Upload folder using huggingface_hub
Browse files- README.md +26 -0
- checkpoint-5660/config.json +34 -0
- checkpoint-5660/model.safetensors +3 -0
- checkpoint-5660/optimizer.pt +3 -0
- checkpoint-5660/rng_state_0.pth +3 -0
- checkpoint-5660/rng_state_1.pth +3 -0
- checkpoint-5660/scheduler.pt +3 -0
- checkpoint-5660/trainer_state.json +1754 -0
- checkpoint-5660/training_args.bin +3 -0
- config.json +34 -0
- model.safetensors +3 -0
- runs/Oct23_07-17-42_bora-Lambda-Vector/events.out.tfevents.1729693064.bora-Lambda-Vector.288568.0 +2 -2
- runs/Oct23_07-17-42_bora-Lambda-Vector/events.out.tfevents.1729693410.bora-Lambda-Vector.288568.1 +3 -0
- special_tokens_map.json +7 -0
- tokenizer.json +0 -0
- tokenizer_config.json +55 -0
- training_args.bin +3 -0
- training_params.json +30 -0
- vocab.txt +0 -0
README.md
ADDED
|
@@ -0,0 +1,26 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
|
| 2 |
+
---
|
| 3 |
+
tags:
|
| 4 |
+
- autotrain
|
| 5 |
+
- text-regression
|
| 6 |
+
base_model: google-bert/bert-base-uncased
|
| 7 |
+
widget:
|
| 8 |
+
- text: "I love AutoTrain"
|
| 9 |
+
---
|
| 10 |
+
|
| 11 |
+
# Model Trained Using AutoTrain
|
| 12 |
+
|
| 13 |
+
- Problem type: Text Regression
|
| 14 |
+
|
| 15 |
+
## Validation Metrics
|
| 16 |
+
loss: 2.4403305053710938
|
| 17 |
+
|
| 18 |
+
mse: 2.4293971061706543
|
| 19 |
+
|
| 20 |
+
mae: 1.1220409870147705
|
| 21 |
+
|
| 22 |
+
r2: 0.6057301759719849
|
| 23 |
+
|
| 24 |
+
rmse: 1.5586522817611694
|
| 25 |
+
|
| 26 |
+
explained_variance: 0.631534218788147
|
checkpoint-5660/config.json
ADDED
|
@@ -0,0 +1,34 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "google-bert/bert-base-uncased",
|
| 3 |
+
"_num_labels": 1,
|
| 4 |
+
"architectures": [
|
| 5 |
+
"BertForSequenceClassification"
|
| 6 |
+
],
|
| 7 |
+
"attention_probs_dropout_prob": 0.1,
|
| 8 |
+
"classifier_dropout": null,
|
| 9 |
+
"gradient_checkpointing": false,
|
| 10 |
+
"hidden_act": "gelu",
|
| 11 |
+
"hidden_dropout_prob": 0.1,
|
| 12 |
+
"hidden_size": 768,
|
| 13 |
+
"id2label": {
|
| 14 |
+
"0": "target"
|
| 15 |
+
},
|
| 16 |
+
"initializer_range": 0.02,
|
| 17 |
+
"intermediate_size": 3072,
|
| 18 |
+
"label2id": {
|
| 19 |
+
"target": 0
|
| 20 |
+
},
|
| 21 |
+
"layer_norm_eps": 1e-12,
|
| 22 |
+
"max_position_embeddings": 512,
|
| 23 |
+
"model_type": "bert",
|
| 24 |
+
"num_attention_heads": 12,
|
| 25 |
+
"num_hidden_layers": 12,
|
| 26 |
+
"pad_token_id": 0,
|
| 27 |
+
"position_embedding_type": "absolute",
|
| 28 |
+
"problem_type": "regression",
|
| 29 |
+
"torch_dtype": "float32",
|
| 30 |
+
"transformers_version": "4.45.0",
|
| 31 |
+
"type_vocab_size": 2,
|
| 32 |
+
"use_cache": true,
|
| 33 |
+
"vocab_size": 30522
|
| 34 |
+
}
|
checkpoint-5660/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f221ffc1102181fde09e11c8c654b2bd8537a14f574df593c45fc36c48e44544
|
| 3 |
+
size 437955572
|
checkpoint-5660/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:b8cb3dec7b7df283088a48f1806815d504d08cd8568c2139c2155a3151f46cc4
|
| 3 |
+
size 876032250
|
checkpoint-5660/rng_state_0.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:0799d8c1b975c7a247ba1488bb2883a7e5bcb9bd836136ecea78f524597c0271
|
| 3 |
+
size 14512
|
checkpoint-5660/rng_state_1.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c0e44fe1f480431433720c71779290f9272fbb0798af9bc57da2205d6f2b5a13
|
| 3 |
+
size 14512
|
checkpoint-5660/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:cb4e29667faa0ed54a12c8f85af04b44c4334f2951f41cc6dfaa7cd556172db1
|
| 3 |
+
size 1064
|
checkpoint-5660/trainer_state.json
ADDED
|
@@ -0,0 +1,1754 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 2.4403305053710938,
|
| 3 |
+
"best_model_checkpoint": "QaagiFilter5/checkpoint-5660",
|
| 4 |
+
"epoch": 10.0,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 5660,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.044169611307420496,
|
| 13 |
+
"grad_norm": 172.7814178466797,
|
| 14 |
+
"learning_rate": 1.678445229681979e-06,
|
| 15 |
+
"loss": 26.1479,
|
| 16 |
+
"step": 25
|
| 17 |
+
},
|
| 18 |
+
{
|
| 19 |
+
"epoch": 0.08833922261484099,
|
| 20 |
+
"grad_norm": 166.59046936035156,
|
| 21 |
+
"learning_rate": 3.886925795053004e-06,
|
| 22 |
+
"loss": 23.212,
|
| 23 |
+
"step": 50
|
| 24 |
+
},
|
| 25 |
+
{
|
| 26 |
+
"epoch": 0.13250883392226148,
|
| 27 |
+
"grad_norm": 128.4106903076172,
|
| 28 |
+
"learning_rate": 6.095406360424029e-06,
|
| 29 |
+
"loss": 18.6127,
|
| 30 |
+
"step": 75
|
| 31 |
+
},
|
| 32 |
+
{
|
| 33 |
+
"epoch": 0.17667844522968199,
|
| 34 |
+
"grad_norm": 92.6707534790039,
|
| 35 |
+
"learning_rate": 8.303886925795053e-06,
|
| 36 |
+
"loss": 11.4991,
|
| 37 |
+
"step": 100
|
| 38 |
+
},
|
| 39 |
+
{
|
| 40 |
+
"epoch": 0.22084805653710246,
|
| 41 |
+
"grad_norm": 27.35597801208496,
|
| 42 |
+
"learning_rate": 1.0512367491166077e-05,
|
| 43 |
+
"loss": 7.0239,
|
| 44 |
+
"step": 125
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.26501766784452296,
|
| 48 |
+
"grad_norm": 71.67853546142578,
|
| 49 |
+
"learning_rate": 1.2720848056537101e-05,
|
| 50 |
+
"loss": 5.5043,
|
| 51 |
+
"step": 150
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.30918727915194344,
|
| 55 |
+
"grad_norm": 37.85909652709961,
|
| 56 |
+
"learning_rate": 1.4929328621908128e-05,
|
| 57 |
+
"loss": 4.803,
|
| 58 |
+
"step": 175
|
| 59 |
+
},
|
| 60 |
+
{
|
| 61 |
+
"epoch": 0.35335689045936397,
|
| 62 |
+
"grad_norm": 73.24124908447266,
|
| 63 |
+
"learning_rate": 1.7137809187279152e-05,
|
| 64 |
+
"loss": 4.7082,
|
| 65 |
+
"step": 200
|
| 66 |
+
},
|
| 67 |
+
{
|
| 68 |
+
"epoch": 0.39752650176678445,
|
| 69 |
+
"grad_norm": 67.75065612792969,
|
| 70 |
+
"learning_rate": 1.9346289752650175e-05,
|
| 71 |
+
"loss": 4.2305,
|
| 72 |
+
"step": 225
|
| 73 |
+
},
|
| 74 |
+
{
|
| 75 |
+
"epoch": 0.4416961130742049,
|
| 76 |
+
"grad_norm": 38.89707565307617,
|
| 77 |
+
"learning_rate": 2.1554770318021204e-05,
|
| 78 |
+
"loss": 3.3087,
|
| 79 |
+
"step": 250
|
| 80 |
+
},
|
| 81 |
+
{
|
| 82 |
+
"epoch": 0.48586572438162545,
|
| 83 |
+
"grad_norm": 46.586524963378906,
|
| 84 |
+
"learning_rate": 2.3763250883392226e-05,
|
| 85 |
+
"loss": 4.019,
|
| 86 |
+
"step": 275
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.5300353356890459,
|
| 90 |
+
"grad_norm": 41.0096435546875,
|
| 91 |
+
"learning_rate": 2.5971731448763255e-05,
|
| 92 |
+
"loss": 3.8054,
|
| 93 |
+
"step": 300
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 0.5742049469964664,
|
| 97 |
+
"grad_norm": 73.49757385253906,
|
| 98 |
+
"learning_rate": 2.8180212014134277e-05,
|
| 99 |
+
"loss": 3.6329,
|
| 100 |
+
"step": 325
|
| 101 |
+
},
|
| 102 |
+
{
|
| 103 |
+
"epoch": 0.6183745583038869,
|
| 104 |
+
"grad_norm": 62.597286224365234,
|
| 105 |
+
"learning_rate": 3.03886925795053e-05,
|
| 106 |
+
"loss": 3.6779,
|
| 107 |
+
"step": 350
|
| 108 |
+
},
|
| 109 |
+
{
|
| 110 |
+
"epoch": 0.6625441696113075,
|
| 111 |
+
"grad_norm": 35.01760482788086,
|
| 112 |
+
