Upload folder using huggingface_hub
Browse files- README.md +27 -0
- added_tokens.json +3 -0
- checkpoint-5193/config.json +42 -0
- checkpoint-5193/model.safetensors +3 -0
- checkpoint-5193/optimizer.pt +3 -0
- checkpoint-5193/rng_state_0.pth +3 -0
- checkpoint-5193/rng_state_1.pth +3 -0
- checkpoint-5193/rng_state_2.pth +3 -0
- checkpoint-5193/rng_state_3.pth +3 -0
- checkpoint-5193/scheduler.pt +3 -0
- checkpoint-5193/trainer_state.json +1509 -0
- checkpoint-5193/training_args.bin +3 -0
- config.json +42 -0
- model.safetensors +3 -0
- runs/Apr29_14-53-13_r-abhishek-autotrain-1bfjaa9n-00d8c-hkw4p/events.out.tfevents.1714402394.r-abhishek-autotrain-1bfjaa9n-00d8c-hkw4p.464.0 +2 -2
- runs/Apr29_14-53-13_r-abhishek-autotrain-1bfjaa9n-00d8c-hkw4p/events.out.tfevents.1714407850.r-abhishek-autotrain-1bfjaa9n-00d8c-hkw4p.464.1 +3 -0
- special_tokens_map.json +15 -0
- spm.model +3 -0
- tokenizer.json +0 -0
- tokenizer_config.json +58 -0
- training_args.bin +3 -0
- training_params.json +28 -0
README.md
ADDED
|
@@ -0,0 +1,27 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
|
| 2 |
+
---
|
| 3 |
+
tags:
|
| 4 |
+
- autotrain
|
| 5 |
+
- text-regression
|
| 6 |
+
widget:
|
| 7 |
+
- text: "I love AutoTrain"
|
| 8 |
+
datasets:
|
| 9 |
+
- autotrain-m96nh-snymb/autotrain-data
|
| 10 |
+
---
|
| 11 |
+
|
| 12 |
+
# Model Trained Using AutoTrain
|
| 13 |
+
|
| 14 |
+
- Problem type: Text Regression
|
| 15 |
+
|
| 16 |
+
## Validation Metrics
|
| 17 |
+
loss: 0.282262921333313
|
| 18 |
+
|
| 19 |
+
mse: 0.2820460796356201
|
| 20 |
+
|
| 21 |
+
mae: 0.4189736545085907
|
| 22 |
+
|
| 23 |
+
r2: 0.74436353679844
|
| 24 |
+
|
| 25 |
+
rmse: 0.5310801267623901
|
| 26 |
+
|
| 27 |
+
explained_variance: 0.7570163011550903
|
added_tokens.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"[MASK]": 128000
|
| 3 |
+
}
|
checkpoint-5193/config.json
ADDED
|
@@ -0,0 +1,42 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "microsoft/deberta-v3-large",
|
| 3 |
+
"_num_labels": 1,
|
| 4 |
+
"architectures": [
|
| 5 |
+
"DebertaV2ForSequenceClassification"
|
| 6 |
+
],
|
| 7 |
+
"attention_probs_dropout_prob": 0.1,
|
| 8 |
+
"hidden_act": "gelu",
|
| 9 |
+
"hidden_dropout_prob": 0.1,
|
| 10 |
+
"hidden_size": 1024,
|
| 11 |
+
"id2label": {
|
| 12 |
+
"0": "target"
|
| 13 |
+
},
|
| 14 |
+
"initializer_range": 0.02,
|
| 15 |
+
"intermediate_size": 4096,
|
| 16 |
+
"label2id": {
|
| 17 |
+
"target": 0
|
| 18 |
+
},
|
| 19 |
+
"layer_norm_eps": 1e-07,
|
| 20 |
+
"max_position_embeddings": 512,
|
| 21 |
+
"max_relative_positions": -1,
|
| 22 |
+
"model_type": "deberta-v2",
|
| 23 |
+
"norm_rel_ebd": "layer_norm",
|
| 24 |
+
"num_attention_heads": 16,
|
| 25 |
+
"num_hidden_layers": 24,
|
| 26 |
+
"pad_token_id": 0,
|
| 27 |
+
"pooler_dropout": 0,
|
| 28 |
+
"pooler_hidden_act": "gelu",
|
| 29 |
+
"pooler_hidden_size": 1024,
|
| 30 |
+
"pos_att_type": [
|
| 31 |
+
"p2c",
|
| 32 |
+
"c2p"
|
| 33 |
+
],
|
| 34 |
+
"position_biased_input": false,
|
| 35 |
+
"position_buckets": 256,
|
| 36 |
+
"relative_attention": true,
|
| 37 |
+
"share_att_key": true,
|
| 38 |
+
"torch_dtype": "float32",
|
| 39 |
+
"transformers_version": "4.40.1",
|
| 40 |
+
"type_vocab_size": 0,
|
| 41 |
+
"vocab_size": 128100
|
| 42 |
+
}
|
checkpoint-5193/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:3d68d3238d7e01448edfe7b1983dcf19cc58c6bfbb7628f33a97a5690b6e914d
|
| 3 |
+
size 1740300340
|
checkpoint-5193/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:6f82c9fa96d35c5b639fda51433bc915f1712871b20849a7d91b7c30ad3c8545
|
| 3 |
+
size 3480832048
|
checkpoint-5193/rng_state_0.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:9692e0df6d5b78451d0faa94b4317f3c3cd1a1bdea676ec7954514c02b9cfcdc
|
| 3 |
+
size 15024
|
checkpoint-5193/rng_state_1.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e60d67b3848a732be46492193f28d5c098caad9f51a52ca97768faf56dca6421
|
| 3 |
+
size 15024
|
checkpoint-5193/rng_state_2.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:02d7e3cf0198b07ae0e4e7c0e7b2fd5824805aaf7c0ec58feda8b2ef7a83c7ef
|
| 3 |
+
size 15024
|
checkpoint-5193/rng_state_3.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:b5c52c0b9880b9a6b76baf795175159ea7ee4f04a9d7ffb3cd483a38f6311db7
|
| 3 |
+
size 15024
|
checkpoint-5193/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:5ccb544db0a45b77c836df5bb36152996c97ff1a522be11d72b179afc3d5ba75
|
| 3 |
+
size 1064
|
checkpoint-5193/trainer_state.json
ADDED
|
@@ -0,0 +1,1509 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.282262921333313,
|
| 3 |
+
"best_model_checkpoint": "autotrain-m96nh-snymb/checkpoint-5193",
|
| 4 |
+
"epoch": 3.0,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 5193,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.014442518775274409,
|
| 13 |
+
"grad_norm": 212.5823516845703,
|
| 14 |
+
"learning_rate": 7.692307692307692e-08,
|
| 15 |
+
"loss": 10.235,
|
| 16 |
+
"step": 25
|
| 17 |
+
},
|
| 18 |
+
{
|
| 19 |
+
"epoch": 0.028885037550548817,
|
| 20 |
+
"grad_norm": 81.24781799316406,
|
| 21 |
+
"learning_rate": 1.7307692307692305e-07,
|
| 22 |
+
"loss": 10.1291,
|
| 23 |
+
"step": 50
|
| 24 |
+
},
|
| 25 |
+
{
|
| 26 |
+
"epoch": 0.043327556325823226,
|
| 27 |
+
"grad_norm": 302.68389892578125,
|
| 28 |
+
"learning_rate": 2.692307692307692e-07,
|
| 29 |
+
"loss": 9.6898,
|
| 30 |
+
"step": 75
|
| 31 |
+
},
|
| 32 |
+
{
|
| 33 |
+
"epoch": 0.057770075101097634,
|
| 34 |
+
"grad_norm": 327.724853515625,
|
| 35 |
+
"learning_rate": 3.615384615384615e-07,
|
| 36 |
+
"loss": 10.4787,
|
| 37 |
+
"step": 100
|
| 38 |
+
},
|
| 39 |
+
{
|
| 40 |
+
"epoch": 0.07221259387637204,
|
| 41 |
+
"grad_norm": 135.6590576171875,
|
| 42 |
+
"learning_rate": 4.538461538461538e-07,
|
| 43 |
+
"loss": 9.5916,
|
| 44 |
+
"step": 125
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.08665511265164645,
|
| 48 |
+
"grad_norm": 108.32626342773438,
|
| 49 |
+
"learning_rate": 5.5e-07,
|
| 50 |
+
"loss": 8.8043,
|
| 51 |
+
"step": 150
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.10109763142692085,
|
| 55 |
+
"grad_norm": 73.20294189453125,
|
| 56 |
+
"learning_rate": 6.461538461538462e-07,
|
| 57 |
+
"loss": 7.1787,
|
| 58 |
+
"step": 175
|
| 59 |
+
},
|
| 60 |
+
{
|
| 61 |
+
"epoch": 0.11554015020219527,
|
| 62 |
+
"grad_norm": 73.47607421875,
|
| 63 |
+
"learning_rate": 7.423076923076923e-07,
|
| 64 |
+
"loss": 5.024,
|
| 65 |
+
"step": 200
|
| 66 |
+
},
|
| 67 |
+
{
|
| 68 |
+
"epoch": 0.12998266897746968,
|
| 69 |
+
"grad_norm": 73.16763305664062,
|
| 70 |
+
"learning_rate": 8.384615384615384e-07,
|
| 71 |
+
"loss": 2.586,
|
| 72 |
+
"step": 225
|
| 73 |
+
},
|
| 74 |
+
{
|
| 75 |
+
"epoch": 0.14442518775274407,
|
| 76 |
+
"grad_norm": 42.77627182006836,
|
| 77 |
+
"learning_rate": 9.346153846153846e-07,
|
| 78 |
+
"loss": 0.9176,
|
| 79 |
+
"step": 250
|
| 80 |
+
},
|
| 81 |
+
{
|
| 82 |
+
"epoch": 0.1588677065280185,
|
| 83 |
+
"grad_norm": 73.87055969238281,
|
| 84 |
+
"learning_rate": 1.0307692307692306e-06,
|
| 85 |
+
"loss": 1.0465,
|
| 86 |
+
"step": 275
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.1733102253032929,
|
| 90 |
+
"grad_norm": 29.49479103088379,
|
| 91 |
+
"learning_rate": 1.1269230769230768e-06,
|
| 92 |
+
"loss": 0.6761,
|
| 93 |
+
"step": 300
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 0.1877527440785673,
