"step": 336
|
| 906 |
+
},
|
| 907 |
+
{
|
| 908 |
+
"epoch": 113.0,
|
| 909 |
+
"eval_loss": 0.6308099627494812,
|
| 910 |
+
"eval_runtime": 10.8495,
|
| 911 |
+
"eval_samples_per_second": 921.699,
|
| 912 |
+
"eval_steps_per_second": 1.198,
|
| 913 |
+
"step": 339
|
| 914 |
+
},
|
| 915 |
+
{
|
| 916 |
+
"epoch": 114.0,
|
| 917 |
+
"eval_loss": 0.6411617398262024,
|
| 918 |
+
"eval_runtime": 10.9089,
|
| 919 |
+
"eval_samples_per_second": 916.679,
|
| 920 |
+
"eval_steps_per_second": 1.192,
|
| 921 |
+
"step": 342
|
| 922 |
+
},
|
| 923 |
+
{
|
| 924 |
+
"epoch": 115.0,
|
| 925 |
+
"eval_loss": 0.6422205567359924,
|
| 926 |
+
"eval_runtime": 10.7825,
|
| 927 |
+
"eval_samples_per_second": 927.428,
|
| 928 |
+
"eval_steps_per_second": 1.206,
|
| 929 |
+
"step": 345
|
| 930 |
+
},
|
| 931 |
+
{
|
| 932 |
+
"epoch": 116.0,
|
| 933 |
+
"eval_loss": 0.6265988349914551,
|
| 934 |
+
"eval_runtime": 10.865,
|
| 935 |
+
"eval_samples_per_second": 920.387,
|
| 936 |
+
"eval_steps_per_second": 1.197,
|
| 937 |
+
"step": 348
|
| 938 |
+
},
|
| 939 |
+
{
|
| 940 |
+
"epoch": 117.0,
|
| 941 |
+
"eval_loss": 0.615440845489502,
|
| 942 |
+
"eval_runtime": 10.8031,
|
| 943 |
+
"eval_samples_per_second": 925.657,
|
| 944 |
+
"eval_steps_per_second": 1.203,
|
| 945 |
+
"step": 351
|
| 946 |
+
},
|
| 947 |
+
{
|
| 948 |
+
"epoch": 118.0,
|
| 949 |
+
"eval_loss": 0.6053263545036316,
|
| 950 |
+
"eval_runtime": 10.8755,
|
| 951 |
+
"eval_samples_per_second": 919.496,
|
| 952 |
+
"eval_steps_per_second": 1.195,
|
| 953 |
+
"step": 354
|
| 954 |
+
},
|
| 955 |
+
{
|
| 956 |
+
"epoch": 119.0,
|
| 957 |
+
"eval_loss": 0.6083167791366577,
|
| 958 |
+
"eval_runtime": 10.861,
|
| 959 |
+
"eval_samples_per_second": 920.722,
|
| 960 |
+
"eval_steps_per_second": 1.197,
|
| 961 |
+
"step": 357
|
| 962 |
+
},
|
| 963 |
+
{
|
| 964 |
+
"epoch": 120.0,
|
| 965 |
+
"eval_loss": 0.607414960861206,
|
| 966 |
+
"eval_runtime": 10.7885,
|
| 967 |
+
"eval_samples_per_second": 926.909,
|
| 968 |
+
"eval_steps_per_second": 1.205,
|
| 969 |
+
"step": 360
|
| 970 |
+
},
|
| 971 |
+
{
|
| 972 |
+
"epoch": 121.0,
|
| 973 |
+
"eval_loss": 0.6041896343231201,
|
| 974 |
+
"eval_runtime": 10.8677,
|
| 975 |
+
"eval_samples_per_second": 920.161,
|
| 976 |
+
"eval_steps_per_second": 1.196,
|
| 977 |
+
"step": 363
|
| 978 |
+
},
|
| 979 |
+
{
|
| 980 |
+
"epoch": 122.0,
|
| 981 |
+
"eval_loss": 0.5942515134811401,
|
| 982 |
+
"eval_runtime": 10.786,
|
| 983 |
+
"eval_samples_per_second": 927.126,
|
| 984 |
+
"eval_steps_per_second": 1.205,
|
| 985 |
+
"step": 366
|
| 986 |
+
},
|
| 987 |
+