"learning_rate": 3.2597173144876325e-05,
|
| 113 |
+
"loss": 3.808,
|
| 114 |
+
"step": 375
|
| 115 |
+
},
|
| 116 |
+
{
|
| 117 |
+
"epoch": 0.7067137809187279,
|
| 118 |
+
"grad_norm": 33.265625,
|
| 119 |
+
"learning_rate": 3.480565371024735e-05,
|
| 120 |
+
"loss": 3.2528,
|
| 121 |
+
"step": 400
|
| 122 |
+
},
|
| 123 |
+
{
|
| 124 |
+
"epoch": 0.7508833922261484,
|
| 125 |
+
"grad_norm": 41.54469299316406,
|
| 126 |
+
"learning_rate": 3.7014134275618377e-05,
|
| 127 |
+
"loss": 3.3373,
|
| 128 |
+
"step": 425
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.7950530035335689,
|
| 132 |
+
"grad_norm": 59.207889556884766,
|
| 133 |
+
"learning_rate": 3.9222614840989406e-05,
|
| 134 |
+
"loss": 2.716,
|
| 135 |
+
"step": 450
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 0.8392226148409894,
|
| 139 |
+
"grad_norm": 37.75099182128906,
|
| 140 |
+
"learning_rate": 4.143109540636043e-05,
|
| 141 |
+
"loss": 3.174,
|
| 142 |
+
"step": 475
|
| 143 |
+
},
|
| 144 |
+
{
|
| 145 |
+
"epoch": 0.8833922261484098,
|
| 146 |
+
"grad_norm": 44.53806686401367,
|
| 147 |
+
"learning_rate": 4.363957597173145e-05,
|
| 148 |
+
"loss": 3.6122,
|
| 149 |
+
"step": 500
|
| 150 |
+
},
|
| 151 |
+
{
|
| 152 |
+
"epoch": 0.9275618374558304,
|
| 153 |
+
"grad_norm": 36.89043426513672,
|
| 154 |
+
"learning_rate": 4.584805653710247e-05,
|
| 155 |
+
"loss": 3.0317,
|
| 156 |
+
"step": 525
|
| 157 |
+
},
|
| 158 |
+
{
|
| 159 |
+
"epoch": 0.9717314487632509,
|
| 160 |
+
"grad_norm": 75.27375030517578,
|
| 161 |
+
"learning_rate": 4.8056537102473495e-05,
|
| 162 |
+
"loss": 2.3758,
|
| 163 |
+
"step": 550
|
| 164 |
+
},
|
| 165 |
+
{
|
| 166 |
+
"epoch": 1.0,
|
| 167 |
+
"eval_explained_variance": 0.5094610452651978,
|
| 168 |
+
"eval_loss": 3.465508222579956,
|
| 169 |
+
"eval_mae": 1.3762876987457275,
|
| 170 |
+
"eval_mse": 3.4517085552215576,
|
| 171 |
+
"eval_r2": 0.4398180842399597,
|
| 172 |
+
"eval_rmse": 1.8578773736953735,
|
| 173 |
+
"eval_runtime": 1.0216,
|
| 174 |
+
"eval_samples_per_second": 2215.123,
|
| 175 |
+
"eval_steps_per_second": 69.498,
|
| 176 |
+
"step": 566
|
| 177 |
+
},
|
| 178 |
+
{
|
| 179 |
+
"epoch": 1.0159010600706713,
|
| 180 |
+
"grad_norm": 26.423126220703125,
|
| 181 |
+
"learning_rate": 4.997055359246172e-05,
|
| 182 |
+
"loss": 2.7841,
|
| 183 |
+
"step": 575
|
| 184 |
+
},
|
| 185 |
+
{
|
| 186 |
+
"epoch": 1.0600706713780919,
|
| 187 |
+
"grad_norm": 69.90520477294922,
|
| 188 |
+
"learning_rate": 4.9725166862976055e-05,
|
| 189 |
+
"loss": 2.1612,
|
| 190 |
+
"step": 600
|
| 191 |
+
},
|
| 192 |
+
{
|
| 193 |
+
"epoch": 1.1042402826855124,
|
| 194 |
+
"grad_norm": 33.07378387451172,
|
| 195 |
+
"learning_rate": 4.9479780133490385e-05,
|
| 196 |
+
"loss": 2.48,
|
| 197 |
+
"step": 625
|
| 198 |
+
},
|
| 199 |
+
{
|
| 200 |
+
"epoch": 1.1484098939929328,
|
| 201 |
+
"grad_norm": 17.648529052734375,
|
| 202 |
+
"learning_rate": 4.9234393404004716e-05,
|
| 203 |
+
"loss": 2.0749,
|
| 204 |
+
"step": 650
|
| 205 |
+
},
|
| 206 |
+
{
|
| 207 |
+
"epoch": 1.1925795053003534,
|
| 208 |
+
"grad_norm": 53.43230438232422,
|
| 209 |
+
"learning_rate": 4.8989006674519046e-05,
|
| 210 |
+
"loss": 2.1101,
|
| 211 |
+
"step": 675
|
| 212 |
+
},
|
| 213 |
+
{
|
| 214 |
+
"epoch": 1.2367491166077738,
|
| 215 |
+
"grad_norm": 35.27627944946289,
|
| 216 |
+
"learning_rate": 4.8743619945033376e-05,
|
| 217 |
+
"loss": 2.2126,
|
| 218 |
+
"step": 700
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 1.2809187279151943,
|
| 222 |
+
"grad_norm": 50.328914642333984,
|
| 223 |
+
"learning_rate": 4.8498233215547706e-05,
|
| 224 |
+
"loss": 2.0551,
|
| 225 |
+
"step": 725
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"epoch": 1.325088339222615,
|
| 229 |
+
"grad_norm": 52.285911560058594,
|
| 230 |
+
"learning_rate": 4.8252846486062036e-05,
|
| 231 |
+
"loss": 1.9554,
|
| 232 |
+
"step": 750
|
| 233 |
+
},
|
| 234 |
+
{
|
| 235 |
+
"epoch": 1.3692579505300353,
|
| 236 |
+
"grad_norm": 43.46308135986328,
|
| 237 |
+
"learning_rate": 4.8007459756576366e-05,
|
| 238 |
+
"loss": 2.3065,
|
| 239 |
+
"step": 775
|
| 240 |
+
},
|
| 241 |
+
{
|
| 242 |
+
"epoch": 1.4134275618374559,
|
| 243 |
+
"grad_norm": 50.10768508911133,
|
| 244 |
+
"learning_rate": 4.7762073027090696e-05,
|
| 245 |
+
"loss": 2.2778,
|
| 246 |
+
"step": 800
|
| 247 |
+
},
|
| 248 |
+
{
|
| 249 |
+
"epoch": 1.4575971731448762,
|
| 250 |
+
"grad_norm": 32.7845344543457,
|
| 251 |
+
"learning_rate": 4.7516686297605026e-05,
|
| 252 |
+
"loss": 2.0533,
|
| 253 |
+
"step": 825
|
| 254 |
+
},
|
| 255 |
+
{
|
| 256 |
+
"epoch": 1.5017667844522968,
|
| 257 |
+
"grad_norm": 24.879047393798828,
|
| 258 |
+
"learning_rate": 4.7271299568119356e-05,
|
| 259 |
+
"loss": 2.1411,
|
| 260 |
+
"step": 850
|
| 261 |
+
},
|
| 262 |
+
{
|
| 263 |
+
"epoch": 1.5459363957597172,
|
| 264 |
+
"grad_norm": 44.34259796142578,
|
| 265 |
+
"learning_rate": 4.702591283863369e-05,
|
| 266 |
+
"loss": 2.2567,
|
| 267 |
+
"step": 875
|
| 268 |
+
},
|
| 269 |
+
{
|
| 270 |
+
"epoch": 1.5901060070671378,
|
| 271 |
+
"grad_norm": 59.14316940307617,
|
| 272 |
+
"learning_rate": 4.6780526109148023e-05,
|
| 273 |
+
"loss": 1.8269,
|
| 274 |
+
"step": 900
|
| 275 |
+
},
|
| 276 |
+
{
|
| 277 |
+
"epoch": 1.6342756183745584,
|
| 278 |
+
"grad_norm": 41.258827209472656,
|
| 279 |
+
"learning_rate": 4.6535139379662354e-05,
|
| 280 |
+
"loss": 2.0853,
|
| 281 |
+
"step": 925
|
| 282 |
+
},
|
| 283 |
+
{
|
| 284 |
+
"epoch": 1.6784452296819787,
|
| 285 |
+
"grad_norm": 22.546058654785156,
|
| 286 |
+
"learning_rate": 4.6289752650176684e-05,
|
| 287 |
+
"loss": 2.1658,
|
| 288 |
+
"step": 950
|
| 289 |
+
},
|
| 290 |
+
{
|
| 291 |
+
"epoch": 1.7226148409893993,
|
| 292 |
+
"grad_norm": 52.796470642089844,
|
| 293 |
+
"learning_rate": 4.6044365920691014e-05,
|
| 294 |
+
"loss": 2.367,
|
| 295 |
+
"step": 975
|
| 296 |
+
},
|
| 297 |
+
{
|
| 298 |
+
"epoch": 1.76678445229682,
|
| 299 |
+
"grad_norm": 29.541540145874023,
|
| 300 |
+
"learning_rate": 4.5798979191205344e-05,
|
| 301 |
+
"loss": 2.1246,
|
| 302 |
+
"step": 1000
|
| 303 |
+
},
|
| 304 |
+
{
|
| 305 |
+
"epoch": 1.8109540636042403,
|
| 306 |
+
"grad_norm": 40.63389587402344,
|
| 307 |
+
"learning_rate": 4.5553592461719674e-05,
|
| 308 |
+
"loss": 2.0685,
|
| 309 |
+
"step": 1025
|
| 310 |
+
},
|
| 311 |
+
{
|
| 312 |
+
"epoch": 1.8551236749116606,
|
| 313 |
+
"grad_norm": 34.64159393310547,
|
| 314 |
+
"learning_rate": 4.5308205732234004e-05,
|
| 315 |
+
"loss": 2.1979,
|
| 316 |
+
"step": 1050
|
| 317 |
+
},
|
| 318 |
+
{
|
| 319 |
+
"epoch": 1.8992932862190812,
|
| 320 |
+
"grad_norm": 27.459083557128906,
|
| 321 |
+
"learning_rate": 4.5062819002748334e-05,
|
| 322 |
+
"loss": 1.9735,
|
| 323 |
+
"step": 1075
|
| 324 |
+
},
|
| 325 |
+
{
|
| 326 |
+
"epoch": 1.9434628975265018,
|
| 327 |
+
"grad_norm": 25.363712310791016,
|
| 328 |
+
"learning_rate": 4.4817432273262664e-05,
|
| 329 |
+
"loss": 2.0339,
|
| 330 |
+
"step": 1100
|
| 331 |