|
| 97 |
+
"grad_norm": 78.30584716796875,
|
| 98 |
+
"learning_rate": 1.2230769230769231e-06,
|
| 99 |
+
"loss": 0.5363,
|
| 100 |
+
"step": 325
|
| 101 |
+
},
|
| 102 |
+
{
|
| 103 |
+
"epoch": 0.2021952628538417,
|
| 104 |
+
"grad_norm": 23.788848876953125,
|
| 105 |
+
"learning_rate": 1.3192307692307692e-06,
|
| 106 |
+
"loss": 0.437,
|
| 107 |
+
"step": 350
|
| 108 |
+
},
|
| 109 |
+
{
|
| 110 |
+
"epoch": 0.21663778162911612,
|
| 111 |
+
"grad_norm": 5.925709247589111,
|
| 112 |
+
"learning_rate": 1.4153846153846155e-06,
|
| 113 |
+
"loss": 0.5172,
|
| 114 |
+
"step": 375
|
| 115 |
+
},
|
| 116 |
+
{
|
| 117 |
+
"epoch": 0.23108030040439054,
|
| 118 |
+
"grad_norm": 32.11520767211914,
|
| 119 |
+
"learning_rate": 1.5115384615384615e-06,
|
| 120 |
+
"loss": 0.509,
|
| 121 |
+
"step": 400
|
| 122 |
+
},
|
| 123 |
+
{
|
| 124 |
+
"epoch": 0.24552281917966493,
|
| 125 |
+
"grad_norm": 14.887007713317871,
|
| 126 |
+
"learning_rate": 1.6076923076923076e-06,
|
| 127 |
+
"loss": 0.6094,
|
| 128 |
+
"step": 425
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.25996533795493937,
|
| 132 |
+
"grad_norm": 31.848352432250977,
|
| 133 |
+
"learning_rate": 1.7038461538461536e-06,
|
| 134 |
+
"loss": 0.5056,
|
| 135 |
+
"step": 450
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 0.27440785673021373,
|
| 139 |
+
"grad_norm": 72.35002136230469,
|
| 140 |
+
"learning_rate": 1.8e-06,
|
| 141 |
+
"loss": 0.4752,
|
| 142 |
+
"step": 475
|
| 143 |
+
},
|
| 144 |
+
{
|
| 145 |
+
"epoch": 0.28885037550548814,
|
| 146 |
+
"grad_norm": 12.196418762207031,
|
| 147 |
+
"learning_rate": 1.896153846153846e-06,
|
| 148 |
+
"loss": 0.375,
|
| 149 |
+
"step": 500
|
| 150 |
+
},
|
| 151 |
+
{
|
| 152 |
+
"epoch": 0.30329289428076256,
|
| 153 |
+
"grad_norm": 25.701213836669922,
|
| 154 |
+
"learning_rate": 1.9923076923076923e-06,
|
| 155 |
+
"loss": 0.4961,
|
| 156 |
+
"step": 525
|
| 157 |
+
},
|
| 158 |
+
{
|
| 159 |
+
"epoch": 0.317735413056037,
|
| 160 |
+
"grad_norm": 41.281166076660156,
|
| 161 |
+
"learning_rate": 1.9901562165632354e-06,
|
| 162 |
+
"loss": 0.4455,
|
| 163 |
+
"step": 550
|
| 164 |
+
},
|
| 165 |
+
{
|
| 166 |
+
"epoch": 0.3321779318313114,
|
| 167 |
+
"grad_norm": 19.428579330444336,
|
| 168 |
+
"learning_rate": 1.979456451958057e-06,
|
| 169 |
+
"loss": 0.6306,
|
| 170 |
+
"step": 575
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 0.3466204506065858,
|
| 174 |
+
"grad_norm": 18.564756393432617,
|
| 175 |
+
"learning_rate": 1.9687566873528784e-06,
|
| 176 |
+
"loss": 0.399,
|
| 177 |
+
"step": 600
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 0.3610629693818602,
|
| 181 |
+
"grad_norm": 41.66643524169922,
|
| 182 |
+
"learning_rate": 1.9580569227476994e-06,
|
| 183 |
+
"loss": 0.4581,
|
| 184 |
+
"step": 625
|
| 185 |
+
},
|
| 186 |
+
{
|
| 187 |
+
"epoch": 0.3755054881571346,
|
| 188 |
+
"grad_norm": 41.7965202331543,
|
| 189 |
+
"learning_rate": 1.947357158142521e-06,
|
| 190 |
+
"loss": 0.4799,
|
| 191 |
+
"step": 650
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"epoch": 0.389948006932409,
|
| 195 |
+
"grad_norm": 9.624334335327148,
|
| 196 |
+
"learning_rate": 1.936657393537342e-06,
|
| 197 |
+
"loss": 0.4632,
|
| 198 |
+
"step": 675
|
| 199 |
+
},
|
| 200 |
+
{
|
| 201 |
+
"epoch": 0.4043905257076834,
|
| 202 |
+
"grad_norm": 11.227400779724121,
|
| 203 |
+
"learning_rate": 1.9259576289321635e-06,
|
| 204 |
+
"loss": 0.3521,
|
| 205 |
+
"step": 700
|
| 206 |
+
},
|
| 207 |
+
{
|
| 208 |
+
"epoch": 0.41883304448295783,
|
| 209 |
+
"grad_norm": 22.89958953857422,
|
| 210 |
+
"learning_rate": 1.915257864326985e-06,
|
| 211 |
+
"loss": 0.4414,
|
| 212 |
+
"step": 725
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 0.43327556325823224,
|
| 216 |
+
"grad_norm": 26.13202476501465,
|
| 217 |
+
"learning_rate": 1.904558099721806e-06,
|
| 218 |
+
"loss": 0.3836,
|
| 219 |
+
"step": 750
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 0.44771808203350666,
|
| 223 |
+
"grad_norm": 55.14106750488281,
|
| 224 |
+
"learning_rate": 1.8938583351166273e-06,
|
| 225 |
+
"loss": 0.5554,
|
| 226 |
+
"step": 775
|
| 227 |
+
},
|
| 228 |
+
{
|
| 229 |
+
"epoch": 0.4621606008087811,
|
| 230 |
+
"grad_norm": 32.7313346862793,
|
| 231 |
+
"learning_rate": 1.8831585705114485e-06,
|
| 232 |
+
"loss": 0.4613,
|
| 233 |
+
"step": 800
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"epoch": 0.47660311958405543,
|
| 237 |
+
"grad_norm": 14.63770866394043,
|
| 238 |
+
"learning_rate": 1.87245880590627e-06,
|
| 239 |
+
"loss": 0.3672,
|
| 240 |
+
"step": 825
|
| 241 |
+
},
|
| 242 |
+
{
|
| 243 |
+
"epoch": 0.49104563835932985,
|
| 244 |
+
"grad_norm": 39.93565368652344,
|
| 245 |
+
"learning_rate": 1.8617590413010913e-06,
|
| 246 |
+
"loss": 0.495,
|
| 247 |
+
"step": 850
|
| 248 |
+
},
|
| 249 |
+
{
|
| 250 |
+
"epoch": 0.5054881571346043,
|
| 251 |
+
"grad_norm": 13.94717025756836,
|
| 252 |
+
"learning_rate": 1.8510592766959126e-06,
|
| 253 |
+
"loss": 0.3883,
|
| 254 |
+
"step": 875
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 0.5199306759098787,
|
| 258 |
+
"grad_norm": 19.33357810974121,
|
| 259 |
+
"learning_rate": 1.8403595120907338e-06,
|
| 260 |
+
"loss": 0.3607,
|
| 261 |
+
"step": 900
|
| 262 |
+
},
|
| 263 |
+
{
|
| 264 |
+
"epoch": 0.5343731946851531,
|
| 265 |
+
"grad_norm": 9.50666618347168,
|
| 266 |
+
"learning_rate": 1.8296597474855553e-06,
|
| 267 |
+
"loss": 0.4309,
|
| 268 |
+
"step": 925
|
| 269 |
+
},
|
| 270 |
+
{
|
| 271 |
+
"epoch": 0.5488157134604275,
|
| 272 |
+
"grad_norm": 99.26518249511719,
|
| 273 |
+
"learning_rate": 1.8189599828803766e-06,
|
| 274 |
+
"loss": 0.4734,
|
| 275 |
+
"step": 950
|
| 276 |
+
},
|
| 277 |
+
{
|
| 278 |
+
"epoch": 0.5632582322357019,
|
| 279 |
+
"grad_norm": 25.336233139038086,
|
| 280 |
+
"learning_rate": 1.8082602182751978e-06,
|
| 281 |
+
"loss": 0.5486,
|
| 282 |
+
"step": 975
|
| 283 |
+
},
|
| 284 |
+
{
|
| 285 |
+
"epoch": 0.5777007510109763,
|
| 286 |
+
"grad_norm": 41.74341583251953,
|
| 287 |
+
"learning_rate": 1.7975604536700191e-06,
|
| 288 |
+
"loss": 0.378,
|
| 289 |
+
"step": 1000
|
| 290 |
+
},
|
| 291 |
+
{
|
| 292 |
+
"epoch": 0.5921432697862508,
|
| 293 |
+
"grad_norm": 37.5980224609375,
|
| 294 |
+
"learning_rate": 1.7868606890648406e-06,
|
| 295 |
+
"loss": 0.3274,
|
| 296 |
+
"step": 1025
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 0.6065857885615251,
|
| 300 |
+
"grad_norm": 46.92686080932617,
|
| 301 |
+
"learning_rate": 1.7761609244596619e-06,
|
| 302 |
+
"loss": 0.3263,
|
| 303 |
+
"step": 1050
|
| 304 |
+
},
|
| 305 |
+
{
|
| 306 |
+
"epoch": 0.6210283073367996,
|
| 307 |
+
"grad_norm": 20.776403427124023,
|
| 308 |
+
"learning_rate": 1.7654611598544831e-06,
|
| 309 |
+
"loss": 0.3563,
|
| 310 |
+
"step": 1075
|
| 311 |
+
},
|
| 312 |
+
{
|
| 313 |
+
"epoch": 0.635470826112074,
|
| 314 |
+
"grad_norm": 26.412818908691406,
|
| 315 |
+
"learning_rate": 1.7547613952493044e-06,
|
| 316 |
+
"loss": 0.354,
|
| 317 |
+
"step": 1100
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 0.6499133448873483,
|
| 321 |
+
"grad_norm": 35.234580993652344,
|
| 322 |
+
"learning_rate": 1.7440616306441259e-06,
|
| 323 |
+
"loss": 0.4225,
|
| 324 |
+
"step": 1125
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 0.6643558636626228,