{
|
| 988 |
+
"epoch": 123.0,
|
| 989 |
+
"eval_loss": 0.5849249958992004,
|
| 990 |
+
"eval_runtime": 10.8776,
|
| 991 |
+
"eval_samples_per_second": 919.318,
|
| 992 |
+
"eval_steps_per_second": 1.195,
|
| 993 |
+
"step": 369
|
| 994 |
+
},
|
| 995 |
+
{
|
| 996 |
+
"epoch": 124.0,
|
| 997 |
+
"eval_loss": 0.5770248770713806,
|
| 998 |
+
"eval_runtime": 10.8665,
|
| 999 |
+
"eval_samples_per_second": 920.261,
|
| 1000 |
+
"eval_steps_per_second": 1.196,
|
| 1001 |
+
"step": 372
|
| 1002 |
+
},
|
| 1003 |
+
{
|
| 1004 |
+
"epoch": 125.0,
|
| 1005 |
+
"eval_loss": 0.5670948028564453,
|
| 1006 |
+
"eval_runtime": 10.7839,
|
| 1007 |
+
"eval_samples_per_second": 927.305,
|
| 1008 |
+
"eval_steps_per_second": 1.205,
|
| 1009 |
+
"step": 375
|
| 1010 |
+
},
|
| 1011 |
+
{
|
| 1012 |
+
"epoch": 126.0,
|
| 1013 |
+
"eval_loss": 0.5637474656105042,
|
| 1014 |
+
"eval_runtime": 10.8663,
|
| 1015 |
+
"eval_samples_per_second": 920.276,
|
| 1016 |
+
"eval_steps_per_second": 1.196,
|
| 1017 |
+
"step": 378
|
| 1018 |
+
},
|
| 1019 |
+
{
|
| 1020 |
+
"epoch": 127.0,
|
| 1021 |
+
"eval_loss": 0.5640723705291748,
|
| 1022 |
+
"eval_runtime": 10.7903,
|
| 1023 |
+
"eval_samples_per_second": 926.754,
|
| 1024 |
+
"eval_steps_per_second": 1.205,
|
| 1025 |
+
"step": 381
|
| 1026 |
+
},
|
| 1027 |
+
{
|
| 1028 |
+
"epoch": 128.0,
|
| 1029 |
+
"eval_loss": 0.5665525197982788,
|
| 1030 |
+
"eval_runtime": 10.8832,
|
| 1031 |
+
"eval_samples_per_second": 918.844,
|
| 1032 |
+
"eval_steps_per_second": 1.194,
|
| 1033 |
+
"step": 384
|
| 1034 |
+
},
|
| 1035 |
+
{
|
| 1036 |
+
"epoch": 129.0,
|
| 1037 |
+
"eval_loss": 0.5674743056297302,
|
| 1038 |
+
"eval_runtime": 10.8693,
|
| 1039 |
+
"eval_samples_per_second": 920.019,
|
| 1040 |
+
"eval_steps_per_second": 1.196,
|
| 1041 |
+
"step": 387
|
| 1042 |
+
},
|
| 1043 |
+
{
|
| 1044 |
+
"epoch": 130.0,
|
| 1045 |
+
"eval_loss": 0.558768630027771,
|
| 1046 |
+
"eval_runtime": 10.7538,
|
| 1047 |
+
"eval_samples_per_second": 929.908,
|
| 1048 |
+
"eval_steps_per_second": 1.209,
|
| 1049 |
+
"step": 390
|
| 1050 |
+
},
|
| 1051 |
+
{
|
| 1052 |
+
"epoch": 131.0,
|
| 1053 |
+
"eval_loss": 0.5466377139091492,
|
| 1054 |
+
"eval_runtime": 10.8966,
|
| 1055 |
+
"eval_samples_per_second": 917.721,
|
| 1056 |
+
"eval_steps_per_second": 1.193,
|
| 1057 |
+
"step": 393
|
| 1058 |
+
},
|
| 1059 |
+
{
|
| 1060 |
+
"epoch": 132.0,
|
| 1061 |
+
"eval_loss": 0.5402641296386719,
|
| 1062 |
+
"eval_runtime": 10.8007,
|
| 1063 |
+
"eval_samples_per_second": 925.864,
|
| 1064 |
+
"eval_steps_per_second": 1.204,
|
| 1065 |
+
"step": 396