+
},
|
| 332 |
+
{
|
| 333 |
+
"epoch": 1.9876325088339222,
|
| 334 |
+
"grad_norm": 45.697757720947266,
|
| 335 |
+
"learning_rate": 4.4572045543776994e-05,
|
| 336 |
+
"loss": 1.967,
|
| 337 |
+
"step": 1125
|
| 338 |
+
},
|
| 339 |
+
{
|
| 340 |
+
"epoch": 2.0,
|
| 341 |
+
"eval_explained_variance": 0.5648437142372131,
|
| 342 |
+
"eval_loss": 3.2432901859283447,
|
| 343 |
+
"eval_mae": 1.3005132675170898,
|
| 344 |
+
"eval_mse": 3.227083444595337,
|
| 345 |
+
"eval_r2": 0.4762727618217468,
|
| 346 |
+
"eval_rmse": 1.7964085340499878,
|
| 347 |
+
"eval_runtime": 1.0461,
|
| 348 |
+
"eval_samples_per_second": 2163.35,
|
| 349 |
+
"eval_steps_per_second": 67.874,
|
| 350 |
+
"step": 1132
|
| 351 |
+
},
|
| 352 |
+
{
|
| 353 |
+
"epoch": 2.0318021201413425,
|
| 354 |
+
"grad_norm": 44.7116813659668,
|
| 355 |
+
"learning_rate": 4.4326658814291324e-05,
|
| 356 |
+
"loss": 1.2539,
|
| 357 |
+
"step": 1150
|
| 358 |
+
},
|
| 359 |
+
{
|
| 360 |
+
"epoch": 2.0759717314487633,
|
| 361 |
+
"grad_norm": 25.487625122070312,
|
| 362 |
+
"learning_rate": 4.408127208480566e-05,
|
| 363 |
+
"loss": 1.1147,
|
| 364 |
+
"step": 1175
|
| 365 |
+
},
|
| 366 |
+
{
|
| 367 |
+
"epoch": 2.1201413427561837,
|
| 368 |
+
"grad_norm": 27.024295806884766,
|
| 369 |
+
"learning_rate": 4.383588535531999e-05,
|
| 370 |
+
"loss": 0.8406,
|
| 371 |
+
"step": 1200
|
| 372 |
+
},
|
| 373 |
+
{
|
| 374 |
+
"epoch": 2.164310954063604,
|
| 375 |
+
"grad_norm": 21.068811416625977,
|
| 376 |
+
"learning_rate": 4.359049862583432e-05,
|
| 377 |
+
"loss": 1.0608,
|
| 378 |
+
"step": 1225
|
| 379 |
+
},
|
| 380 |
+
{
|
| 381 |
+
"epoch": 2.208480565371025,
|
| 382 |
+
"grad_norm": 17.862035751342773,
|
| 383 |
+
"learning_rate": 4.334511189634865e-05,
|
| 384 |
+
"loss": 1.1741,
|
| 385 |
+
"step": 1250
|
| 386 |
+
},
|
| 387 |
+
{
|
| 388 |
+
"epoch": 2.2526501766784452,
|
| 389 |
+
"grad_norm": 44.87358856201172,
|
| 390 |
+
"learning_rate": 4.3099725166862975e-05,
|
| 391 |
+
"loss": 0.8678,
|
| 392 |
+
"step": 1275
|
| 393 |
+
},
|
| 394 |
+
{
|
| 395 |
+
"epoch": 2.2968197879858656,
|
| 396 |
+
"grad_norm": 28.953048706054688,
|
| 397 |
+
"learning_rate": 4.2854338437377305e-05,
|
| 398 |
+
"loss": 0.8326,
|
| 399 |
+
"step": 1300
|
| 400 |
+
},
|
| 401 |
+
{
|
| 402 |
+
"epoch": 2.340989399293286,
|
| 403 |
+
"grad_norm": 32.61209487915039,
|
| 404 |
+
"learning_rate": 4.2608951707891635e-05,
|
| 405 |
+
"loss": 0.9226,
|
| 406 |
+
"step": 1325
|
| 407 |
+
},
|
| 408 |
+
{
|
| 409 |
+
"epoch": 2.385159010600707,
|
| 410 |
+
"grad_norm": 24.571086883544922,
|
| 411 |
+
"learning_rate": 4.2363564978405965e-05,
|
| 412 |
+
"loss": 0.8819,
|
| 413 |
+
"step": 1350
|
| 414 |
+
},
|
| 415 |
+
{
|
| 416 |
+
"epoch": 2.429328621908127,
|
| 417 |
+
"grad_norm": 23.368595123291016,
|
| 418 |
+
"learning_rate": 4.2118178248920295e-05,
|
| 419 |
+
"loss": 0.8845,
|
| 420 |
+
"step": 1375
|
| 421 |
+
},
|
| 422 |
+
{
|
| 423 |
+
"epoch": 2.4734982332155475,
|
| 424 |
+
"grad_norm": 35.36629867553711,
|
| 425 |
+
"learning_rate": 4.1872791519434626e-05,
|
| 426 |
+
"loss": 1.0751,
|
| 427 |
+
"step": 1400
|
| 428 |
+
},
|
| 429 |
+
{
|
| 430 |
+
"epoch": 2.5176678445229683,
|
| 431 |
+
"grad_norm": 15.506020545959473,
|
| 432 |
+
"learning_rate": 4.162740478994896e-05,
|
| 433 |
+
"loss": 0.933,
|
| 434 |
+
"step": 1425
|
| 435 |
+
},
|
| 436 |
+
{
|
| 437 |
+
"epoch": 2.5618374558303887,
|
| 438 |
+
"grad_norm": 40.057884216308594,
|
| 439 |
+
"learning_rate": 4.138201806046329e-05,
|
| 440 |
+
"loss": 0.9503,
|
| 441 |
+
"step": 1450
|
| 442 |
+
},
|
| 443 |
+
{
|
| 444 |
+
"epoch": 2.606007067137809,
|
| 445 |
+
"grad_norm": 20.372051239013672,
|
| 446 |
+
"learning_rate": 4.113663133097762e-05,
|
| 447 |
+
"loss": 0.8741,
|
| 448 |
+
"step": 1475
|
| 449 |
+
},
|
| 450 |
+
{
|
| 451 |
+
"epoch": 2.65017667844523,
|
| 452 |
+
"grad_norm": 10.090834617614746,
|
| 453 |
+
"learning_rate": 4.089124460149195e-05,
|
| 454 |
+
"loss": 0.9391,
|
| 455 |
+
"step": 1500
|
| 456 |
+
},
|
| 457 |
+
{
|
| 458 |
+
"epoch": 2.6943462897526502,
|
| 459 |
+
"grad_norm": 14.696582794189453,
|
| 460 |
+
"learning_rate": 4.064585787200628e-05,
|
| 461 |
+
"loss": 0.9666,
|
| 462 |
+
"step": 1525
|
| 463 |
+
},
|
| 464 |
+
{
|
| 465 |
+
"epoch": 2.7385159010600706,
|
| 466 |
+
"grad_norm": 31.863920211791992,
|
| 467 |
+
"learning_rate": 4.040047114252061e-05,
|
| 468 |
+
"loss": 1.0567,
|
| 469 |
+
"step": 1550
|
| 470 |
+
},
|
| 471 |
+
{
|
| 472 |
+
"epoch": 2.7826855123674914,
|
| 473 |
+
"grad_norm": 36.8994255065918,
|
| 474 |
+
"learning_rate": 4.015508441303494e-05,
|
| 475 |
+
"loss": 1.0003,
|
| 476 |
+
"step": 1575
|
| 477 |
+
},
|
| 478 |
+
{
|
| 479 |
+
"epoch": 2.8268551236749118,
|
| 480 |
+
"grad_norm": 21.43227195739746,
|
| 481 |
+
"learning_rate": 3.990969768354927e-05,
|
| 482 |
+
"loss": 1.0532,
|
| 483 |
+
"step": 1600
|
| 484 |
+
},
|
| 485 |
+
{
|
| 486 |
+
"epoch": 2.871024734982332,
|
| 487 |
+
"grad_norm": 35.63019561767578,
|
| 488 |
+
"learning_rate": 3.96643109540636e-05,
|
| 489 |
+
"loss": 0.8816,
|
| 490 |
+
"step": 1625
|
| 491 |
+
},
|
| 492 |
+
{
|
| 493 |
+
"epoch": 2.9151943462897525,
|
| 494 |
+
"grad_norm": 9.274225234985352,
|
| 495 |
+
"learning_rate": 3.941892422457793e-05,
|
| 496 |
+
"loss": 1.0889,
|
| 497 |
+
"step": 1650
|
| 498 |
+
},
|
| 499 |
+
{
|
| 500 |
+
"epoch": 2.9593639575971733,
|
| 501 |
+
"grad_norm": 22.54602813720703,
|
| 502 |
+
"learning_rate": 3.9173537495092263e-05,
|
| 503 |
+
"loss": 0.9442,
|
| 504 |
+
"step": 1675
|
| 505 |
+
},
|
| 506 |
+
{
|
| 507 |
+
"epoch": 3.0,
|
| 508 |
+
"eval_explained_variance": 0.5853231549263,
|
| 509 |
+
"eval_loss": 2.897214651107788,
|
| 510 |
+
"eval_mae": 1.210811972618103,
|
| 511 |
+
"eval_mse": 2.8827266693115234,
|
| 512 |
+
"eval_r2": 0.5321588516235352,
|
| 513 |
+
"eval_rmse": 1.6978594064712524,
|
| 514 |
+
"eval_runtime": 1.0405,
|
| 515 |
+
"eval_samples_per_second": 2174.844,
|
| 516 |
+
"eval_steps_per_second": 68.234,
|
| 517 |
+
"step": 1698
|
| 518 |
+
},
|
| 519 |
+
{
|
| 520 |
+
"epoch": 3.0035335689045937,
|
| 521 |
+
"grad_norm": 20.036935806274414,
|
| 522 |
+
"learning_rate": 3.8928150765606594e-05,
|
| 523 |
+
"loss": 0.9184,
|
| 524 |
+
"step": 1700
|
| 525 |
+
},
|
| 526 |
+
{
|
| 527 |
+
"epoch": 3.047703180212014,
|
| 528 |
+
"grad_norm": 34.86088180541992,
|
| 529 |
+
"learning_rate": 3.868276403612093e-05,
|
| 530 |
+
"loss": 0.5967,
|
| 531 |
+
"step": 1725
|
| 532 |
+
},
|
| 533 |
+
{
|
| 534 |
+
"epoch": 3.091872791519435,
|
| 535 |
+
"grad_norm": 19.049938201904297,
|
| 536 |
+
"learning_rate": 3.843737730663526e-05,
|
| 537 |
+
"loss": 0.5294,
|
| 538 |
+
"step": 1750
|
| 539 |
+
},
|
| 540 |
+
{
|
| 541 |
+
"epoch": 3.136042402826855,
|
| 542 |
+
"grad_norm": 12.164793014526367,
|
| 543 |
+
"learning_rate": 3.819199057714959e-05,
|
| 544 |
+
"loss": 0.6298,
|
| 545 |
+
"step": 1775
|
| 546 |
+
},
|
| 547 |
+
{
|
| 548 |
+
"epoch": 3.1802120141342756,