|
| 328 |
+
"grad_norm": 38.81793212890625,
|
| 329 |
+
"learning_rate": 1.7333618660389472e-06,
|
| 330 |
+
"loss": 0.3189,
|
| 331 |
+
"step": 1150
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 0.6787983824378971,
|
| 335 |
+
"grad_norm": 26.322595596313477,
|
| 336 |
+
"learning_rate": 1.7226621014337684e-06,
|
| 337 |
+
"loss": 0.3566,
|
| 338 |
+
"step": 1175
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 0.6932409012131716,
|
| 342 |
+
"grad_norm": 19.16035270690918,
|
| 343 |
+
"learning_rate": 1.7119623368285897e-06,
|
| 344 |
+
"loss": 0.3212,
|
| 345 |
+
"step": 1200
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 0.707683419988446,
|
| 349 |
+
"grad_norm": 26.820486068725586,
|
| 350 |
+
"learning_rate": 1.7012625722234112e-06,
|
| 351 |
+
"loss": 0.3318,
|
| 352 |
+
"step": 1225
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 0.7221259387637204,
|
| 356 |
+
"grad_norm": 9.583172798156738,
|
| 357 |
+
"learning_rate": 1.6905628076182322e-06,
|
| 358 |
+
"loss": 0.3306,
|
| 359 |
+
"step": 1250
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 0.7365684575389948,
|
| 363 |
+
"grad_norm": 67.81623840332031,
|
| 364 |
+
"learning_rate": 1.6798630430130535e-06,
|
| 365 |
+
"loss": 0.3503,
|
| 366 |
+
"step": 1275
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 0.7510109763142692,
|
| 370 |
+
"grad_norm": 26.901622772216797,
|
| 371 |
+
"learning_rate": 1.6691632784078748e-06,
|
| 372 |
+
"loss": 0.3112,
|
| 373 |
+
"step": 1300
|
| 374 |
+
},
|
| 375 |
+
{
|
| 376 |
+
"epoch": 0.7654534950895436,
|
| 377 |
+
"grad_norm": 19.61265754699707,
|
| 378 |
+
"learning_rate": 1.6584635138026963e-06,
|
| 379 |
+
"loss": 0.3354,
|
| 380 |
+
"step": 1325
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 0.779896013864818,
|
| 384 |
+
"grad_norm": 26.634790420532227,
|
| 385 |
+
"learning_rate": 1.6477637491975175e-06,
|
| 386 |
+
"loss": 0.4247,
|
| 387 |
+
"step": 1350
|
| 388 |
+
},
|
| 389 |
+
{
|
| 390 |
+
"epoch": 0.7943385326400925,
|
| 391 |
+
"grad_norm": 12.382301330566406,
|
| 392 |
+
"learning_rate": 1.6370639845923388e-06,
|
| 393 |
+
"loss": 0.3341,
|
| 394 |
+
"step": 1375
|
| 395 |
+
},
|
| 396 |
+
{
|
| 397 |
+
"epoch": 0.8087810514153668,
|
| 398 |
+
"grad_norm": 17.31427574157715,
|
| 399 |
+
"learning_rate": 1.62636421998716e-06,
|
| 400 |
+
"loss": 0.3274,
|
| 401 |
+
"step": 1400
|
| 402 |
+
},
|
| 403 |
+
{
|
| 404 |
+
"epoch": 0.8232235701906413,
|
| 405 |
+
"grad_norm": 11.828612327575684,
|
| 406 |
+
"learning_rate": 1.6156644553819815e-06,
|
| 407 |
+
"loss": 0.3575,
|
| 408 |
+
"step": 1425
|
| 409 |
+
},
|
| 410 |
+
{
|
| 411 |
+
"epoch": 0.8376660889659157,
|
| 412 |
+
"grad_norm": 12.682332038879395,
|
| 413 |
+
"learning_rate": 1.6049646907768028e-06,
|
| 414 |
+
"loss": 0.2944,
|
| 415 |
+
"step": 1450
|
| 416 |
+
},
|
| 417 |
+
{
|
| 418 |
+
"epoch": 0.85210860774119,
|
| 419 |
+
"grad_norm": 14.496199607849121,
|
| 420 |
+
"learning_rate": 1.594264926171624e-06,
|
| 421 |
+
"loss": 0.3447,
|
| 422 |
+
"step": 1475
|
| 423 |
+
},
|
| 424 |
+
{
|
| 425 |
+
"epoch": 0.8665511265164645,
|
| 426 |
+
"grad_norm": 20.546905517578125,
|
| 427 |
+
"learning_rate": 1.5835651615664453e-06,
|
| 428 |
+
"loss": 0.3162,
|
| 429 |
+
"step": 1500
|
| 430 |
+
},
|
| 431 |
+
{
|
| 432 |
+
"epoch": 0.8809936452917388,
|
| 433 |
+
"grad_norm": 12.139562606811523,
|
| 434 |
+
"learning_rate": 1.5728653969612668e-06,
|
| 435 |
+
"loss": 0.3553,
|
| 436 |
+
"step": 1525
|
| 437 |
+
},
|
| 438 |
+
{
|
| 439 |
+
"epoch": 0.8954361640670133,
|
| 440 |
+
"grad_norm": 38.39575958251953,
|
| 441 |
+
"learning_rate": 1.562165632356088e-06,
|
| 442 |
+
"loss": 0.3929,
|
| 443 |
+
"step": 1550
|
| 444 |
+
},
|
| 445 |
+
{
|
| 446 |
+
"epoch": 0.9098786828422877,
|
| 447 |
+
"grad_norm": 9.470243453979492,
|
| 448 |
+
"learning_rate": 1.5514658677509094e-06,
|
| 449 |
+
"loss": 0.2997,
|
| 450 |
+
"step": 1575
|
| 451 |
+
},
|
| 452 |
+
{
|
| 453 |
+
"epoch": 0.9243212016175621,
|
| 454 |
+
"grad_norm": 14.675395965576172,
|
| 455 |
+
"learning_rate": 1.5407661031457306e-06,
|
| 456 |
+
"loss": 0.3275,
|
| 457 |
+
"step": 1600
|
| 458 |
+
},
|
| 459 |
+
{
|
| 460 |
+
"epoch": 0.9387637203928365,
|
| 461 |
+
"grad_norm": 3.3243517875671387,
|
| 462 |
+
"learning_rate": 1.5300663385405521e-06,
|
| 463 |
+
"loss": 0.426,
|
| 464 |
+
"step": 1625
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 0.9532062391681109,
|
| 468 |
+
"grad_norm": 12.924434661865234,
|
| 469 |
+
"learning_rate": 1.5193665739353734e-06,
|
| 470 |
+
"loss": 0.2976,
|
| 471 |
+
"step": 1650
|
| 472 |
+
},
|
| 473 |
+
{
|
| 474 |
+
"epoch": 0.9676487579433853,
|
| 475 |
+
"grad_norm": 46.80762481689453,
|
| 476 |
+
"learning_rate": 1.5086668093301947e-06,
|
| 477 |
+
"loss": 0.3499,
|
| 478 |
+
"step": 1675
|
| 479 |
+
},
|
| 480 |
+
{
|
| 481 |
+
"epoch": 0.9820912767186597,
|
| 482 |
+
"grad_norm": 15.062408447265625,
|
| 483 |
+
"learning_rate": 1.497967044725016e-06,
|
| 484 |
+
"loss": 0.3225,
|
| 485 |
+
"step": 1700
|
| 486 |
+
},
|
| 487 |
+
{
|
| 488 |
+
"epoch": 0.9965337954939342,
|
| 489 |
+
"grad_norm": 22.281009674072266,
|
| 490 |
+
"learning_rate": 1.4872672801198374e-06,
|
| 491 |
+
"loss": 0.4113,
|
| 492 |
+
"step": 1725
|
| 493 |
+
},
|
| 494 |
+
{
|
| 495 |
+
"epoch": 1.0,
|
| 496 |
+
"eval_explained_variance": 0.7363581657409668,
|
| 497 |
+
"eval_loss": 0.3119768798351288,
|
| 498 |
+
"eval_mae": 0.4405648112297058,
|
| 499 |
+
"eval_mse": 0.31180134415626526,
|
| 500 |
+
"eval_r2": 0.7173944470629221,
|
| 501 |
+
"eval_rmse": 0.5583917498588562,
|
| 502 |
+
"eval_runtime": 80.5336,
|
| 503 |
+
"eval_samples_per_second": 42.988,
|
| 504 |
+
"eval_steps_per_second": 2.695,
|
| 505 |
+
"step": 1731
|
| 506 |
+
},
|
| 507 |
+
{
|
| 508 |
+
"epoch": 1.0109763142692085,
|
| 509 |
+
"grad_norm": 27.233137130737305,
|
| 510 |
+
"learning_rate": 1.4765675155146587e-06,
|
| 511 |
+
"loss": 0.3579,
|
| 512 |
+
"step": 1750
|
| 513 |
+
},
|
| 514 |
+
{
|
| 515 |
+
"epoch": 1.025418833044483,
|
| 516 |
+
"grad_norm": 23.662677764892578,
|
| 517 |
+
"learning_rate": 1.46586775090948e-06,
|
| 518 |
+
"loss": 0.3746,
|
| 519 |
+
"step": 1775
|
| 520 |
+
},
|
| 521 |
+
{
|
| 522 |
+
"epoch": 1.0398613518197575,
|
| 523 |
+
"grad_norm": 27.650678634643555,
|
| 524 |
+
"learning_rate": 1.4551679863043012e-06,
|
| 525 |
+
"loss": 0.3723,
|
| 526 |
+
"step": 1800
|
| 527 |
+
},
|
| 528 |
+
{
|
| 529 |
+
"epoch": 1.0543038705950318,
|
| 530 |
+
"grad_norm": 38.274314880371094,
|
| 531 |
+
"learning_rate": 1.4444682216991227e-06,
|
| 532 |
+
"loss": 0.3705,
|
| 533 |
+
"step": 1825
|
| 534 |
+
},
|
| 535 |
+
{
|
| 536 |
+
"epoch": 1.0687463893703062,
|
| 537 |
+
"grad_norm": 43.43634796142578,
|
| 538 |
+
"learning_rate": 1.433768457093944e-06,
|
| 539 |
+
"loss": 0.2841,
|
| 540 |
+
"step": 1850
|
| 541 |
+
},
|
| 542 |
+
{
|
| 543 |
+
"epoch": 1.0831889081455806,
|
| 544 |
+
"grad_norm": 19.07291030883789,
|
| 545 |
+
"learning_rate": 1.4230686924887652e-06,
|
| 546 |
+
"loss": 0.4436,
|
| 547 |
+
"step": 1875
|
| 548 |
+
},
|
| 549 |
+
{
|
| 550 |
+
"epoch": 1.097631426920855,
|
| 551 |
+
"grad_norm": 36.660614013671875,
|
| 552 |
+
"learning_rate": 1.4123689278835865e-06,
|
| 553 |
+
"loss": 0.3031,
|
| 554 |
+
"step": 1900
|