|
| 1066 |
+
},
|
| 1067 |
+
{
|
| 1068 |
+
"epoch": 133.0,
|
| 1069 |
+
"eval_loss": 0.540850043296814,
|
| 1070 |
+
"eval_runtime": 10.8814,
|
| 1071 |
+
"eval_samples_per_second": 918.996,
|
| 1072 |
+
"eval_steps_per_second": 1.195,
|
| 1073 |
+
"step": 399
|
| 1074 |
+
},
|
| 1075 |
+
{
|
| 1076 |
+
"epoch": 134.0,
|
| 1077 |
+
"eval_loss": 0.5402743220329285,
|
| 1078 |
+
"eval_runtime": 10.8778,
|
| 1079 |
+
"eval_samples_per_second": 919.303,
|
| 1080 |
+
"eval_steps_per_second": 1.195,
|
| 1081 |
+
"step": 402
|
| 1082 |
+
},
|
| 1083 |
+
{
|
| 1084 |
+
"epoch": 135.0,
|
| 1085 |
+
"eval_loss": 0.5400083065032959,
|
| 1086 |
+
"eval_runtime": 10.7631,
|
| 1087 |
+
"eval_samples_per_second": 929.098,
|
| 1088 |
+
"eval_steps_per_second": 1.208,
|
| 1089 |
+
"step": 405
|
| 1090 |
+
},
|
| 1091 |
+
{
|
| 1092 |
+
"epoch": 136.0,
|
| 1093 |
+
"eval_loss": 0.5397944450378418,
|
| 1094 |
+
"eval_runtime": 10.8575,
|
| 1095 |
+
"eval_samples_per_second": 921.018,
|
| 1096 |
+
"eval_steps_per_second": 1.197,
|
| 1097 |
+
"step": 408
|
| 1098 |
+
},
|
| 1099 |
+
{
|
| 1100 |
+
"epoch": 137.0,
|
| 1101 |
+
"eval_loss": 0.5391473770141602,
|
| 1102 |
+
"eval_runtime": 10.8073,
|
| 1103 |
+
"eval_samples_per_second": 925.299,
|
| 1104 |
+
"eval_steps_per_second": 1.203,
|
| 1105 |
+
"step": 411
|
| 1106 |
+
},
|
| 1107 |
+
{
|
| 1108 |
+
"epoch": 138.0,
|
| 1109 |
+
"eval_loss": 0.5366548895835876,
|
| 1110 |
+
"eval_runtime": 10.8743,
|
| 1111 |
+
"eval_samples_per_second": 919.601,
|
| 1112 |
+
"eval_steps_per_second": 1.195,
|
| 1113 |
+
"step": 414
|
| 1114 |
+
},
|
| 1115 |
+
{
|
| 1116 |
+
"epoch": 139.0,
|
| 1117 |
+
"eval_loss": 0.5325959920883179,
|
| 1118 |
+
"eval_runtime": 10.8603,
|
| 1119 |
+
"eval_samples_per_second": 920.781,
|
| 1120 |
+
"eval_steps_per_second": 1.197,
|
| 1121 |
+
"step": 417
|
| 1122 |
+
},
|
| 1123 |
+
{
|
| 1124 |
+
"epoch": 140.0,
|
| 1125 |
+
"eval_loss": 0.5299940705299377,
|
| 1126 |
+
"eval_runtime": 10.7821,
|
| 1127 |
+
"eval_samples_per_second": 927.467,
|
| 1128 |
+
"eval_steps_per_second": 1.206,
|
| 1129 |
+
"step": 420
|
| 1130 |
+
},
|
| 1131 |
+
{
|
| 1132 |
+
"epoch": 141.0,
|
| 1133 |
+
"eval_loss": 0.5305985808372498,
|
| 1134 |
+
"eval_runtime": 10.8496,
|
| 1135 |
+
"eval_samples_per_second": 921.697,
|
| 1136 |
+
"eval_steps_per_second": 1.198,
|
| 1137 |
+
"step": 423
|
| 1138 |
+
},
|
| 1139 |
+
{
|
| 1140 |
+
"epoch": 142.0,
|
| 1141 |
+
"eval_loss": 0.5292918682098389,
|
| 1142 |
+
"eval_runtime": 10.7928,
|
| 1143 |
+
"eval_samples_per_second": 926.542,
|
| 1144 |
+