|
| 549 |
+
"grad_norm": 34.6705207824707,
|
| 550 |
+
"learning_rate": 3.794660384766392e-05,
|
| 551 |
+
"loss": 0.6278,
|
| 552 |
+
"step": 1800
|
| 553 |
+
},
|
| 554 |
+
{
|
| 555 |
+
"epoch": 3.224381625441696,
|
| 556 |
+
"grad_norm": 14.351198196411133,
|
| 557 |
+
"learning_rate": 3.770121711817825e-05,
|
| 558 |
+
"loss": 0.5648,
|
| 559 |
+
"step": 1825
|
| 560 |
+
},
|
| 561 |
+
{
|
| 562 |
+
"epoch": 3.2685512367491167,
|
| 563 |
+
"grad_norm": 30.73033905029297,
|
| 564 |
+
"learning_rate": 3.745583038869258e-05,
|
| 565 |
+
"loss": 0.6436,
|
| 566 |
+
"step": 1850
|
| 567 |
+
},
|
| 568 |
+
{
|
| 569 |
+
"epoch": 3.312720848056537,
|
| 570 |
+
"grad_norm": 22.45409393310547,
|
| 571 |
+
"learning_rate": 3.721044365920691e-05,
|
| 572 |
+
"loss": 0.6032,
|
| 573 |
+
"step": 1875
|
| 574 |
+
},
|
| 575 |
+
{
|
| 576 |
+
"epoch": 3.3568904593639575,
|
| 577 |
+
"grad_norm": 12.688070297241211,
|
| 578 |
+
"learning_rate": 3.696505692972124e-05,
|
| 579 |
+
"loss": 0.6173,
|
| 580 |
+
"step": 1900
|
| 581 |
+
},
|
| 582 |
+
{
|
| 583 |
+
"epoch": 3.4010600706713783,
|
| 584 |
+
"grad_norm": 21.9694766998291,
|
| 585 |
+
"learning_rate": 3.671967020023557e-05,
|
| 586 |
+
"loss": 0.515,
|
| 587 |
+
"step": 1925
|
| 588 |
+
},
|
| 589 |
+
{
|
| 590 |
+
"epoch": 3.4452296819787986,
|
| 591 |
+
"grad_norm": 11.734025001525879,
|
| 592 |
+
"learning_rate": 3.64742834707499e-05,
|
| 593 |
+
"loss": 0.5558,
|
| 594 |
+
"step": 1950
|
| 595 |
+
},
|
| 596 |
+
{
|
| 597 |
+
"epoch": 3.489399293286219,
|
| 598 |
+
"grad_norm": 22.58125877380371,
|
| 599 |
+
"learning_rate": 3.622889674126423e-05,
|
| 600 |
+
"loss": 0.5682,
|
| 601 |
+
"step": 1975
|
| 602 |
+
},
|
| 603 |
+
{
|
| 604 |
+
"epoch": 3.53356890459364,
|
| 605 |
+
"grad_norm": 19.323246002197266,
|
| 606 |
+
"learning_rate": 3.598351001177857e-05,
|
| 607 |
+
"loss": 0.5542,
|
| 608 |
+
"step": 2000
|
| 609 |
+
},
|
| 610 |
+
{
|
| 611 |
+
"epoch": 3.57773851590106,
|
| 612 |
+
"grad_norm": 18.396697998046875,
|
| 613 |
+
"learning_rate": 3.57381232822929e-05,
|
| 614 |
+
"loss": 0.6233,
|
| 615 |
+
"step": 2025
|
| 616 |
+
},
|
| 617 |
+
{
|
| 618 |
+
"epoch": 3.6219081272084805,
|
| 619 |
+
"grad_norm": 29.500776290893555,
|
| 620 |
+
"learning_rate": 3.549273655280723e-05,
|
| 621 |
+
"loss": 0.6947,
|
| 622 |
+
"step": 2050
|
| 623 |
+
},
|
| 624 |
+
{
|
| 625 |
+
"epoch": 3.666077738515901,
|
| 626 |
+
"grad_norm": 14.776251792907715,
|
| 627 |
+
"learning_rate": 3.524734982332156e-05,
|
| 628 |
+
"loss": 0.5729,
|
| 629 |
+
"step": 2075
|
| 630 |
+
},
|
| 631 |
+
{
|
| 632 |
+
"epoch": 3.7102473498233217,
|
| 633 |
+
"grad_norm": 30.837646484375,
|
| 634 |
+
"learning_rate": 3.500196309383589e-05,
|
| 635 |
+
"loss": 0.669,
|
| 636 |
+
"step": 2100
|
| 637 |
+
},
|
| 638 |
+
{
|
| 639 |
+
"epoch": 3.754416961130742,
|
| 640 |
+
"grad_norm": 22.036334991455078,
|
| 641 |
+
"learning_rate": 3.475657636435022e-05,
|
| 642 |
+
"loss": 0.5365,
|
| 643 |
+
"step": 2125
|
| 644 |
+
},
|
| 645 |
+
{
|
| 646 |
+
"epoch": 3.7985865724381624,
|
| 647 |
+
"grad_norm": 7.9163079261779785,
|
| 648 |
+
"learning_rate": 3.451118963486455e-05,
|
| 649 |
+
"loss": 0.548,
|
| 650 |
+
"step": 2150
|
| 651 |
+
},
|
| 652 |
+
{
|
| 653 |
+
"epoch": 3.842756183745583,
|
| 654 |
+
"grad_norm": 27.34328842163086,
|
| 655 |
+
"learning_rate": 3.426580290537888e-05,
|
| 656 |
+
"loss": 0.5585,
|
| 657 |
+
"step": 2175
|
| 658 |
+
},
|
| 659 |
+
{
|
| 660 |
+
"epoch": 3.8869257950530036,
|
| 661 |
+
"grad_norm": 20.00174903869629,
|
| 662 |
+
"learning_rate": 3.402041617589321e-05,
|
| 663 |
+
"loss": 0.6514,
|
| 664 |
+
"step": 2200
|
| 665 |
+
},
|
| 666 |
+
{
|
| 667 |
+
"epoch": 3.931095406360424,
|
| 668 |
+
"grad_norm": 15.843454360961914,
|
| 669 |
+
"learning_rate": 3.377502944640754e-05,
|
| 670 |
+
"loss": 0.6258,
|
| 671 |
+
"step": 2225
|
| 672 |
+
},
|
| 673 |
+
{
|
| 674 |
+
"epoch": 3.9752650176678443,
|
| 675 |
+
"grad_norm": 31.284502029418945,
|
| 676 |
+
"learning_rate": 3.352964271692187e-05,
|
| 677 |
+
"loss": 0.5995,
|
| 678 |
+
"step": 2250
|
| 679 |
+
},
|
| 680 |
+
{
|
| 681 |
+
"epoch": 4.0,
|
| 682 |
+
"eval_explained_variance": 0.6056773662567139,
|
| 683 |
+
"eval_loss": 2.727660655975342,
|
| 684 |
+
"eval_mae": 1.1792982816696167,
|
| 685 |
+
"eval_mse": 2.715883255004883,
|
| 686 |
+
"eval_r2": 0.5592360496520996,
|
| 687 |
+
"eval_rmse": 1.6479936838150024,
|
| 688 |
+
"eval_runtime": 1.0286,
|
| 689 |
+
"eval_samples_per_second": 2200.076,
|
| 690 |
+
"eval_steps_per_second": 69.026,
|
| 691 |
+
"step": 2264
|
| 692 |
+
},
|
| 693 |
+
{
|
| 694 |
+
"epoch": 4.019434628975265,
|
| 695 |
+
"grad_norm": 15.487521171569824,
|
| 696 |
+
"learning_rate": 3.32842559874362e-05,
|
| 697 |
+
"loss": 0.5058,
|
| 698 |
+
"step": 2275
|
| 699 |
+
},
|
| 700 |
+
{
|
| 701 |
+
"epoch": 4.063604240282685,
|
| 702 |
+
"grad_norm": 8.639957427978516,
|
| 703 |
+
"learning_rate": 3.3038869257950536e-05,
|
| 704 |
+
"loss": 0.4803,
|
| 705 |
+
"step": 2300
|
| 706 |
+
},
|
| 707 |
+
{
|
| 708 |
+
"epoch": 4.107773851590106,
|
| 709 |
+
"grad_norm": 7.802794456481934,
|
| 710 |
+
"learning_rate": 3.2793482528464867e-05,
|
| 711 |
+
"loss": 0.351,
|
| 712 |
+
"step": 2325
|
| 713 |
+
},
|
| 714 |
+
{
|
| 715 |
+
"epoch": 4.151943462897527,
|
| 716 |
+
"grad_norm": 10.31377124786377,
|
| 717 |
+
"learning_rate": 3.25480957989792e-05,
|
| 718 |
+
"loss": 0.453,
|
| 719 |
+
"step": 2350
|
| 720 |
+
},
|
| 721 |
+
{
|
| 722 |
+
"epoch": 4.196113074204947,
|
| 723 |
+
"grad_norm": 18.428464889526367,
|
| 724 |
+
"learning_rate": 3.230270906949353e-05,
|
| 725 |
+
"loss": 0.4612,
|
| 726 |
+
"step": 2375
|
| 727 |
+
},
|
| 728 |
+
{
|
| 729 |
+
"epoch": 4.240282685512367,
|
| 730 |
+
"grad_norm": 15.8595609664917,
|
| 731 |
+
"learning_rate": 3.205732234000786e-05,
|
| 732 |
+
"loss": 0.3915,
|
| 733 |
+
"step": 2400
|
| 734 |
+
},
|
| 735 |
+
{
|
| 736 |
+
"epoch": 4.284452296819788,
|
| 737 |
+
"grad_norm": 17.2500057220459,
|
| 738 |
+
"learning_rate": 3.181193561052219e-05,
|
| 739 |
+
"loss": 0.3635,
|
| 740 |
+
"step": 2425
|
| 741 |
+
},
|
| 742 |
+
{
|
| 743 |
+
"epoch": 4.328621908127208,
|
| 744 |
+
"grad_norm": 13.921929359436035,
|
| 745 |
+
"learning_rate": 3.156654888103652e-05,
|
| 746 |
+
"loss": 0.3373,
|
| 747 |
+
"step": 2450
|
| 748 |
+
},
|
| 749 |
+
{
|
| 750 |
+
"epoch": 4.372791519434629,
|
| 751 |
+
"grad_norm": 14.69737434387207,
|
| 752 |
+
"learning_rate": 3.132116215155085e-05,
|
| 753 |
+
"loss": 0.4055,
|
| 754 |
+
"step": 2475
|
| 755 |
+
},
|
| 756 |
+
{
|
| 757 |
+
"epoch": 4.41696113074205,
|
| 758 |
+
"grad_norm": 23.352657318115234,
|
| 759 |
+
"learning_rate": 3.107577542206518e-05,
|
| 760 |
+
"loss": 0.4435,
|
| 761 |
+
"step": 2500
|
| 762 |
+
},
|
| 763 |
+
{
|
| 764 |
+
"epoch": 4.46113074204947,
|
| 765 |
+
"grad_norm": 10.377140045166016,
|
| 766 |
+
"learning_rate": 3.083038869257951e-05,
|
| 767 |
+
"loss": 0.3323,