| 555 |
+
},
|
| 556 |
+
{
|
| 557 |
+
"epoch": 1.1120739456961295,
|
| 558 |
+
"grad_norm": 10.8666353225708,
|
| 559 |
+
"learning_rate": 1.401669163278408e-06,
|
| 560 |
+
"loss": 0.2895,
|
| 561 |
+
"step": 1925
|
| 562 |
+
},
|
| 563 |
+
{
|
| 564 |
+
"epoch": 1.1265164644714039,
|
| 565 |
+
"grad_norm": 34.650394439697266,
|
| 566 |
+
"learning_rate": 1.3909693986732293e-06,
|
| 567 |
+
"loss": 0.3506,
|
| 568 |
+
"step": 1950
|
| 569 |
+
},
|
| 570 |
+
{
|
| 571 |
+
"epoch": 1.1409589832466782,
|
| 572 |
+
"grad_norm": 43.247623443603516,
|
| 573 |
+
"learning_rate": 1.3802696340680503e-06,
|
| 574 |
+
"loss": 0.3555,
|
| 575 |
+
"step": 1975
|
| 576 |
+
},
|
| 577 |
+
{
|
| 578 |
+
"epoch": 1.1554015020219526,
|
| 579 |
+
"grad_norm": 18.716602325439453,
|
| 580 |
+
"learning_rate": 1.3695698694628716e-06,
|
| 581 |
+
"loss": 0.3931,
|
| 582 |
+
"step": 2000
|
| 583 |
+
},
|
| 584 |
+
{
|
| 585 |
+
"epoch": 1.169844020797227,
|
| 586 |
+
"grad_norm": 31.356761932373047,
|
| 587 |
+
"learning_rate": 1.358870104857693e-06,
|
| 588 |
+
"loss": 0.2976,
|
| 589 |
+
"step": 2025
|
| 590 |
+
},
|
| 591 |
+
{
|
| 592 |
+
"epoch": 1.1842865395725015,
|
| 593 |
+
"grad_norm": 18.609111785888672,
|
| 594 |
+
"learning_rate": 1.3481703402525143e-06,
|
| 595 |
+
"loss": 0.3163,
|
| 596 |
+
"step": 2050
|
| 597 |
+
},
|
| 598 |
+
{
|
| 599 |
+
"epoch": 1.1987290583477759,
|
| 600 |
+
"grad_norm": 31.023008346557617,
|
| 601 |
+
"learning_rate": 1.3374705756473356e-06,
|
| 602 |
+
"loss": 0.3454,
|
| 603 |
+
"step": 2075
|
| 604 |
+
},
|
| 605 |
+
{
|
| 606 |
+
"epoch": 1.2131715771230502,
|
| 607 |
+
"grad_norm": 27.93479347229004,
|
| 608 |
+
"learning_rate": 1.3267708110421569e-06,
|
| 609 |
+
"loss": 0.3452,
|
| 610 |
+
"step": 2100
|
| 611 |
+
},
|
| 612 |
+
{
|
| 613 |
+
"epoch": 1.2276140958983246,
|
| 614 |
+
"grad_norm": 23.254547119140625,
|
| 615 |
+
"learning_rate": 1.3160710464369784e-06,
|
| 616 |
+
"loss": 0.3486,
|
| 617 |
+
"step": 2125
|
| 618 |
+
},
|
| 619 |
+
{
|
| 620 |
+
"epoch": 1.242056614673599,
|
| 621 |
+
"grad_norm": 45.776458740234375,
|
| 622 |
+
"learning_rate": 1.3053712818317996e-06,
|
| 623 |
+
"loss": 0.3586,
|
| 624 |
+
"step": 2150
|
| 625 |
+
},
|
| 626 |
+
{
|
| 627 |
+
"epoch": 1.2564991334488735,
|
| 628 |
+
"grad_norm": 14.92525863647461,
|
| 629 |
+
"learning_rate": 1.294671517226621e-06,
|
| 630 |
+
"loss": 0.3338,
|
| 631 |
+
"step": 2175
|
| 632 |
+
},
|
| 633 |
+
{
|
| 634 |
+
"epoch": 1.270941652224148,
|
| 635 |
+
"grad_norm": 20.12270736694336,
|
| 636 |
+
"learning_rate": 1.2839717526214422e-06,
|
| 637 |
+
"loss": 0.3437,
|
| 638 |
+
"step": 2200
|
| 639 |
+
},
|
| 640 |
+
{
|
| 641 |
+
"epoch": 1.2853841709994223,
|
| 642 |
+
"grad_norm": 41.65699005126953,
|
| 643 |
+
"learning_rate": 1.2732719880162636e-06,
|
| 644 |
+
"loss": 0.3264,
|
| 645 |
+
"step": 2225
|
| 646 |
+
},
|
| 647 |
+
{
|
| 648 |
+
"epoch": 1.2998266897746968,
|
| 649 |
+
"grad_norm": 32.03495788574219,
|
| 650 |
+
"learning_rate": 1.262572223411085e-06,
|
| 651 |
+
"loss": 0.3404,
|
| 652 |
+
"step": 2250
|
| 653 |
+
},
|
| 654 |
+
{
|
| 655 |
+
"epoch": 1.314269208549971,
|
| 656 |
+
"grad_norm": 4.864631175994873,
|
| 657 |
+
"learning_rate": 1.2518724588059062e-06,
|
| 658 |
+
"loss": 0.296,
|
| 659 |
+
"step": 2275
|
| 660 |
+
},
|
| 661 |
+
{
|
| 662 |
+
"epoch": 1.3287117273252456,
|
| 663 |
+
"grad_norm": 10.562322616577148,
|
| 664 |
+
"learning_rate": 1.2411726942007275e-06,
|
| 665 |
+
"loss": 0.3442,
|
| 666 |
+
"step": 2300
|
| 667 |
+
},
|
| 668 |
+
{
|
| 669 |
+
"epoch": 1.34315424610052,
|
| 670 |
+
"grad_norm": 33.48724365234375,
|
| 671 |
+
"learning_rate": 1.230472929595549e-06,
|
| 672 |
+
"loss": 0.257,
|
| 673 |
+
"step": 2325
|
| 674 |
+
},
|
| 675 |
+
{
|
| 676 |
+
"epoch": 1.3575967648757943,
|
| 677 |
+
"grad_norm": 19.912137985229492,
|
| 678 |
+
"learning_rate": 1.2197731649903702e-06,
|
| 679 |
+
"loss": 0.2968,
|
| 680 |
+
"step": 2350
|
| 681 |
+
},
|
| 682 |
+
{
|
| 683 |
+
"epoch": 1.3720392836510689,
|
| 684 |
+
"grad_norm": 22.246639251708984,
|
| 685 |
+
"learning_rate": 1.2090734003851915e-06,
|
| 686 |
+
"loss": 0.2793,
|
| 687 |
+
"step": 2375
|
| 688 |
+
},
|
| 689 |
+
{
|
| 690 |
+
"epoch": 1.3864818024263432,
|
| 691 |
+
"grad_norm": 18.22015380859375,
|
| 692 |
+
"learning_rate": 1.1983736357800127e-06,
|
| 693 |
+
"loss": 0.3079,
|
| 694 |
+
"step": 2400
|
| 695 |
+
},
|
| 696 |
+
{
|
| 697 |
+
"epoch": 1.4009243212016176,
|
| 698 |
+
"grad_norm": 15.965062141418457,
|
| 699 |
+
"learning_rate": 1.1876738711748342e-06,
|
| 700 |
+
"loss": 0.335,
|
| 701 |
+
"step": 2425
|
| 702 |
+
},
|
| 703 |
+
{
|
| 704 |
+
"epoch": 1.415366839976892,
|
| 705 |
+
"grad_norm": 20.45452117919922,
|
| 706 |
+
"learning_rate": 1.1769741065696555e-06,
|
| 707 |
+
"loss": 0.3061,
|
| 708 |
+
"step": 2450
|
| 709 |
+
},
|
| 710 |
+
{
|
| 711 |
+
"epoch": 1.4298093587521663,
|
| 712 |
+
"grad_norm": 13.89696216583252,
|
| 713 |
+
"learning_rate": 1.1662743419644768e-06,
|
| 714 |
+
"loss": 0.275,
|
| 715 |
+
"step": 2475
|
| 716 |
+
},
|
| 717 |
+
{
|
| 718 |
+
"epoch": 1.4442518775274409,
|
| 719 |
+
"grad_norm": 35.64567947387695,
|
| 720 |
+
"learning_rate": 1.155574577359298e-06,
|
| 721 |
+
"loss": 0.3471,
|
| 722 |
+
"step": 2500
|
| 723 |
+
},
|
| 724 |
+
{
|
| 725 |
+
"epoch": 1.4586943963027152,
|
| 726 |
+
"grad_norm": 14.65186882019043,
|
| 727 |
+
"learning_rate": 1.1448748127541195e-06,
|
| 728 |
+
"loss": 0.314,
|
| 729 |
+
"step": 2525
|
| 730 |
+
},
|
| 731 |
+
{
|
| 732 |
+
"epoch": 1.4731369150779896,
|
| 733 |
+
"grad_norm": 14.541102409362793,
|
| 734 |
+
"learning_rate": 1.1341750481489408e-06,
|
| 735 |
+
"loss": 0.2916,
|
| 736 |
+
"step": 2550
|
| 737 |
+
},
|
| 738 |
+
{
|
| 739 |
+
"epoch": 1.487579433853264,
|
| 740 |
+
"grad_norm": 37.96781539916992,
|
| 741 |
+
"learning_rate": 1.123475283543762e-06,
|
| 742 |
+
"loss": 0.3849,
|
| 743 |
+
"step": 2575
|
| 744 |
+
},
|
| 745 |
+
{
|
| 746 |
+
"epoch": 1.5020219526285383,
|
| 747 |
+
"grad_norm": 16.675336837768555,
|
| 748 |
+
"learning_rate": 1.1127755189385833e-06,
|
| 749 |
+
"loss": 0.2672,
|
| 750 |
+
"step": 2600
|
| 751 |
+
},
|
| 752 |
+
{
|
| 753 |
+
"epoch": 1.516464471403813,
|
| 754 |
+
"grad_norm": 28.15886116027832,
|
| 755 |
+
"learning_rate": 1.1020757543334048e-06,
|
| 756 |
+
"loss": 0.3031,
|
| 757 |
+
"step": 2625
|
| 758 |
+
},
|
| 759 |
+
{
|
| 760 |
+
"epoch": 1.5309069901790873,
|
| 761 |
+
"grad_norm": 28.914554595947266,
|
| 762 |
+
"learning_rate": 1.091375989728226e-06,
|
| 763 |
+
"loss": 0.2585,
|
| 764 |
+
"step": 2650
|
| 765 |
+
},
|
| 766 |
+
{
|
| 767 |
+
"epoch": 1.5453495089543616,
|
| 768 |
+
"grad_norm": 31.952404022216797,
|
| 769 |
+
"learning_rate": 1.0806762251230471e-06,
|
| 770 |
+
"loss": 0.3405,
|
| 771 |
+
"step": 2675
|
| 772 |
+
},
|
| 773 |
+
{
|
| 774 |
+
"epoch": 1.5597920277296362,
|
| 775 |
+
"grad_norm": 62.01006317138672,
|
| 776 |
+
"learning_rate": 1.0699764605178684e-06,
|
| 777 |
+
"loss": 0.2892,
|
| 778 |
+
"step": 2700
|
| 779 |
+
},
|
| 780 |
+
{
|
| 781 |
+
"epoch": 1.5742345465049103,
|
| 782 |
+
"grad_norm": 15.054553985595703,
|
| 783 |
+
"learning_rate": 1.0592766959126899e-06,
|
| 784 |
+
"loss": 0.2902,
|
| 785 |
+
"step": 2725
|
| 786 |
+
},
|
| 787 |
+
{
|
| 788 |