"eval_steps_per_second": 1.205,
|
| 1145 |
+
"step": 426
|
| 1146 |
+
},
|
| 1147 |
+
{
|
| 1148 |
+
"epoch": 143.0,
|
| 1149 |
+
"eval_loss": 0.5267909169197083,
|
| 1150 |
+
"eval_runtime": 10.8928,
|
| 1151 |
+
"eval_samples_per_second": 918.04,
|
| 1152 |
+
"eval_steps_per_second": 1.193,
|
| 1153 |
+
"step": 429
|
| 1154 |
+
},
|
| 1155 |
+
{
|
| 1156 |
+
"epoch": 144.0,
|
| 1157 |
+
"eval_loss": 0.523544430732727,
|
| 1158 |
+
"eval_runtime": 10.8792,
|
| 1159 |
+
"eval_samples_per_second": 919.184,
|
| 1160 |
+
"eval_steps_per_second": 1.195,
|
| 1161 |
+
"step": 432
|
| 1162 |
+
},
|
| 1163 |
+
{
|
| 1164 |
+
"epoch": 145.0,
|
| 1165 |
+
"eval_loss": 0.5205994248390198,
|
| 1166 |
+
"eval_runtime": 10.7887,
|
| 1167 |
+
"eval_samples_per_second": 926.892,
|
| 1168 |
+
"eval_steps_per_second": 1.205,
|
| 1169 |
+
"step": 435
|
| 1170 |
+
},
|
| 1171 |
+
{
|
| 1172 |
+
"epoch": 146.0,
|
| 1173 |
+
"eval_loss": 0.519413411617279,
|
| 1174 |
+
"eval_runtime": 10.8656,
|
| 1175 |
+
"eval_samples_per_second": 920.336,
|
| 1176 |
+
"eval_steps_per_second": 1.196,
|
| 1177 |
+
"step": 438
|
| 1178 |
+
},
|
| 1179 |
+
{
|
| 1180 |
+
"epoch": 147.0,
|
| 1181 |
+
"eval_loss": 0.519058108329773,
|
| 1182 |
+
"eval_runtime": 10.7991,
|
| 1183 |
+
"eval_samples_per_second": 926.005,
|
| 1184 |
+
"eval_steps_per_second": 1.204,
|
| 1185 |
+
"step": 441
|
| 1186 |
+
},
|
| 1187 |
+
{
|
| 1188 |
+
"epoch": 148.0,
|
| 1189 |
+
"eval_loss": 0.5188890099525452,
|
| 1190 |
+
"eval_runtime": 10.8757,
|
| 1191 |
+
"eval_samples_per_second": 919.48,
|
| 1192 |
+
"eval_steps_per_second": 1.195,
|
| 1193 |
+
"step": 444
|
| 1194 |
+
},
|
| 1195 |
+
{
|
| 1196 |
+
"epoch": 149.0,
|
| 1197 |
+
"eval_loss": 0.5188961625099182,
|
| 1198 |
+
"eval_runtime": 10.8633,
|
| 1199 |
+
"eval_samples_per_second": 920.533,
|
| 1200 |
+
"eval_steps_per_second": 1.197,
|
| 1201 |
+
"step": 447
|
| 1202 |
+
},
|
| 1203 |
+
{
|
| 1204 |
+
"epoch": 150.0,
|
| 1205 |
+
"eval_loss": 0.5188751220703125,
|
| 1206 |
+
"eval_runtime": 10.8003,
|
| 1207 |
+
"eval_samples_per_second": 925.904,
|
| 1208 |
+
"eval_steps_per_second": 1.204,
|
| 1209 |
+
"step": 450
|
| 1210 |
+
}
|
| 1211 |
+
],
|
| 1212 |
+
"logging_steps": 500,
|
| 1213 |
+
"max_steps": 450,
|
| 1214 |
+
"num_train_epochs": 150,
|
| 1215 |
+
"save_steps": 500,
|
| 1216 |
+
"total_flos": 8613277286400000.0,
|
| 1217 |
+
"trial_name": null,
|
| 1218 |
+
"trial_params": null
|
| 1219 |
+
}
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:b7d1481f2fa08349db1651d00c965505bd70a011714273a45d642a0b3eac76b5
|
| 3 |
+
size 4792
|