|
| 768 |
+
"step": 2525
|
| 769 |
+
},
|
| 770 |
+
{
|
| 771 |
+
"epoch": 4.5053003533568905,
|
| 772 |
+
"grad_norm": 9.900633811950684,
|
| 773 |
+
"learning_rate": 3.058500196309384e-05,
|
| 774 |
+
"loss": 0.3799,
|
| 775 |
+
"step": 2550
|
| 776 |
+
},
|
| 777 |
+
{
|
| 778 |
+
"epoch": 4.549469964664311,
|
| 779 |
+
"grad_norm": 11.127573013305664,
|
| 780 |
+
"learning_rate": 3.0339615233608164e-05,
|
| 781 |
+
"loss": 0.3736,
|
| 782 |
+
"step": 2575
|
| 783 |
+
},
|
| 784 |
+
{
|
| 785 |
+
"epoch": 4.593639575971731,
|
| 786 |
+
"grad_norm": 16.36042594909668,
|
| 787 |
+
"learning_rate": 3.0094228504122494e-05,
|
| 788 |
+
"loss": 0.4345,
|
| 789 |
+
"step": 2600
|
| 790 |
+
},
|
| 791 |
+
{
|
| 792 |
+
"epoch": 4.637809187279152,
|
| 793 |
+
"grad_norm": 25.30819320678711,
|
| 794 |
+
"learning_rate": 2.9848841774636828e-05,
|
| 795 |
+
"loss": 0.3901,
|
| 796 |
+
"step": 2625
|
| 797 |
+
},
|
| 798 |
+
{
|
| 799 |
+
"epoch": 4.681978798586572,
|
| 800 |
+
"grad_norm": 23.80220603942871,
|
| 801 |
+
"learning_rate": 2.9603455045151158e-05,
|
| 802 |
+
"loss": 0.4355,
|
| 803 |
+
"step": 2650
|
| 804 |
+
},
|
| 805 |
+
{
|
| 806 |
+
"epoch": 4.726148409893993,
|
| 807 |
+
"grad_norm": 11.138365745544434,
|
| 808 |
+
"learning_rate": 2.9358068315665488e-05,
|
| 809 |
+
"loss": 0.3541,
|
| 810 |
+
"step": 2675
|
| 811 |
+
},
|
| 812 |
+
{
|
| 813 |
+
"epoch": 4.770318021201414,
|
| 814 |
+
"grad_norm": 20.883947372436523,
|
| 815 |
+
"learning_rate": 2.9112681586179818e-05,
|
| 816 |
+
"loss": 0.3744,
|
| 817 |
+
"step": 2700
|
| 818 |
+
},
|
| 819 |
+
{
|
| 820 |
+
"epoch": 4.814487632508834,
|
| 821 |
+
"grad_norm": 8.56249713897705,
|
| 822 |
+
"learning_rate": 2.8867294856694148e-05,
|
| 823 |
+
"loss": 0.3599,
|
| 824 |
+
"step": 2725
|
| 825 |
+
},
|
| 826 |
+
{
|
| 827 |
+
"epoch": 4.858657243816254,
|
| 828 |
+
"grad_norm": 20.723438262939453,
|
| 829 |
+
"learning_rate": 2.862190812720848e-05,
|
| 830 |
+
"loss": 0.4961,
|
| 831 |
+
"step": 2750
|
| 832 |
+
},
|
| 833 |
+
{
|
| 834 |
+
"epoch": 4.902826855123675,
|
| 835 |
+
"grad_norm": 10.906144142150879,
|
| 836 |
+
"learning_rate": 2.8376521397722812e-05,
|
| 837 |
+
"loss": 0.3921,
|
| 838 |
+
"step": 2775
|
| 839 |
+
},
|
| 840 |
+
{
|
| 841 |
+
"epoch": 4.946996466431095,
|
| 842 |
+
"grad_norm": 10.464198112487793,
|
| 843 |
+
"learning_rate": 2.8131134668237142e-05,
|
| 844 |
+
"loss": 0.3569,
|
| 845 |
+
"step": 2800
|
| 846 |
+
},
|
| 847 |
+
{
|
| 848 |
+
"epoch": 4.991166077738516,
|
| 849 |
+
"grad_norm": 60.609703063964844,
|
| 850 |
+
"learning_rate": 2.7885747938751472e-05,
|
| 851 |
+
"loss": 0.4576,
|
| 852 |
+
"step": 2825
|
| 853 |
+
},
|
| 854 |
+
{
|
| 855 |
+
"epoch": 5.0,
|
| 856 |
+
"eval_explained_variance": 0.6176950931549072,
|
| 857 |
+
"eval_loss": 2.477543830871582,
|
| 858 |
+
"eval_mae": 1.134692907333374,
|
| 859 |
+
"eval_mse": 2.467003345489502,
|
| 860 |
+
"eval_r2": 0.5996270179748535,
|
| 861 |
+
"eval_rmse": 1.5706697702407837,
|
| 862 |
+
"eval_runtime": 1.0254,
|
| 863 |
+
"eval_samples_per_second": 2206.885,
|
| 864 |
+
"eval_steps_per_second": 69.239,
|
| 865 |
+
"step": 2830
|
| 866 |
+
},
|
| 867 |
+
{
|
| 868 |
+
"epoch": 5.035335689045937,
|
| 869 |
+
"grad_norm": 15.79285717010498,
|
| 870 |
+
"learning_rate": 2.7640361209265802e-05,
|
| 871 |
+
"loss": 0.3458,
|
| 872 |
+
"step": 2850
|
| 873 |
+
},
|
| 874 |
+
{
|
| 875 |
+
"epoch": 5.079505300353357,
|
| 876 |
+
"grad_norm": 9.123896598815918,
|
| 877 |
+
"learning_rate": 2.7394974479780132e-05,
|
| 878 |
+
"loss": 0.2756,
|
| 879 |
+
"step": 2875
|
| 880 |
+
},
|
| 881 |
+
{
|
| 882 |
+
"epoch": 5.123674911660777,
|
| 883 |
+
"grad_norm": 10.223502159118652,
|
| 884 |
+
"learning_rate": 2.7149587750294462e-05,
|
| 885 |
+
"loss": 0.2661,
|
| 886 |
+
"step": 2900
|
| 887 |
+
},
|
| 888 |
+
{
|
| 889 |
+
"epoch": 5.167844522968198,
|
| 890 |
+
"grad_norm": 15.297872543334961,
|
| 891 |
+
"learning_rate": 2.6904201020808796e-05,
|
| 892 |
+
"loss": 0.2333,
|
| 893 |
+
"step": 2925
|
| 894 |
+
},
|
| 895 |
+
{
|
| 896 |
+
"epoch": 5.212014134275618,
|
| 897 |
+
"grad_norm": 6.982729911804199,
|
| 898 |
+
"learning_rate": 2.6658814291323126e-05,
|
| 899 |
+
"loss": 0.2596,
|
| 900 |
+
"step": 2950
|
| 901 |
+
},
|
| 902 |
+
{
|
| 903 |
+
"epoch": 5.256183745583039,
|
| 904 |
+
"grad_norm": 14.569358825683594,
|
| 905 |
+
"learning_rate": 2.6413427561837456e-05,
|
| 906 |
+
"loss": 0.2551,
|
| 907 |
+
"step": 2975
|
| 908 |
+
},
|
| 909 |
+
{
|
| 910 |
+
"epoch": 5.30035335689046,
|
| 911 |
+
"grad_norm": 10.034513473510742,
|
| 912 |
+
"learning_rate": 2.6168040832351786e-05,
|
| 913 |
+
"loss": 0.2628,
|
| 914 |
+
"step": 3000
|
| 915 |
+
},
|
| 916 |
+
{
|
| 917 |
+
"epoch": 5.34452296819788,
|
| 918 |
+
"grad_norm": 10.558284759521484,
|
| 919 |
+
"learning_rate": 2.5922654102866116e-05,
|
| 920 |
+
"loss": 0.2863,
|
| 921 |
+
"step": 3025
|
| 922 |
+
},
|
| 923 |
+
{
|
| 924 |
+
"epoch": 5.3886925795053005,
|
| 925 |
+
"grad_norm": 16.27973747253418,
|
| 926 |
+
"learning_rate": 2.5677267373380446e-05,
|
| 927 |
+
"loss": 0.2675,
|
| 928 |
+
"step": 3050
|
| 929 |
+
},
|
| 930 |
+
{
|
| 931 |
+
"epoch": 5.432862190812721,
|
| 932 |
+
"grad_norm": 14.27253532409668,
|
| 933 |
+
"learning_rate": 2.543188064389478e-05,
|
| 934 |
+
"loss": 0.283,
|
| 935 |
+
"step": 3075
|
| 936 |
+
},
|
| 937 |
+
{
|
| 938 |
+
"epoch": 5.477031802120141,
|
| 939 |
+
"grad_norm": 19.350168228149414,
|
| 940 |
+
"learning_rate": 2.518649391440911e-05,
|
| 941 |
+
"loss": 0.3148,
|
| 942 |
+
"step": 3100
|
| 943 |
+
},
|
| 944 |
+
{
|
| 945 |
+
"epoch": 5.521201413427562,
|
| 946 |
+
"grad_norm": 8.544435501098633,
|
| 947 |
+
"learning_rate": 2.494110718492344e-05,
|
| 948 |
+
"loss": 0.2565,
|
| 949 |
+
"step": 3125
|
| 950 |
+
},
|
| 951 |
+
{
|
| 952 |
+
"epoch": 5.565371024734983,
|
| 953 |
+
"grad_norm": 17.02871322631836,
|
| 954 |
+
"learning_rate": 2.469572045543777e-05,
|
| 955 |
+
"loss": 0.3099,
|
| 956 |
+
"step": 3150
|
| 957 |
+
},
|
| 958 |
+
{
|
| 959 |
+
"epoch": 5.609540636042403,
|
| 960 |
+
"grad_norm": 10.150726318359375,
|
| 961 |
+
"learning_rate": 2.44503337259521e-05,
|
| 962 |
+
"loss": 0.2532,
|
| 963 |
+
"step": 3175
|
| 964 |
+
},
|
| 965 |
+
{
|
| 966 |
+
"epoch": 5.6537102473498235,
|
| 967 |
+
"grad_norm": 14.919329643249512,
|
| 968 |
+
"learning_rate": 2.420494699646643e-05,
|
| 969 |
+
"loss": 0.2592,
|
| 970 |
+
"step": 3200
|
| 971 |
+
},
|
| 972 |
+
{
|
| 973 |
+
"epoch": 5.6978798586572434,
|
| 974 |
+
"grad_norm": 10.145200729370117,
|
| 975 |
+
"learning_rate": 2.3959560266980764e-05,
|
| 976 |
+
"loss": 0.2884,
|
| 977 |
+
"step": 3225
|
| 978 |
+
},
|
| 979 |
+
{
|
| 980 |
+
"epoch": 5.742049469964664,
|
| 981 |
+
"grad_norm": 26.694957733154297,
|
| 982 |
+
"learning_rate": 2.3714173537495094e-05,
|
| 983 |
+
"loss": 0.3092,
|
| 984 |
+
"step": 3250
|
| 985 |
+
},
|
| 986 |
+
{
|
| 987 |
+
"epoch": 5.786219081272085,
|