+
"epoch": 1.588677065280185,
|
| 789 |
+
"grad_norm": 39.178443908691406,
|
| 790 |
+
"learning_rate": 1.0485769313075112e-06,
|
| 791 |
+
"loss": 0.2743,
|
| 792 |
+
"step": 2750
|
| 793 |
+
},
|
| 794 |
+
{
|
| 795 |
+
"epoch": 1.6031195840554593,
|
| 796 |
+
"grad_norm": 43.06193923950195,
|
| 797 |
+
"learning_rate": 1.0378771667023324e-06,
|
| 798 |
+
"loss": 0.2982,
|
| 799 |
+
"step": 2775
|
| 800 |
+
},
|
| 801 |
+
{
|
| 802 |
+
"epoch": 1.6175621028307337,
|
| 803 |
+
"grad_norm": 43.87297821044922,
|
| 804 |
+
"learning_rate": 1.0271774020971537e-06,
|
| 805 |
+
"loss": 0.2902,
|
| 806 |
+
"step": 2800
|
| 807 |
+
},
|
| 808 |
+
{
|
| 809 |
+
"epoch": 1.6320046216060082,
|
| 810 |
+
"grad_norm": 21.78912925720215,
|
| 811 |
+
"learning_rate": 1.0164776374919752e-06,
|
| 812 |
+
"loss": 0.3465,
|
| 813 |
+
"step": 2825
|
| 814 |
+
},
|
| 815 |
+
{
|
| 816 |
+
"epoch": 1.6464471403812824,
|
| 817 |
+
"grad_norm": 15.053204536437988,
|
| 818 |
+
"learning_rate": 1.0057778728867964e-06,
|
| 819 |
+
"loss": 0.3213,
|
| 820 |
+
"step": 2850
|
| 821 |
+
},
|
| 822 |
+
{
|
| 823 |
+
"epoch": 1.660889659156557,
|
| 824 |
+
"grad_norm": 21.79863166809082,
|
| 825 |
+
"learning_rate": 9.950781082816177e-07,
|
| 826 |
+
"loss": 0.3278,
|
| 827 |
+
"step": 2875
|
| 828 |
+
},
|
| 829 |
+
{
|
| 830 |
+
"epoch": 1.6753321779318313,
|
| 831 |
+
"grad_norm": 58.025299072265625,
|
| 832 |
+
"learning_rate": 9.843783436764392e-07,
|
| 833 |
+
"loss": 0.2828,
|
| 834 |
+
"step": 2900
|
| 835 |
+
},
|
| 836 |
+
{
|
| 837 |
+
"epoch": 1.6897746967071057,
|
| 838 |
+
"grad_norm": 22.137096405029297,
|
| 839 |
+
"learning_rate": 9.736785790712605e-07,
|
| 840 |
+
"loss": 0.3023,
|
| 841 |
+
"step": 2925
|
| 842 |
+
},
|
| 843 |
+
{
|
| 844 |
+
"epoch": 1.7042172154823803,
|
| 845 |
+
"grad_norm": 19.531232833862305,
|
| 846 |
+
"learning_rate": 9.629788144660817e-07,
|
| 847 |
+
"loss": 0.3039,
|
| 848 |
+
"step": 2950
|
| 849 |
+
},
|
| 850 |
+
{
|
| 851 |
+
"epoch": 1.7186597342576544,
|
| 852 |
+
"grad_norm": 59.77436065673828,
|
| 853 |
+
"learning_rate": 9.52279049860903e-07,
|
| 854 |
+
"loss": 0.3376,
|
| 855 |
+
"step": 2975
|
| 856 |
+
},
|
| 857 |
+
{
|
| 858 |
+
"epoch": 1.733102253032929,
|
| 859 |
+
"grad_norm": 27.803564071655273,
|
| 860 |
+
"learning_rate": 9.415792852557243e-07,
|
| 861 |
+
"loss": 0.2839,
|
| 862 |
+
"step": 3000
|
| 863 |
+
},
|
| 864 |
+
{
|
| 865 |
+
"epoch": 1.7475447718082033,
|
| 866 |
+
"grad_norm": 21.773244857788086,
|
| 867 |
+
"learning_rate": 9.308795206505456e-07,
|
| 868 |
+
"loss": 0.3372,
|
| 869 |
+
"step": 3025
|
| 870 |
+
},
|
| 871 |
+
{
|
| 872 |
+
"epoch": 1.7619872905834777,
|
| 873 |
+
"grad_norm": 11.421875953674316,
|
| 874 |
+
"learning_rate": 9.201797560453669e-07,
|
| 875 |
+
"loss": 0.3754,
|
| 876 |
+
"step": 3050
|
| 877 |
+
},
|
| 878 |
+
{
|
| 879 |
+
"epoch": 1.7764298093587523,
|
| 880 |
+
"grad_norm": 14.211411476135254,
|
| 881 |
+
"learning_rate": 9.094799914401883e-07,
|
| 882 |
+
"loss": 0.3214,
|
| 883 |
+
"step": 3075
|
| 884 |
+
},
|
| 885 |
+
{
|
| 886 |
+
"epoch": 1.7908723281340264,
|
| 887 |
+
"grad_norm": 43.777278900146484,
|
| 888 |
+
"learning_rate": 8.987802268350096e-07,
|
| 889 |
+
"loss": 0.3508,
|
| 890 |
+
"step": 3100
|
| 891 |
+
},
|
| 892 |
+
{
|
| 893 |
+
"epoch": 1.805314846909301,
|
| 894 |
+
"grad_norm": 38.14100646972656,
|
| 895 |
+
"learning_rate": 8.880804622298309e-07,
|
| 896 |
+
"loss": 0.2535,
|
| 897 |
+
"step": 3125
|
| 898 |
+
},
|
| 899 |
+
{
|
| 900 |
+
"epoch": 1.8197573656845754,
|
| 901 |
+
"grad_norm": 15.347945213317871,
|
| 902 |
+
"learning_rate": 8.773806976246522e-07,
|
| 903 |
+
"loss": 0.3121,
|
| 904 |
+
"step": 3150
|
| 905 |
+
},
|
| 906 |
+
{
|
| 907 |
+
"epoch": 1.8341998844598497,
|
| 908 |
+
"grad_norm": 8.05485725402832,
|
| 909 |
+
"learning_rate": 8.666809330194736e-07,
|
| 910 |
+
"loss": 0.3227,
|
| 911 |
+
"step": 3175
|
| 912 |
+
},
|
| 913 |
+
{
|
| 914 |
+
"epoch": 1.8486424032351243,
|
| 915 |
+
"grad_norm": 11.664706230163574,
|
| 916 |
+
"learning_rate": 8.559811684142948e-07,
|
| 917 |
+
"loss": 0.3061,
|
| 918 |
+
"step": 3200
|
| 919 |
+
},
|
| 920 |
+
{
|
| 921 |
+
"epoch": 1.8630849220103987,
|
| 922 |
+
"grad_norm": 7.515502452850342,
|
| 923 |
+
"learning_rate": 8.452814038091161e-07,
|
| 924 |
+
"loss": 0.2753,
|
| 925 |
+
"step": 3225
|
| 926 |
+
},
|
| 927 |
+
{
|
| 928 |
+
"epoch": 1.877527440785673,
|
| 929 |
+
"grad_norm": 30.233638763427734,
|
| 930 |
+
"learning_rate": 8.345816392039374e-07,
|
| 931 |
+
"loss": 0.3518,
|
| 932 |
+
"step": 3250
|
| 933 |
+
},
|
| 934 |
+
{
|
| 935 |
+
"epoch": 1.8919699595609474,
|
| 936 |
+
"grad_norm": 16.609712600708008,
|
| 937 |
+
"learning_rate": 8.238818745987588e-07,
|
| 938 |
+
"loss": 0.3087,
|
| 939 |
+
"step": 3275
|
| 940 |
+
},
|
| 941 |
+
{
|
| 942 |
+
"epoch": 1.9064124783362217,
|
| 943 |
+
"grad_norm": 12.235444068908691,
|
| 944 |
+
"learning_rate": 8.1318210999358e-07,
|
| 945 |
+
"loss": 0.3224,
|
| 946 |
+
"step": 3300
|
| 947 |
+
},
|
| 948 |
+
{
|
| 949 |
+
"epoch": 1.9208549971114963,
|
| 950 |
+
"grad_norm": 36.453224182128906,
|
| 951 |
+
"learning_rate": 8.024823453884014e-07,
|
| 952 |
+
"loss": 0.3311,
|
| 953 |
+
"step": 3325
|
| 954 |
+
},
|
| 955 |
+
{
|
| 956 |
+
"epoch": 1.9352975158867707,
|
| 957 |
+
"grad_norm": 21.512168884277344,
|
| 958 |
+
"learning_rate": 7.917825807832227e-07,
|
| 959 |
+
"loss": 0.2857,
|
| 960 |
+
"step": 3350
|
| 961 |
+
},
|
| 962 |
+
{
|
| 963 |
+
"epoch": 1.949740034662045,
|
| 964 |
+
"grad_norm": 9.703317642211914,
|
| 965 |
+
"learning_rate": 7.81082816178044e-07,
|
| 966 |
+
"loss": 0.2662,
|
| 967 |
+
"step": 3375
|
| 968 |
+
},
|
| 969 |
+
{
|
| 970 |
+
"epoch": 1.9641825534373196,
|
| 971 |
+
"grad_norm": 17.714481353759766,
|
| 972 |
+
"learning_rate": 7.703830515728653e-07,
|
| 973 |
+
"loss": 0.291,
|
| 974 |
+
"step": 3400
|
| 975 |
+
},
|
| 976 |
+
{
|
| 977 |
+
"epoch": 1.9786250722125938,
|
| 978 |
+
"grad_norm": 22.379777908325195,
|
| 979 |
+
"learning_rate": 7.596832869676867e-07,
|
| 980 |
+
"loss": 0.3267,
|
| 981 |
+
"step": 3425
|
| 982 |
+
},
|
| 983 |
+
{
|
| 984 |
+
"epoch": 1.9930675909878683,
|
| 985 |
+
"grad_norm": 8.563464164733887,
|
| 986 |
+
"learning_rate": 7.48983522362508e-07,
|
| 987 |
+
"loss": 0.2939,
|
| 988 |
+
"step": 3450
|
| 989 |
+
},
|
| 990 |
+
{
|
| 991 |
+
"epoch": 2.0,
|
| 992 |
+
"eval_explained_variance": 0.7469815611839294,
|
| 993 |
+
"eval_loss": 0.31648534536361694,
|
| 994 |
+
"eval_mae": 0.4420657455921173,
|
| 995 |
+
"eval_mse": 0.3162277936935425,
|
| 996 |
+
"eval_r2": 0.7133824489512686,
|
| 997 |
+
"eval_rmse": 0.5623413324356079,
|
| 998 |
+
"eval_runtime": 80.5259,
|
| 999 |
+
"eval_samples_per_second": 42.992,
|
| 1000 |
+
"eval_steps_per_second": 2.695,
|
| 1001 |
+
"step": 3462
|
| 1002 |
+
},
|
| 1003 |
+
{
|
| 1004 |
+
"epoch": 2.0075101097631425,
|
| 1005 |
+
"grad_norm": 31.694887161254883,
|
| 1006 |
+
"learning_rate": 7.382837577573293e-07,
|
| 1007 |
+
"loss": 0.2527,
|
| 1008 |
+
"step": 3475
|
| 1009 |
+
},
|
| 1010 |
+
{
|
| 1011 |
+
"epoch": 2.021952628538417,
|
| 1012 |
+
"grad_norm": 24.721397399902344,
|
| 1013 |
+
"learning_rate": 7.275839931521506e-07,
|
| 1014 |
+
"loss": 0.2815,
|
| 1015 |
+
"step": 3500
|
| 1016 |
+
},
|
| 1017 |