| 988 |
+
"grad_norm": 9.009997367858887,
|
| 989 |
+
"learning_rate": 2.3468786808009424e-05,
|
| 990 |
+
"loss": 0.2768,
|
| 991 |
+
"step": 3275
|
| 992 |
+
},
|
| 993 |
+
{
|
| 994 |
+
"epoch": 5.830388692579505,
|
| 995 |
+
"grad_norm": 14.945252418518066,
|
| 996 |
+
"learning_rate": 2.3223400078523754e-05,
|
| 997 |
+
"loss": 0.3033,
|
| 998 |
+
"step": 3300
|
| 999 |
+
},
|
| 1000 |
+
{
|
| 1001 |
+
"epoch": 5.874558303886926,
|
| 1002 |
+
"grad_norm": 22.651411056518555,
|
| 1003 |
+
"learning_rate": 2.2978013349038084e-05,
|
| 1004 |
+
"loss": 0.302,
|
| 1005 |
+
"step": 3325
|
| 1006 |
+
},
|
| 1007 |
+
{
|
| 1008 |
+
"epoch": 5.918727915194347,
|
| 1009 |
+
"grad_norm": 35.13420867919922,
|
| 1010 |
+
"learning_rate": 2.2732626619552418e-05,
|
| 1011 |
+
"loss": 0.2968,
|
| 1012 |
+
"step": 3350
|
| 1013 |
+
},
|
| 1014 |
+
{
|
| 1015 |
+
"epoch": 5.9628975265017665,
|
| 1016 |
+
"grad_norm": 8.667001724243164,
|
| 1017 |
+
"learning_rate": 2.2487239890066748e-05,
|
| 1018 |
+
"loss": 0.2591,
|
| 1019 |
+
"step": 3375
|
| 1020 |
+
},
|
| 1021 |
+
{
|
| 1022 |
+
"epoch": 6.0,
|
| 1023 |
+
"eval_explained_variance": 0.6250333786010742,
|
| 1024 |
+
"eval_loss": 2.4686992168426514,
|
| 1025 |
+
"eval_mae": 1.1278386116027832,
|
| 1026 |
+
"eval_mse": 2.4570138454437256,
|
| 1027 |
+
"eval_r2": 0.6012482643127441,
|
| 1028 |
+
"eval_rmse": 1.5674865245819092,
|
| 1029 |
+
"eval_runtime": 1.0575,
|
| 1030 |
+
"eval_samples_per_second": 2139.964,
|
| 1031 |
+
"eval_steps_per_second": 67.14,
|
| 1032 |
+
"step": 3396
|
| 1033 |
+
},
|
| 1034 |
+
{
|
| 1035 |
+
"epoch": 6.007067137809187,
|
| 1036 |
+
"grad_norm": 7.289055824279785,
|
| 1037 |
+
"learning_rate": 2.2241853160581078e-05,
|
| 1038 |
+
"loss": 0.2845,
|
| 1039 |
+
"step": 3400
|
| 1040 |
+
},
|
| 1041 |
+
{
|
| 1042 |
+
"epoch": 6.051236749116608,
|
| 1043 |
+
"grad_norm": 9.033802032470703,
|
| 1044 |
+
"learning_rate": 2.1996466431095408e-05,
|
| 1045 |
+
"loss": 0.2348,
|
| 1046 |
+
"step": 3425
|
| 1047 |
+
},
|
| 1048 |
+
{
|
| 1049 |
+
"epoch": 6.095406360424028,
|
| 1050 |
+
"grad_norm": 8.453237533569336,
|
| 1051 |
+
"learning_rate": 2.175107970160974e-05,
|
| 1052 |
+
"loss": 0.1626,
|
| 1053 |
+
"step": 3450
|
| 1054 |
+
},
|
| 1055 |
+
{
|
| 1056 |
+
"epoch": 6.139575971731449,
|
| 1057 |
+
"grad_norm": 12.62721061706543,
|
| 1058 |
+
"learning_rate": 2.150569297212407e-05,
|
| 1059 |
+
"loss": 0.1882,
|
| 1060 |
+
"step": 3475
|
| 1061 |
+
},
|
| 1062 |
+
{
|
| 1063 |
+
"epoch": 6.18374558303887,
|
| 1064 |
+
"grad_norm": 4.496434688568115,
|
| 1065 |
+
"learning_rate": 2.1260306242638402e-05,
|
| 1066 |
+
"loss": 0.2002,
|
| 1067 |
+
"step": 3500
|
| 1068 |
+
},
|
| 1069 |
+
{
|
| 1070 |
+
"epoch": 6.22791519434629,
|
| 1071 |
+
"grad_norm": 10.905092239379883,
|
| 1072 |
+
"learning_rate": 2.1014919513152732e-05,
|
| 1073 |
+
"loss": 0.1872,
|
| 1074 |
+
"step": 3525
|
| 1075 |
+
},
|
| 1076 |
+
{
|
| 1077 |
+
"epoch": 6.27208480565371,
|
| 1078 |
+
"grad_norm": 6.897697448730469,
|
| 1079 |
+
"learning_rate": 2.076953278366706e-05,
|
| 1080 |
+
"loss": 0.1842,
|
| 1081 |
+
"step": 3550
|
| 1082 |
+
},
|
| 1083 |
+
{
|
| 1084 |
+
"epoch": 6.316254416961131,
|
| 1085 |
+
"grad_norm": 18.04530906677246,
|
| 1086 |
+
"learning_rate": 2.052414605418139e-05,
|
| 1087 |
+
"loss": 0.1848,
|
| 1088 |
+
"step": 3575
|
| 1089 |
+
},
|
| 1090 |
+
{
|
| 1091 |
+
"epoch": 6.360424028268551,
|
| 1092 |
+
"grad_norm": 28.663089752197266,
|
| 1093 |
+
"learning_rate": 2.027875932469572e-05,
|
| 1094 |
+
"loss": 0.2338,
|
| 1095 |
+
"step": 3600
|
| 1096 |
+
},
|
| 1097 |
+
{
|
| 1098 |
+
"epoch": 6.404593639575972,
|
| 1099 |
+
"grad_norm": 7.388453960418701,
|
| 1100 |
+
"learning_rate": 2.0033372595210052e-05,
|
| 1101 |
+
"loss": 0.1731,
|
| 1102 |
+
"step": 3625
|
| 1103 |
+
},
|
| 1104 |
+
{
|
| 1105 |
+
"epoch": 6.448763250883392,
|
| 1106 |
+
"grad_norm": 8.80778694152832,
|
| 1107 |
+
"learning_rate": 1.9787985865724383e-05,
|
| 1108 |
+
"loss": 0.1948,
|
| 1109 |
+
"step": 3650
|
| 1110 |
+
},
|
| 1111 |
+
{
|
| 1112 |
+
"epoch": 6.492932862190813,
|
| 1113 |
+
"grad_norm": 15.45908260345459,
|
| 1114 |
+
"learning_rate": 1.9542599136238713e-05,
|
| 1115 |
+
"loss": 0.2057,
|
| 1116 |
+
"step": 3675
|
| 1117 |
+
},
|
| 1118 |
+
{
|
| 1119 |
+
"epoch": 6.5371024734982335,
|
| 1120 |
+
"grad_norm": 9.21857738494873,
|
| 1121 |
+
"learning_rate": 1.9297212406753043e-05,
|
| 1122 |
+
"loss": 0.2214,
|
| 1123 |
+
"step": 3700
|
| 1124 |
+
},
|
| 1125 |
+
{
|
| 1126 |
+
"epoch": 6.581272084805653,
|
| 1127 |
+
"grad_norm": 9.719182968139648,
|
| 1128 |
+
"learning_rate": 1.9051825677267373e-05,
|
| 1129 |
+
"loss": 0.1998,
|
| 1130 |
+
"step": 3725
|
| 1131 |
+
},
|
| 1132 |
+
{
|
| 1133 |
+
"epoch": 6.625441696113074,
|
| 1134 |
+
"grad_norm": 7.005542278289795,
|
| 1135 |
+
"learning_rate": 1.8806438947781703e-05,
|
| 1136 |
+
"loss": 0.1841,
|
| 1137 |
+
"step": 3750
|
| 1138 |
+
},
|
| 1139 |
+
{
|
| 1140 |
+
"epoch": 6.669611307420495,
|
| 1141 |
+
"grad_norm": 7.91580867767334,
|
| 1142 |
+
"learning_rate": 1.8561052218296036e-05,
|
| 1143 |
+
"loss": 0.1772,
|
| 1144 |
+
"step": 3775
|
| 1145 |
+
},
|
| 1146 |
+
{
|
| 1147 |
+
"epoch": 6.713780918727915,
|
| 1148 |
+
"grad_norm": 11.724370002746582,
|
| 1149 |
+
"learning_rate": 1.8315665488810367e-05,
|
| 1150 |
+
"loss": 0.1448,
|
| 1151 |
+
"step": 3800
|
| 1152 |
+
},
|
| 1153 |
+
{
|
| 1154 |
+
"epoch": 6.757950530035336,
|
| 1155 |
+
"grad_norm": 17.918067932128906,
|
| 1156 |
+
"learning_rate": 1.8070278759324697e-05,
|
| 1157 |
+
"loss": 0.2006,
|
| 1158 |
+
"step": 3825
|
| 1159 |
+
},
|
| 1160 |
+
{
|
| 1161 |
+
"epoch": 6.8021201413427566,
|
| 1162 |
+
"grad_norm": 11.593217849731445,
|
| 1163 |
+
"learning_rate": 1.7824892029839027e-05,
|
| 1164 |
+
"loss": 0.2018,
|
| 1165 |
+
"step": 3850
|
| 1166 |
+
},
|
| 1167 |
+
{
|
| 1168 |
+
"epoch": 6.8462897526501765,
|
| 1169 |
+
"grad_norm": 12.630243301391602,
|
| 1170 |
+
"learning_rate": 1.7579505300353357e-05,
|
| 1171 |
+
"loss": 0.1784,
|
| 1172 |
+
"step": 3875
|
| 1173 |
+
},
|
| 1174 |
+
{
|
| 1175 |
+
"epoch": 6.890459363957597,
|
| 1176 |
+
"grad_norm": 7.918076992034912,
|
| 1177 |
+
"learning_rate": 1.7334118570867687e-05,
|
| 1178 |
+
"loss": 0.1864,
|
| 1179 |
+
"step": 3900
|
| 1180 |
+
},
|
| 1181 |
+
{
|
| 1182 |
+
"epoch": 6.934628975265017,
|
| 1183 |
+
"grad_norm": 9.690032005310059,
|
| 1184 |
+
"learning_rate": 1.708873184138202e-05,
|
| 1185 |
+
"loss": 0.225,
|
| 1186 |
+
"step": 3925
|
| 1187 |
+
},
|
| 1188 |
+
{
|
| 1189 |
+
"epoch": 6.978798586572438,
|
| 1190 |
+
"grad_norm": 19.622501373291016,
|
| 1191 |
+
"learning_rate": 1.684334511189635e-05,
|
| 1192 |
+
"loss": 0.1898,
|
| 1193 |
+
"step": 3950
|
| 1194 |
+
},
|
| 1195 |
+
{
|
| 1196 |
+
"epoch": 7.0,
|
| 1197 |
+
"eval_explained_variance": 0.6274623870849609,
|
| 1198 |
+
"eval_loss": 2.4948439598083496,
|
| 1199 |
+
"eval_mae": 1.140474796295166,
|
| 1200 |
+
"eval_mse": 2.484145164489746,
|
| 1201 |
+
"eval_r2": 0.5968450903892517,
|
| 1202 |
+
"eval_rmse": 1.5761171579360962,
|
| 1203 |
+
"eval_runtime": 1.0544,
|
| 1204 |
+
"eval_samples_per_second": 2146.215,
|
| 1205 |
+
"eval_steps_per_second": 67.336,
|
| 1206 |
+
"step": 3962
|
| 1207 |
+
},
|
| 1208 |
+
{
|
| 1209 |
+
"epoch": 7.022968197879859,
|
| 1210 |
+
"grad_norm": 6.990258693695068,
|
| 1211 |
+
"learning_rate": 1.659795838241068e-05,
|
| 1212 |
+
"loss": 0.1784,
|
| 1213 |
+
"step": 3975
|
| 1214 |
+
},
|
| 1215 |
+
{
|
| 1216 |
+
"epoch": 7.067137809187279,
|
| 1217 |
+
"grad_norm": 5.022929668426514,
|
| 1218 |
+
"learning_rate": 1.635257165292501e-05,
|
| 1219 |
+
"loss": 0.1471,
|
| 1220 |
+
"step": 4000
|
| 1221 |
+
},
|
| 1222 |
+
{
|
| 1223 |
+
"epoch": 7.1113074204946995,
|
| 1224 |
+
"grad_norm": 6.776317119598389,
|
| 1225 |
+
"learning_rate": 1.610718492343934e-05,
|
| 1226 |
+
"loss": 0.1333,
|
| 1227 |
+
"step": 4025
|
| 1228 |
+
},
|
| 1229 |
+
{
|
| 1230 |
+
"epoch": 7.15547703180212,
|
| 1231 |
+
"grad_norm": 13.689594268798828,
|
| 1232 |
+
"learning_rate": 1.586179819395367e-05,
|
| 1233 |
+
"loss": 0.1339,
|
| 1234 |
+
"step": 4050
|
| 1235 |
+
},
|
| 1236 |
+
{
|
| 1237 |
+
"epoch": 7.19964664310954,
|
| 1238 |
+
"grad_norm": 6.960967063903809,
|
| 1239 |
+
"learning_rate": 1.5616411464468005e-05,
|
| 1240 |
+
"loss": 0.1278,
|
| 1241 |
+
"step": 4075
|
| 1242 |
+
},
|
| 1243 |
+
{
|
| 1244 |
+
"epoch": 7.243816254416961,
|
| 1245 |
+
"grad_norm": 9.860389709472656,
|
| 1246 |
+
"learning_rate": 1.5371024734982335e-05,
|
| 1247 |
+
"loss": 0.1419,
|
| 1248 |
+
"step": 4100
|
| 1249 |
+
},
|
| 1250 |
+
{
|
| 1251 |
+
"epoch": 7.287985865724382,
|
| 1252 |
+
"grad_norm": 12.38211727142334,
|
| 1253 |
+
"learning_rate": 1.5125638005496665e-05,
|
| 1254 |
+
"loss": 0.154,
|
| 1255 |
+
"step": 4125
|
| 1256 |
+
},
|
| 1257 |
+
{
|
| 1258 |
+
"epoch": 7.332155477031802,
|
| 1259 |
+
"grad_norm": 5.641261577606201,
|
| 1260 |
+
"learning_rate": 1.4880251276010995e-05,
|
| 1261 |
+
"loss": 0.1556,
|
| 1262 |
+
"step": 4150
|
| 1263 |
+
},
|
| 1264 |
+
{
|
| 1265 |
+
"epoch": 7.376325088339223,
|
| 1266 |
+
"grad_norm": 12.16318416595459,
|
| 1267 |
+
"learning_rate": 1.4634864546525327e-05,
|
| 1268 |
+
"loss": 0.1574,
|
| 1269 |
+
"step": 4175
|
| 1270 |
+
},
|
| 1271 |
+
{
|
| 1272 |
+
"epoch": 7.420494699646643,
|
| 1273 |
+
"grad_norm": 6.69941520690918,
|
| 1274 |
+
"learning_rate": 1.4389477817039653e-05,
|
| 1275 |
+
"loss": 0.1461,
|
| 1276 |
+
"step": 4200
|
| 1277 |
+
},
|
| 1278 |
+
{
|
| 1279 |
+
"epoch": 7.464664310954063,
|
| 1280 |
+
"grad_norm": 6.084545612335205,
|
| 1281 |
+
"learning_rate": 1.4144091087553985e-05,
|
| 1282 |
+
"loss": 0.1448,
|
| 1283 |
+
"step": 4225
|
| 1284 |
+
},
|
| 1285 |
+
{
|
| 1286 |
+
"epoch": 7.508833922261484,
|
| 1287 |
+
"grad_norm": 10.190794944763184,
|
| 1288 |
+
"learning_rate": 1.3898704358068315e-05,
|
| 1289 |
+
"loss": 0.1739,
|
| 1290 |
+
"step": 4250
|
| 1291 |
+
},
|
| 1292 |
+
{
|
| 1293 |
+
"epoch": 7.553003533568905,
|
| 1294 |
+
"grad_norm": 5.71160364151001,
|
| 1295 |
+
"learning_rate": 1.3653317628582645e-05,
|
| 1296 |
+
"loss": 0.1162,
|
| 1297 |
+
"step": 4275
|
| 1298 |
+
},
|
| 1299 |
+
{
|
| 1300 |
+
"epoch": 7.597173144876325,
|
| 1301 |
+
"grad_norm": 7.204683303833008,
|
| 1302 |
+
"learning_rate": 1.3407930899096977e-05,
|
| 1303 |
+
"loss": 0.1351,
|
| 1304 |
+
"step": 4300
|
| 1305 |
+
},
|
| 1306 |
+
{
|
| 1307 |
+
"epoch": 7.641342756183746,
|
| 1308 |
+
"grad_norm": 11.425016403198242,
|
| 1309 |
+
"learning_rate": 1.3162544169611307e-05,
|
| 1310 |
+
"loss": 0.1327,
|
| 1311 |
+
"step": 4325
|
| 1312 |
+
},
|
| 1313 |
+
{
|
| 1314 |
+
"epoch": 7.685512367491166,
|
| 1315 |
+
"grad_norm": 5.907583713531494,
|
| 1316 |
+
"learning_rate": 1.2917157440125637e-05,
|
| 1317 |
+
"loss": 0.1308,
|
| 1318 |
+
"step": 4350
|
| 1319 |
+
},
|
| 1320 |
+
{
|
| 1321 |
+
"epoch": 7.729681978798586,
|
| 1322 |
+
"grad_norm": 7.888271808624268,
|
| 1323 |
+
"learning_rate": 1.267177071063997e-05,
|
| 1324 |
+
"loss": 0.1427,
|
| 1325 |
+
"step": 4375
|
| 1326 |
+
},
|
| 1327 |
+
{
|
| 1328 |
+
"epoch": 7.773851590106007,
|
| 1329 |
+
"grad_norm": 5.342504024505615,
|
| 1330 |
+
"learning_rate": 1.24263839811543e-05,
|
| 1331 |
+
"loss": 0.1437,
|
| 1332 |
+
"step": 4400
|
| 1333 |
+
},
|
| 1334 |
+
{
|
| 1335 |
+
"epoch": 7.818021201413428,
|
| 1336 |
+
"grad_norm": 4.982570171356201,
|
| 1337 |
+
"learning_rate": 1.218099725166863e-05,
|
| 1338 |
+
"loss": 0.1237,
|
| 1339 |
+
"step": 4425
|
| 1340 |
+
},
|
| 1341 |
+
{
|
| 1342 |
+
"epoch": 7.862190812720848,
|
| 1343 |
+
"grad_norm": 3.8644771575927734,
|
| 1344 |
+
"learning_rate": 1.1935610522182961e-05,
|
| 1345 |
+
"loss": 0.1225,
|
| 1346 |
+
"step": 4450
|
| 1347 |
+
},
|
| 1348 |
+
{
|
| 1349 |
+
"epoch": 7.906360424028269,
|
| 1350 |
+
"grad_norm": 3.926877498626709,
|
| 1351 |
+
"learning_rate": 1.1690223792697291e-05,
|
| 1352 |
+
"loss": 0.1262,
|
| 1353 |
+
"step": 4475
|
| 1354 |
+
},
|
| 1355 |
+
{
|
| 1356 |
+
"epoch": 7.950530035335689,
|
| 1357 |
+
"grad_norm": 6.653122425079346,
|
| 1358 |
+
"learning_rate": 1.1444837063211621e-05,
|
| 1359 |
+
"loss": 0.1333,
|
| 1360 |
+
"step": 4500
|
| 1361 |
+
},
|
| 1362 |
+
{
|
| 1363 |
+
"epoch": 7.9946996466431095,
|
| 1364 |
+
"grad_norm": 7.844146251678467,
|
| 1365 |
+
"learning_rate": 1.1199450333725953e-05,
|
| 1366 |
+
"loss": 0.1308,
|
| 1367 |
+
"step": 4525
|
| 1368 |
+
},
|
| 1369 |
+
{
|
| 1370 |
+
"epoch": 8.0,
|
| 1371 |
+
"eval_explained_variance": 0.6261004209518433,
|
| 1372 |
+
"eval_loss": 2.495262622833252,
|
| 1373 |
+
"eval_mae": 1.1328405141830444,
|
| 1374 |
+
"eval_mse": 2.48473858833313,
|
| 1375 |
+
"eval_r2": 0.5967487692832947,
|
| 1376 |
+
"eval_rmse": 1.5763053894042969,
|
| 1377 |
+
"eval_runtime": 1.0666,
|
| 1378 |
+
"eval_samples_per_second": 2121.718,
|
| 1379 |
+
"eval_steps_per_second": 66.567,
|
| 1380 |
+
"step": 4528
|
| 1381 |
+
},
|
| 1382 |
+
{
|
| 1383 |
+
"epoch": 8.03886925795053,
|
| 1384 |
+
"grad_norm": 24.181594848632812,
|
| 1385 |
+
"learning_rate": 1.0954063604240283e-05,
|
| 1386 |
+
"loss": 0.1091,
|
| 1387 |
+
"step": 4550
|
| 1388 |
+
},
|
| 1389 |
+
{
|
| 1390 |
+
"epoch": 8.083038869257951,
|
| 1391 |
+
"grad_norm": 3.6817755699157715,
|
| 1392 |
+
"learning_rate": 1.0708676874754613e-05,
|
| 1393 |
+
"loss": 0.1275,
|
| 1394 |
+
"step": 4575
|
| 1395 |
+
},
|
| 1396 |
+
{
|
| 1397 |
+
"epoch": 8.12720848056537,
|
| 1398 |
+
"grad_norm": 3.9178013801574707,
|
| 1399 |
+
"learning_rate": 1.0463290145268945e-05,
|
| 1400 |
+
"loss": 0.0956,
|
| 1401 |
+
"step": 4600
|
| 1402 |
+
},
|
| 1403 |
+
{
|
| 1404 |
+
"epoch": 8.171378091872791,
|
| 1405 |
+
"grad_norm": 4.021216869354248,
|
| 1406 |
+
"learning_rate": 1.0217903415783275e-05,
|
| 1407 |
+
"loss": 0.103,
|
| 1408 |
+
"step": 4625
|
| 1409 |
+
},
|
| 1410 |
+
{
|
| 1411 |
+
"epoch": 8.215547703180212,