+
{
|
| 1018 |
+
"epoch": 2.0363951473136916,
|
| 1019 |
+
"grad_norm": 33.44636917114258,
|
| 1020 |
+
"learning_rate": 7.16884228546972e-07,
|
| 1021 |
+
"loss": 0.3586,
|
| 1022 |
+
"step": 3525
|
| 1023 |
+
},
|
| 1024 |
+
{
|
| 1025 |
+
"epoch": 2.050837666088966,
|
| 1026 |
+
"grad_norm": 44.72824478149414,
|
| 1027 |
+
"learning_rate": 7.061844639417933e-07,
|
| 1028 |
+
"loss": 0.3163,
|
| 1029 |
+
"step": 3550
|
| 1030 |
+
},
|
| 1031 |
+
{
|
| 1032 |
+
"epoch": 2.0652801848642404,
|
| 1033 |
+
"grad_norm": 22.65967559814453,
|
| 1034 |
+
"learning_rate": 6.954846993366146e-07,
|
| 1035 |
+
"loss": 0.2523,
|
| 1036 |
+
"step": 3575
|
| 1037 |
+
},
|
| 1038 |
+
{
|
| 1039 |
+
"epoch": 2.079722703639515,
|
| 1040 |
+
"grad_norm": 9.611360549926758,
|
| 1041 |
+
"learning_rate": 6.847849347314358e-07,
|
| 1042 |
+
"loss": 0.2609,
|
| 1043 |
+
"step": 3600
|
| 1044 |
+
},
|
| 1045 |
+
{
|
| 1046 |
+
"epoch": 2.094165222414789,
|
| 1047 |
+
"grad_norm": 19.328899383544922,
|
| 1048 |
+
"learning_rate": 6.740851701262572e-07,
|
| 1049 |
+
"loss": 0.3327,
|
| 1050 |
+
"step": 3625
|
| 1051 |
+
},
|
| 1052 |
+
{
|
| 1053 |
+
"epoch": 2.1086077411900637,
|
| 1054 |
+
"grad_norm": 17.564197540283203,
|
| 1055 |
+
"learning_rate": 6.633854055210784e-07,
|
| 1056 |
+
"loss": 0.2777,
|
| 1057 |
+
"step": 3650
|
| 1058 |
+
},
|
| 1059 |
+
{
|
| 1060 |
+
"epoch": 2.123050259965338,
|
| 1061 |
+
"grad_norm": 35.05995178222656,
|
| 1062 |
+
"learning_rate": 6.526856409158998e-07,
|
| 1063 |
+
"loss": 0.2516,
|
| 1064 |
+
"step": 3675
|
| 1065 |
+
},
|
| 1066 |
+
{
|
| 1067 |
+
"epoch": 2.1374927787406124,
|
| 1068 |
+
"grad_norm": 17.389116287231445,
|
| 1069 |
+
"learning_rate": 6.419858763107211e-07,
|
| 1070 |
+
"loss": 0.2387,
|
| 1071 |
+
"step": 3700
|
| 1072 |
+
},
|
| 1073 |
+
{
|
| 1074 |
+
"epoch": 2.151935297515887,
|
| 1075 |
+
"grad_norm": 18.017724990844727,
|
| 1076 |
+
"learning_rate": 6.312861117055425e-07,
|
| 1077 |
+
"loss": 0.2764,
|
| 1078 |
+
"step": 3725
|
| 1079 |
+
},
|
| 1080 |
+
{
|
| 1081 |
+
"epoch": 2.166377816291161,
|
| 1082 |
+
"grad_norm": 13.583812713623047,
|
| 1083 |
+
"learning_rate": 6.205863471003637e-07,
|
| 1084 |
+
"loss": 0.2717,
|
| 1085 |
+
"step": 3750
|
| 1086 |
+
},
|
| 1087 |
+
{
|
| 1088 |
+
"epoch": 2.1808203350664357,
|
| 1089 |
+
"grad_norm": 18.499242782592773,
|
| 1090 |
+
"learning_rate": 6.098865824951851e-07,
|
| 1091 |
+
"loss": 0.2439,
|
| 1092 |
+
"step": 3775
|
| 1093 |
+
},
|
| 1094 |
+
{
|
| 1095 |
+
"epoch": 2.19526285384171,
|
| 1096 |
+
"grad_norm": 17.863845825195312,
|
| 1097 |
+
"learning_rate": 5.991868178900064e-07,
|
| 1098 |
+
"loss": 0.2498,
|
| 1099 |
+
"step": 3800
|
| 1100 |
+
},
|
| 1101 |
+
{
|
| 1102 |
+
"epoch": 2.2097053726169844,
|
| 1103 |
+
"grad_norm": 42.79360580444336,
|
| 1104 |
+
"learning_rate": 5.884870532848277e-07,
|
| 1105 |
+
"loss": 0.2986,
|
| 1106 |
+
"step": 3825
|
| 1107 |
+
},
|
| 1108 |
+
{
|
| 1109 |
+
"epoch": 2.224147891392259,
|
| 1110 |
+
"grad_norm": 18.10019302368164,
|
| 1111 |
+
"learning_rate": 5.77787288679649e-07,
|
| 1112 |
+
"loss": 0.2692,
|
| 1113 |
+
"step": 3850
|
| 1114 |
+
},
|
| 1115 |
+
{
|
| 1116 |
+
"epoch": 2.238590410167533,
|
| 1117 |
+
"grad_norm": 52.769935607910156,
|
| 1118 |
+
"learning_rate": 5.670875240744704e-07,
|
| 1119 |
+
"loss": 0.3265,
|
| 1120 |
+
"step": 3875
|
| 1121 |
+
},
|
| 1122 |
+
{
|
| 1123 |
+
"epoch": 2.2530329289428077,
|
| 1124 |
+
"grad_norm": 42.038516998291016,
|
| 1125 |
+
"learning_rate": 5.563877594692917e-07,
|
| 1126 |
+
"loss": 0.3196,
|
| 1127 |
+
"step": 3900
|
| 1128 |
+
},
|
| 1129 |
+
{
|
| 1130 |
+
"epoch": 2.267475447718082,
|
| 1131 |
+
"grad_norm": 14.1666898727417,
|
| 1132 |
+
"learning_rate": 5.45687994864113e-07,
|
| 1133 |
+
"loss": 0.2888,
|
| 1134 |
+
"step": 3925
|
| 1135 |
+
},
|
| 1136 |
+
{
|
| 1137 |
+
"epoch": 2.2819179664933564,
|
| 1138 |
+
"grad_norm": 16.471778869628906,
|
| 1139 |
+
"learning_rate": 5.349882302589342e-07,
|
| 1140 |
+
"loss": 0.2782,
|
| 1141 |
+
"step": 3950
|
| 1142 |
+
},
|
| 1143 |
+
{
|
| 1144 |
+
"epoch": 2.296360485268631,
|
| 1145 |
+
"grad_norm": 9.197157859802246,
|
| 1146 |
+
"learning_rate": 5.242884656537556e-07,
|
| 1147 |
+
"loss": 0.3127,
|
| 1148 |
+
"step": 3975
|
| 1149 |
+
},
|
| 1150 |
+
{
|
| 1151 |
+
"epoch": 2.310803004043905,
|
| 1152 |
+
"grad_norm": 19.208568572998047,
|
| 1153 |
+
"learning_rate": 5.135887010485768e-07,
|
| 1154 |
+
"loss": 0.2572,
|
| 1155 |
+
"step": 4000
|
| 1156 |
+
},
|
| 1157 |
+
{
|
| 1158 |
+
"epoch": 2.3252455228191797,
|
| 1159 |
+
"grad_norm": 5.966078758239746,
|
| 1160 |
+
"learning_rate": 5.028889364433982e-07,
|
| 1161 |
+
"loss": 0.2631,
|
| 1162 |
+
"step": 4025
|
| 1163 |
+
},
|
| 1164 |
+
{
|
| 1165 |
+
"epoch": 2.339688041594454,
|
| 1166 |
+
"grad_norm": 27.037731170654297,
|
| 1167 |
+
"learning_rate": 4.921891718382196e-07,
|
| 1168 |
+
"loss": 0.2794,
|
| 1169 |
+
"step": 4050
|
| 1170 |
+
},
|
| 1171 |
+
{
|
| 1172 |
+
"epoch": 2.3541305603697285,
|
| 1173 |
+
"grad_norm": 39.20252990722656,
|
| 1174 |
+
"learning_rate": 4.814894072330409e-07,
|
| 1175 |
+
"loss": 0.2656,
|
| 1176 |
+
"step": 4075
|
| 1177 |
+
},
|
| 1178 |
+
{
|
| 1179 |
+
"epoch": 2.368573079145003,
|
| 1180 |
+
"grad_norm": 32.399147033691406,
|
| 1181 |
+
"learning_rate": 4.7078964262786213e-07,
|
| 1182 |
+
"loss": 0.2654,
|
| 1183 |
+
"step": 4100
|
| 1184 |
+
},
|
| 1185 |
+
{
|
| 1186 |
+
"epoch": 2.383015597920277,
|
| 1187 |
+
"grad_norm": 23.706451416015625,
|
| 1188 |
+
"learning_rate": 4.6008987802268346e-07,
|
| 1189 |
+
"loss": 0.2887,
|
| 1190 |
+
"step": 4125
|
| 1191 |
+
},
|
| 1192 |
+
{
|
| 1193 |
+
"epoch": 2.3974581166955518,
|
| 1194 |
+
"grad_norm": 15.86970043182373,
|
| 1195 |
+
"learning_rate": 4.493901134175048e-07,
|
| 1196 |
+
"loss": 0.2568,
|
| 1197 |
+
"step": 4150
|
| 1198 |
+
},
|
| 1199 |
+
{
|
| 1200 |
+
"epoch": 2.4119006354708263,
|
| 1201 |
+
"grad_norm": 27.933916091918945,
|
| 1202 |
+
"learning_rate": 4.386903488123261e-07,
|
| 1203 |
+
"loss": 0.2536,
|
| 1204 |
+
"step": 4175
|
| 1205 |
+
},
|
| 1206 |
+
{
|
| 1207 |
+
"epoch": 2.4263431542461005,
|
| 1208 |
+
"grad_norm": 16.812334060668945,
|
| 1209 |
+
"learning_rate": 4.279905842071474e-07,
|
| 1210 |
+
"loss": 0.2542,
|
| 1211 |
+
"step": 4200
|
| 1212 |
+
},
|
| 1213 |
+
{
|
| 1214 |
+
"epoch": 2.440785673021375,
|
| 1215 |
+
"grad_norm": 38.82505416870117,
|
| 1216 |
+
"learning_rate": 4.172908196019687e-07,
|
| 1217 |
+
"loss": 0.2872,
|
| 1218 |
+
"step": 4225
|
| 1219 |
+
},
|
| 1220 |
+
{
|
| 1221 |
+
"epoch": 2.455228191796649,
|
| 1222 |
+
"grad_norm": 23.149492263793945,
|
| 1223 |
+
"learning_rate": 4.0659105499679e-07,
|
| 1224 |
+
"loss": 0.2818,
|
| 1225 |
+
"step": 4250
|
| 1226 |
+
},
|
| 1227 |
+
{
|
| 1228 |
+
"epoch": 2.4696707105719238,
|
| 1229 |
+
"grad_norm": 43.19930648803711,
|
| 1230 |
+
"learning_rate": 3.9589129039161134e-07,
|
| 1231 |
+
"loss": 0.213,
|
| 1232 |
+
"step": 4275
|
| 1233 |
+
},
|
| 1234 |
+
{
|
| 1235 |
+
"epoch": 2.484113229347198,
|
| 1236 |
+
"grad_norm": 23.671152114868164,
|
| 1237 |
+
"learning_rate": 3.8519152578643266e-07,
|
| 1238 |
+
"loss": 0.3698,
|
| 1239 |
+
"step": 4300
|
| 1240 |
+
},
|
| 1241 |
+
{
|
| 1242 |
+
"epoch": 2.4985557481224725,
|
| 1243 |
+
"grad_norm": 88.69607543945312,
|
| 1244 |
+
"learning_rate": 3.74491761181254e-07,
|
| 1245 |
+
"loss": 0.2968,
|
| 1246 |
+
"step": 4325
|
| 1247 |
+
},
|
| 1248 |
+
{
|
| 1249 |
+
"epoch": 2.512998266897747,
|
| 1250 |
+
"grad_norm": 9.653864860534668,
|
| 1251 |
+
"learning_rate": 3.637919965760753e-07,
|
| 1252 |
+
"loss": 0.2777,
|
| 1253 |
+
"step": 4350
|
| 1254 |
+
},
|
| 1255 |
+
{
|
| 1256 |
+
"epoch": 2.527440785673021,
|
| 1257 |
+
"grad_norm": 11.768026351928711,
|
| 1258 |
+
"learning_rate": 3.5309223197089663e-07,
|
| 1259 |
+
"loss": 0.3236,
|
| 1260 |
+
"step": 4375
|
| 1261 |
+
},
|
| 1262 |
+
{
|
| 1263 |
+
"epoch": 2.541883304448296,
|
| 1264 |
+
"grad_norm": 15.171217918395996,
|
| 1265 |
+
"learning_rate": 3.423924673657179e-07,
|
| 1266 |
+
"loss": 0.3077,
|
| 1267 |
+
"step": 4400
|
| 1268 |
+
},
|
| 1269 |
+
{
|
| 1270 |
+
"epoch": 2.5563258232235704,
|
| 1271 |
+
"grad_norm": 9.879386901855469,
|
| 1272 |
+
"learning_rate": 3.316927027605392e-07,
|
| 1273 |
+
"loss": 0.2684,
|
| 1274 |
+
"step": 4425
|
| 1275 |
+
},
|
| 1276 |
+
{
|
| 1277 |
+
"epoch": 2.5707683419988445,
|
| 1278 |
+
"grad_norm": 9.355985641479492,
|
| 1279 |
+
"learning_rate": 3.2099293815536054e-07,
|
| 1280 |
+
"loss": 0.2526,
|
| 1281 |
+
"step": 4450
|
| 1282 |
+
},
|
| 1283 |
+
{
|
| 1284 |
+
"epoch": 2.585210860774119,
|
| 1285 |
+
"grad_norm": 4.87063455581665,
|
| 1286 |
+
"learning_rate": 3.1029317355018186e-07,
|
| 1287 |
+
"loss": 0.2689,
|
| 1288 |
+
"step": 4475
|
| 1289 |
+
},
|
| 1290 |
+
{
|
| 1291 |
+
"epoch": 2.5996533795493937,
|
| 1292 |
+
"grad_norm": 20.083267211914062,
|
| 1293 |
+
"learning_rate": 2.995934089450032e-07,
|
| 1294 |
+
"loss": 0.259,
|
| 1295 |
+
"step": 4500
|
| 1296 |
+
},
|
| 1297 |
+
{
|
| 1298 |
+
"epoch": 2.614095898324668,
|
| 1299 |
+
"grad_norm": 12.317808151245117,
|
| 1300 |
+
"learning_rate": 2.888936443398245e-07,
|
| 1301 |
+
"loss": 0.2232,
|
| 1302 |
+
"step": 4525
|
| 1303 |
+
},
|
| 1304 |
+
{
|
| 1305 |
+
"epoch": 2.628538417099942,
|
| 1306 |
+
"grad_norm": 28.255945205688477,
|
| 1307 |
+
"learning_rate": 2.7819387973464583e-07,
|
| 1308 |
+
"loss": 0.2466,
|
| 1309 |
+
"step": 4550
|
| 1310 |
+
},
|
| 1311 |
+
{
|
| 1312 |
+
"epoch": 2.6429809358752165,
|
| 1313 |
+
"grad_norm": 72.9136734008789,
|
| 1314 |
+
"learning_rate": 2.674941151294671e-07,
|
| 1315 |
+
"loss": 0.2693,
|
| 1316 |
+
"step": 4575
|
| 1317 |
+
},
|
| 1318 |
+
{
|
| 1319 |
+
"epoch": 2.657423454650491,
|
| 1320 |
+
"grad_norm": 44.20970153808594,
|
| 1321 |
+
"learning_rate": 2.567943505242884e-07,
|
| 1322 |
+
"loss": 0.2527,
|
| 1323 |
+
"step": 4600
|
| 1324 |
+
},
|
| 1325 |
+
{
|
| 1326 |
+
"epoch": 2.6718659734257653,
|
| 1327 |
+
"grad_norm": 17.912519454956055,
|
| 1328 |
+
"learning_rate": 2.460945859191098e-07,
|
| 1329 |
+
"loss": 0.2706,
|
| 1330 |
+
"step": 4625
|
| 1331 |
+
},
|
| 1332 |
+
{
|
| 1333 |
+
"epoch": 2.68630849220104,
|
| 1334 |
+
"grad_norm": 48.64137649536133,
|
| 1335 |
+
"learning_rate": 2.3539482131393107e-07,
|
| 1336 |
+
"loss": 0.2795,
|
| 1337 |
+
"step": 4650
|
| 1338 |
+
},
|
| 1339 |
+
{
|
| 1340 |
+
"epoch": 2.7007510109763144,
|
| 1341 |
+
"grad_norm": 39.6313362121582,
|
| 1342 |
+
"learning_rate": 2.246950567087524e-07,
|
| 1343 |
+
"loss": 0.2669,
|
| 1344 |
+
"step": 4675
|
| 1345 |
+
},
|
| 1346 |
+
{
|
| 1347 |
+
"epoch": 2.7151935297515886,
|
| 1348 |
+
"grad_norm": 12.336877822875977,
|
| 1349 |
+
"learning_rate": 2.139952921035737e-07,
|
| 1350 |
+
"loss": 0.2791,
|
| 1351 |
+
"step": 4700
|
| 1352 |
+
},
|
| 1353 |
+
{
|
| 1354 |
+
"epoch": 2.729636048526863,
|
| 1355 |
+
"grad_norm": 11.376914024353027,
|
| 1356 |
+
"learning_rate": 2.03295527498395e-07,
|
| 1357 |
+
"loss": 0.2599,
|
| 1358 |
+
"step": 4725
|
| 1359 |
+
},
|
| 1360 |
+
{
|
| 1361 |
+
"epoch": 2.7440785673021377,
|
| 1362 |
+
"grad_norm": 26.109207153320312,
|
| 1363 |
+
"learning_rate": 1.9259576289321633e-07,
|
| 1364 |
+
"loss": 0.3112,
|
| 1365 |
+
"step": 4750
|
| 1366 |
+
},
|
| 1367 |
+
{
|
| 1368 |
+
"epoch": 2.758521086077412,
|
| 1369 |
+
"grad_norm": 13.475809097290039,
|
| 1370 |
+
"learning_rate": 1.8189599828803765e-07,
|
| 1371 |
+
"loss": 0.2797,
|
| 1372 |
+
"step": 4775
|
| 1373 |
+
},
|
| 1374 |
+
{
|
| 1375 |
+
"epoch": 2.7729636048526864,
|
| 1376 |
+
"grad_norm": 23.833911895751953,
|
| 1377 |
+
"learning_rate": 1.7119623368285895e-07,
|
| 1378 |
+
"loss": 0.264,
|
| 1379 |
+
"step": 4800
|
| 1380 |
+
},
|
| 1381 |
+
{
|
| 1382 |
+
"epoch": 2.7874061236279606,
|
| 1383 |
+
"grad_norm": 7.037588119506836,
|
| 1384 |
+
"learning_rate": 1.6049646907768027e-07,
|
| 1385 |
+
"loss": 0.3202,
|
| 1386 |
+
"step": 4825
|
| 1387 |
+
},
|
| 1388 |
+
{
|
| 1389 |
+
"epoch": 2.801848642403235,
|
| 1390 |
+
"grad_norm": 36.29332733154297,
|
| 1391 |
+
"learning_rate": 1.497967044725016e-07,
|
| 1392 |
+
"loss": 0.286,
|
| 1393 |
+
"step": 4850
|
| 1394 |
+
},
|
| 1395 |
+
{
|
| 1396 |
+
"epoch": 2.8162911611785093,
|
| 1397 |
+
"grad_norm": 10.196252822875977,
|
| 1398 |
+
"learning_rate": 1.3909693986732292e-07,
|
| 1399 |
+
"loss": 0.3008,
|
| 1400 |
+
"step": 4875
|
| 1401 |
+
},
|
| 1402 |
+
{
|
| 1403 |
+
"epoch": 2.830733679953784,
|
| 1404 |
+
"grad_norm": 27.923114776611328,
|
| 1405 |
+
"learning_rate": 1.283971752621442e-07,
|
| 1406 |
+
"loss": 0.2778,
|
| 1407 |
+
"step": 4900
|
| 1408 |
+
},
|
| 1409 |
+
{
|
| 1410 |
+
"epoch": 2.8451761987290585,
|
| 1411 |
+
"grad_norm": 5.924576282501221,
|
| 1412 |
+
"learning_rate": 1.1769741065696553e-07,
|
| 1413 |
+
"loss": 0.2237,
|
| 1414 |
+
"step": 4925
|
| 1415 |
+
},
|
| 1416 |
+
{
|
| 1417 |
+
"epoch": 2.8596187175043326,
|
| 1418 |
+
"grad_norm": 16.432357788085938,
|
| 1419 |
+
"learning_rate": 1.0699764605178686e-07,
|
| 1420 |
+
"loss": 0.2532,
|
| 1421 |
+
"step": 4950
|
| 1422 |
+
},
|
| 1423 |
+
{
|
| 1424 |
+
"epoch": 2.874061236279607,
|
| 1425 |
+
"grad_norm": 10.846713066101074,
|
| 1426 |
+
"learning_rate": 9.629788144660816e-08,
|
| 1427 |
+
"loss": 0.2277,
|
| 1428 |
+
"step": 4975
|
| 1429 |
+
},
|
| 1430 |
+
{
|
| 1431 |
+
"epoch": 2.8885037550548818,
|
| 1432 |
+
"grad_norm": 21.979785919189453,
|
| 1433 |
+
"learning_rate": 8.559811684142947e-08,
|
| 1434 |
+
"loss": 0.3065,
|
| 1435 |
+
"step": 5000
|
| 1436 |
+
},
|
| 1437 |
+
{
|
| 1438 |
+
"epoch": 2.902946273830156,
|
| 1439 |
+
"grad_norm": 17.25764274597168,
|
| 1440 |
+
"learning_rate": 7.48983522362508e-08,
|
| 1441 |
+
"loss": 0.2449,
|
| 1442 |
+
"step": 5025
|
| 1443 |
+
},
|
| 1444 |
+
{
|
| 1445 |
+
"epoch": 2.9173887926054305,
|
| 1446 |
+
"grad_norm": 20.356718063354492,
|
| 1447 |
+
"learning_rate": 6.41985876310721e-08,
|
| 1448 |
+
"loss": 0.2357,
|
| 1449 |
+
"step": 5050
|
| 1450 |
+
},
|
| 1451 |
+
{
|
| 1452 |
+
"epoch": 2.9318313113807046,
|
| 1453 |
+
"grad_norm": 18.828088760375977,
|
| 1454 |
+
"learning_rate": 5.349882302589343e-08,
|
| 1455 |
+
"loss": 0.2329,
|
| 1456 |
+
"step": 5075
|
| 1457 |
+
},
|
| 1458 |
+
{
|
| 1459 |
+
"epoch": 2.946273830155979,
|
| 1460 |
+
"grad_norm": 10.398417472839355,
|
| 1461 |
+
"learning_rate": 4.279905842071474e-08,
|
| 1462 |
+
"loss": 0.2279,
|
| 1463 |
+
"step": 5100
|
| 1464 |
+
},
|
| 1465 |
+
{
|
| 1466 |
+
"epoch": 2.9607163489312533,
|
| 1467 |
+
"grad_norm": 40.09988784790039,
|
| 1468 |
+
"learning_rate": 3.209929381553605e-08,
|
| 1469 |
+
"loss": 0.2337,
|
| 1470 |
+
"step": 5125
|
| 1471 |
+
},
|
| 1472 |
+
{
|
| 1473 |
+
"epoch": 2.975158867706528,
|
| 1474 |
+
"grad_norm": 22.620283126831055,
|
| 1475 |
+
"learning_rate": 2.139952921035737e-08,
|
| 1476 |
+
"loss": 0.2783,
|
| 1477 |
+
"step": 5150
|
| 1478 |
+
},
|
| 1479 |
+
{
|
| 1480 |
+
"epoch": 2.9896013864818025,
|
| 1481 |
+
"grad_norm": 33.585792541503906,
|
| 1482 |
+
"learning_rate": 1.0699764605178684e-08,
|
| 1483 |
+
"loss": 0.2726,
|
| 1484 |
+
"step": 5175
|
| 1485 |
+
},
|
| 1486 |
+
{
|
| 1487 |
+
"epoch": 3.0,
|
| 1488 |
+
"eval_explained_variance": 0.7570163011550903,
|
| 1489 |
+
"eval_loss": 0.282262921333313,
|
| 1490 |
+
"eval_mae": 0.4189736545085907,
|
| 1491 |
+
"eval_mse": 0.2820460796356201,
|
| 1492 |
+
"eval_r2": 0.74436353679844,
|
| 1493 |
+
"eval_rmse": 0.5310801267623901,
|
| 1494 |
+
"eval_runtime": 80.5385,
|
| 1495 |
+
"eval_samples_per_second": 42.986,
|
| 1496 |
+
"eval_steps_per_second": 2.694,
|
| 1497 |
+
"step": 5193
|
| 1498 |
+
}
|
| 1499 |
+
],
|
| 1500 |
+
"logging_steps": 25,
|
| 1501 |
+
"max_steps": 5193,
|
| 1502 |
+
"num_input_tokens_seen": 0,
|
| 1503 |
+
"num_train_epochs": 3,
|
| 1504 |
+
"save_steps": 500,
|
| 1505 |
+
"total_flos": 7.743257397795226e+16,
|
| 1506 |
+
"train_batch_size": 2,
|
| 1507 |
+
"trial_name": null,
|
| 1508 |
+
"trial_params": null
|
| 1509 |
+
}
|
checkpoint-5193/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:270da9b099b2d02a736480bd012939e9b5f02ea692f21c4e08c4925dc1f458dc
|
| 3 |
+
size 5048
|
config.json
ADDED
|
@@ -0,0 +1,42 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "microsoft/deberta-v3-large",
|
| 3 |
+
"_num_labels": 1,
|
| 4 |
+
"architectures": [
|
| 5 |
+
"DebertaV2ForSequenceClassification"
|
| 6 |
+
],
|
| 7 |
+
"attention_probs_dropout_prob": 0.1,
|
| 8 |
+
"hidden_act": "gelu",
|
| 9 |
+
"hidden_dropout_prob": 0.1,
|
| 10 |
+
"hidden_size": 1024,
|
| 11 |
+
"id2label": {
|
| 12 |
+
"0": "target"
|
| 13 |
+
},
|
| 14 |
+
"initializer_range": 0.02,
|
| 15 |
+
"intermediate_size": 4096,
|
| 16 |
+
"label2id": {
|
| 17 |
+
"target": 0
|
| 18 |
+
},
|
| 19 |
+
"layer_norm_eps": 1e-07,
|
| 20 |
+
"max_position_embeddings": 512,
|
| 21 |
+
"max_relative_positions": -1,
|
| 22 |
+
"model_type": "deberta-v2",
|
| 23 |
+
"norm_rel_ebd": "layer_norm",
|
| 24 |
+
"num_attention_heads": 16,
|
| 25 |
+
"num_hidden_layers": 24,
|
| 26 |
+
"pad_token_id": 0,
|
| 27 |
+
"pooler_dropout": 0,
|
| 28 |
+
"pooler_hidden_act": "gelu",
|
| 29 |
+
"pooler_hidden_size": 1024,
|
| 30 |
+
"pos_att_type": [
|
| 31 |
+
"p2c",
|
| 32 |
+
"c2p"
|
| 33 |
+
],
|
| 34 |
+
"position_biased_input": false,
|
| 35 |
+
"position_buckets": 256,
|
| 36 |
+
"relative_attention": true,
|
| 37 |
+
"share_att_key": true,
|
| 38 |
+
"torch_dtype": "float32",
|
| 39 |
+
"transformers_version": "4.40.1",
|
| 40 |
+
"type_vocab_size": 0,
|
| 41 |
+
"vocab_size": 128100
|
| 42 |
+
}
|
model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:3d68d3238d7e01448edfe7b1983dcf19cc58c6bfbb7628f33a97a5690b6e914d
|
| 3 |
+
size 1740300340
|
runs/Apr29_14-53-13_r-abhishek-autotrain-1bfjaa9n-00d8c-hkw4p/events.out.tfevents.1714402394.r-abhishek-autotrain-1bfjaa9n-00d8c-hkw4p.464.0
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c1d3054238590a97382b063b811087cd93da94519af24987471b0223a836fdde
|
| 3 |
+
size 50525
|
runs/Apr29_14-53-13_r-abhishek-autotrain-1bfjaa9n-00d8c-hkw4p/events.out.tfevents.1714407850.r-abhishek-autotrain-1bfjaa9n-00d8c-hkw4p.464.1
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:81b9b3dff0d2329581b485dd6e9867790b2a9b813bd88a0ec2b931a09c093664
|
| 3 |
+
size 609
|
special_tokens_map.json
ADDED
|
@@ -0,0 +1,15 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": "[CLS]",
|
| 3 |
+
"cls_token": "[CLS]",
|
| 4 |
+
"eos_token": "[SEP]",
|
| 5 |
+
"mask_token": "[MASK]",
|
| 6 |
+
"pad_token": "[PAD]",
|
| 7 |
+
"sep_token": "[SEP]",
|
| 8 |
+
"unk_token": {
|
| 9 |
+
"content": "[UNK]",
|
| 10 |
+
"lstrip": false,
|
| 11 |
+
"normalized": true,
|
| 12 |
+
"rstrip": false,
|
| 13 |
+
"single_word": false
|
| 14 |
+
}
|
| 15 |
+
}
|
spm.model
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c679fbf93643d19aab7ee10c0b99e460bdbc02fedf34b92b05af343b4af586fd
|
| 3 |
+
size 2464616
|
tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
tokenizer_config.json
ADDED
|
@@ -0,0 +1,58 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"added_tokens_decoder": {
|
| 3 |
+
"0": {
|
| 4 |
+
"content": "[PAD]",
|
| 5 |
+
"lstrip": false,
|
| 6 |
+
"normalized": false,
|
| 7 |
+
"rstrip": false,
|
| 8 |
+
"single_word": false,
|
| 9 |
+
"special": true
|
| 10 |
+
},
|
| 11 |
+
"1": {
|
| 12 |
+
"content": "[CLS]",
|
| 13 |
+
"lstrip": false,
|
| 14 |
+
"normalized": false,
|
| 15 |
+
"rstrip": false,
|
| 16 |
+
"single_word": false,
|
| 17 |
+
"special": true
|
| 18 |
+
},
|
| 19 |
+
"2": {
|
| 20 |
+
"content": "[SEP]",
|
| 21 |
+
"lstrip": false,
|
| 22 |
+
"normalized": false,
|
| 23 |
+
"rstrip": false,
|
| 24 |
+
"single_word": false,
|
| 25 |
+
"special": true
|
| 26 |
+
},
|
| 27 |
+
"3": {
|
| 28 |
+
"content": "[UNK]",
|
| 29 |
+
"lstrip": false,
|
| 30 |
+
"normalized": true,
|
| 31 |
+
"rstrip": false,
|
| 32 |
+
"single_word": false,
|
| 33 |
+
"special": true
|
| 34 |
+
},
|
| 35 |
+
"128000": {
|
| 36 |
+
"content": "[MASK]",
|
| 37 |
+
"lstrip": false,
|
| 38 |
+
"normalized": false,
|
| 39 |
+
"rstrip": false,
|
| 40 |
+
"single_word": false,
|
| 41 |
+
"special": true
|
| 42 |
+
}
|
| 43 |
+
},
|
| 44 |
+
"bos_token": "[CLS]",
|
| 45 |
+
"clean_up_tokenization_spaces": true,
|
| 46 |
+
"cls_token": "[CLS]",
|
| 47 |
+
"do_lower_case": false,
|
| 48 |
+
"eos_token": "[SEP]",
|
| 49 |
+
"mask_token": "[MASK]",
|
| 50 |
+
"model_max_length": 1000000000000000019884624838656,
|
| 51 |
+
"pad_token": "[PAD]",
|
| 52 |
+
"sep_token": "[SEP]",
|
| 53 |
+
"sp_model_kwargs": {},
|
| 54 |
+
"split_by_punct": false,
|
| 55 |
+
"tokenizer_class": "DebertaV2Tokenizer",
|
| 56 |
+
"unk_token": "[UNK]",
|
| 57 |
+
"vocab_type": "spm"
|
| 58 |
+
}
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:270da9b099b2d02a736480bd012939e9b5f02ea692f21c4e08c4925dc1f458dc
|
| 3 |
+
size 5048
|
training_params.json
ADDED
|
@@ -0,0 +1,28 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"data_path": "autotrain-m96nh-snymb/autotrain-data",
|
| 3 |
+
"model": "microsoft/deberta-v3-large",
|
| 4 |
+
"lr": 2e-06,
|
| 5 |
+
"epochs": 3,
|
| 6 |
+
"max_seq_length": 1024,
|
| 7 |
+
"batch_size": 2,
|
| 8 |
+
"warmup_ratio": 0.1,
|
| 9 |
+
"gradient_accumulation": 1,
|
| 10 |
+
"optimizer": "adamw_torch",
|
| 11 |
+
"scheduler": "linear",
|
| 12 |
+
"weight_decay": 0.0,
|
| 13 |
+
"max_grad_norm": 1.0,
|
| 14 |
+
"seed": 42,
|
| 15 |
+
"train_split": "train",
|
| 16 |
+
"valid_split": "validation",
|
| 17 |
+
"text_column": "autotrain_text",
|
| 18 |
+
"target_column": "autotrain_label",
|
| 19 |
+
"logging_steps": -1,
|
| 20 |
+
"project_name": "autotrain-m96nh-snymb",
|
| 21 |
+
"auto_find_batch_size": false,
|
| 22 |
+
"mixed_precision": "fp16",
|
| 23 |
+
"save_total_limit": 1,
|
| 24 |
+
"push_to_hub": true,
|
| 25 |
+
"evaluation_strategy": "epoch",
|
| 26 |
+
"username": "abhishek",
|
| 27 |
+
"log": "tensorboard"
|
| 28 |
+
}
|