|
| 1412 |
+
"grad_norm": 11.239612579345703,
|
| 1413 |
+
"learning_rate": 9.972516686297605e-06,
|
| 1414 |
+
"loss": 0.0989,
|
| 1415 |
+
"step": 4650
|
| 1416 |
+
},
|
| 1417 |
+
{
|
| 1418 |
+
"epoch": 8.259717314487633,
|
| 1419 |
+
"grad_norm": 5.062624931335449,
|
| 1420 |
+
"learning_rate": 9.727129956811937e-06,
|
| 1421 |
+
"loss": 0.0998,
|
| 1422 |
+
"step": 4675
|
| 1423 |
+
},
|
| 1424 |
+
{
|
| 1425 |
+
"epoch": 8.303886925795053,
|
| 1426 |
+
"grad_norm": 4.972472190856934,
|
| 1427 |
+
"learning_rate": 9.481743227326266e-06,
|
| 1428 |
+
"loss": 0.0991,
|
| 1429 |
+
"step": 4700
|
| 1430 |
+
},
|
| 1431 |
+
{
|
| 1432 |
+
"epoch": 8.348056537102474,
|
| 1433 |
+
"grad_norm": 4.167801380157471,
|
| 1434 |
+
"learning_rate": 9.236356497840597e-06,
|
| 1435 |
+
"loss": 0.1016,
|
| 1436 |
+
"step": 4725
|
| 1437 |
+
},
|
| 1438 |
+
{
|
| 1439 |
+
"epoch": 8.392226148409893,
|
| 1440 |
+
"grad_norm": 4.731273651123047,
|
| 1441 |
+
"learning_rate": 8.990969768354928e-06,
|
| 1442 |
+
"loss": 0.1025,
|
| 1443 |
+
"step": 4750
|
| 1444 |
+
},
|
| 1445 |
+
{
|
| 1446 |
+
"epoch": 8.436395759717314,
|
| 1447 |
+
"grad_norm": 5.190533638000488,
|
| 1448 |
+
"learning_rate": 8.745583038869258e-06,
|
| 1449 |
+
"loss": 0.1012,
|
| 1450 |
+
"step": 4775
|
| 1451 |
+
},
|
| 1452 |
+
{
|
| 1453 |
+
"epoch": 8.480565371024735,
|
| 1454 |
+
"grad_norm": 7.925145149230957,
|
| 1455 |
+
"learning_rate": 8.50019630938359e-06,
|
| 1456 |
+
"loss": 0.1126,
|
| 1457 |
+
"step": 4800
|
| 1458 |
+
},
|
| 1459 |
+
{
|
| 1460 |
+
"epoch": 8.524734982332156,
|
| 1461 |
+
"grad_norm": 6.021143913269043,
|
| 1462 |
+
"learning_rate": 8.25480957989792e-06,
|
| 1463 |
+
"loss": 0.0804,
|
| 1464 |
+
"step": 4825
|
| 1465 |
+
},
|
| 1466 |
+
{
|
| 1467 |
+
"epoch": 8.568904593639576,
|
| 1468 |
+
"grad_norm": 4.904351711273193,
|
| 1469 |
+
"learning_rate": 8.00942285041225e-06,
|
| 1470 |
+
"loss": 0.0984,
|
| 1471 |
+
"step": 4850
|
| 1472 |
+
},
|
| 1473 |
+
{
|
| 1474 |
+
"epoch": 8.613074204946997,
|
| 1475 |
+
"grad_norm": 5.331979274749756,
|
| 1476 |
+
"learning_rate": 7.764036120926581e-06,
|
| 1477 |
+
"loss": 0.1087,
|
| 1478 |
+
"step": 4875
|
| 1479 |
+
},
|
| 1480 |
+
{
|
| 1481 |
+
"epoch": 8.657243816254416,
|
| 1482 |
+
"grad_norm": 7.438701629638672,
|
| 1483 |
+
"learning_rate": 7.518649391440912e-06,
|
| 1484 |
+
"loss": 0.11,
|
| 1485 |
+
"step": 4900
|
| 1486 |
+
},
|
| 1487 |
+
{
|
| 1488 |
+
"epoch": 8.701413427561837,
|
| 1489 |
+
"grad_norm": 5.5875420570373535,
|
| 1490 |
+
"learning_rate": 7.2732626619552425e-06,
|
| 1491 |
+
"loss": 0.1211,
|
| 1492 |
+
"step": 4925
|
| 1493 |
+
},
|
| 1494 |
+
{
|
| 1495 |
+
"epoch": 8.745583038869258,
|
| 1496 |
+
"grad_norm": 5.041488170623779,
|
| 1497 |
+
"learning_rate": 7.027875932469573e-06,
|
| 1498 |
+
"loss": 0.0783,
|
| 1499 |
+
"step": 4950
|
| 1500 |
+
},
|
| 1501 |
+
{
|
| 1502 |
+
"epoch": 8.789752650176679,
|
| 1503 |
+
"grad_norm": 3.2430458068847656,
|
| 1504 |
+
"learning_rate": 6.782489202983904e-06,
|
| 1505 |
+
"loss": 0.0982,
|
| 1506 |
+
"step": 4975
|
| 1507 |
+
},
|
| 1508 |
+
{
|
| 1509 |
+
"epoch": 8.8339222614841,
|
| 1510 |
+
"grad_norm": 4.588315010070801,
|
| 1511 |
+
"learning_rate": 6.5371024734982345e-06,
|
| 1512 |
+
"loss": 0.1137,
|
| 1513 |
+
"step": 5000
|
| 1514 |
+
},
|
| 1515 |
+
{
|
| 1516 |
+
"epoch": 8.878091872791519,
|
| 1517 |
+
"grad_norm": 5.235021114349365,
|
| 1518 |
+
"learning_rate": 6.291715744012564e-06,
|
| 1519 |
+
"loss": 0.097,
|
| 1520 |
+
"step": 5025
|
| 1521 |
+
},
|
| 1522 |
+
{
|
| 1523 |
+
"epoch": 8.92226148409894,
|
| 1524 |
+
"grad_norm": 4.683666229248047,
|
| 1525 |
+
"learning_rate": 6.046329014526895e-06,
|
| 1526 |
+
"loss": 0.0836,
|
| 1527 |
+
"step": 5050
|
| 1528 |
+
},
|
| 1529 |
+
{
|
| 1530 |
+
"epoch": 8.96643109540636,
|
| 1531 |
+
"grad_norm": 3.5004100799560547,
|
| 1532 |
+
"learning_rate": 5.800942285041226e-06,
|
| 1533 |
+
"loss": 0.0884,
|
| 1534 |
+
"step": 5075
|
| 1535 |
+
},
|
| 1536 |
+
{
|
| 1537 |
+
"epoch": 9.0,
|
| 1538 |
+
"eval_explained_variance": 0.6279962062835693,
|
| 1539 |
+
"eval_loss": 2.4493231773376465,
|
| 1540 |
+
"eval_mae": 1.1165964603424072,
|
| 1541 |
+
"eval_mse": 2.43858003616333,
|
| 1542 |
+
"eval_r2": 0.6042398810386658,
|
| 1543 |
+
"eval_rmse": 1.5615953207015991,
|
| 1544 |
+
"eval_runtime": 1.0409,
|
| 1545 |
+
"eval_samples_per_second": 2174.145,
|
| 1546 |
+
"eval_steps_per_second": 68.212,
|
| 1547 |
+
"step": 5094
|
| 1548 |
+
},
|
| 1549 |
+
{
|
| 1550 |
+
"epoch": 9.010600706713781,
|
| 1551 |
+
"grad_norm": 6.5793352127075195,
|
| 1552 |
+
"learning_rate": 5.555555555555556e-06,
|
| 1553 |
+
"loss": 0.0868,
|
| 1554 |
+
"step": 5100
|
| 1555 |
+
},
|
| 1556 |
+
{
|
| 1557 |
+
"epoch": 9.054770318021202,
|
| 1558 |
+
"grad_norm": 2.6503169536590576,
|
| 1559 |
+
"learning_rate": 5.310168826069886e-06,
|
| 1560 |
+
"loss": 0.0657,
|
| 1561 |
+
"step": 5125
|
| 1562 |
+
},
|
| 1563 |
+
{
|
| 1564 |
+
"epoch": 9.098939929328623,
|
| 1565 |
+
"grad_norm": 2.9820001125335693,
|
| 1566 |
+
"learning_rate": 5.064782096584217e-06,
|
| 1567 |
+
"loss": 0.0699,
|
| 1568 |
+
"step": 5150
|
| 1569 |
+
},
|
| 1570 |
+
{
|
| 1571 |
+
"epoch": 9.143109540636042,
|
| 1572 |
+
"grad_norm": 3.706174373626709,
|
| 1573 |
+
"learning_rate": 4.819395367098548e-06,
|
| 1574 |
+
"loss": 0.0811,
|
| 1575 |
+
"step": 5175
|
| 1576 |
+
},
|
| 1577 |
+
{
|
| 1578 |
+
"epoch": 9.187279151943462,
|
| 1579 |
+
"grad_norm": 2.652496099472046,
|
| 1580 |
+
"learning_rate": 4.574008637612878e-06,
|
| 1581 |
+
"loss": 0.0658,
|
| 1582 |
+
"step": 5200
|
| 1583 |
+
},
|
| 1584 |
+
{
|
| 1585 |
+
"epoch": 9.231448763250883,
|
| 1586 |
+
"grad_norm": 4.267024040222168,
|
| 1587 |
+
"learning_rate": 4.328621908127209e-06,
|
| 1588 |
+
"loss": 0.0706,
|
| 1589 |
+
"step": 5225
|
| 1590 |
+
},
|
| 1591 |
+
{
|
| 1592 |
+
"epoch": 9.275618374558304,
|
| 1593 |
+
"grad_norm": 2.936169147491455,
|
| 1594 |
+
"learning_rate": 4.08323517864154e-06,
|
| 1595 |
+
"loss": 0.0661,
|
| 1596 |
+
"step": 5250
|
| 1597 |
+
},
|
| 1598 |
+
{
|
| 1599 |
+
"epoch": 9.319787985865725,
|
| 1600 |
+
"grad_norm": 3.2708146572113037,
|
| 1601 |
+
"learning_rate": 3.83784844915587e-06,
|
| 1602 |
+
"loss": 0.0829,
|
| 1603 |
+
"step": 5275
|
| 1604 |
+
},
|
| 1605 |
+
{
|
| 1606 |
+
"epoch": 9.363957597173146,
|
| 1607 |
+
"grad_norm": 5.624586582183838,
|
| 1608 |
+
"learning_rate": 3.5924617196702e-06,
|
| 1609 |
+
"loss": 0.0847,
|
| 1610 |
+
"step": 5300
|
| 1611 |
+
},
|
| 1612 |
+
{
|
| 1613 |
+
"epoch": 9.408127208480565,
|
| 1614 |
+
"grad_norm": 6.347949981689453,
|
| 1615 |
+
"learning_rate": 3.347074990184531e-06,
|
| 1616 |
+
"loss": 0.081,
|
| 1617 |
+
"step": 5325
|
| 1618 |
+
},
|
| 1619 |
+
{
|
| 1620 |
+
"epoch": 9.452296819787986,
|
| 1621 |
+
"grad_norm": 3.8003103733062744,
|
| 1622 |
+
"learning_rate": 3.1016882606988615e-06,
|
| 1623 |
+
"loss": 0.0657,
|
| 1624 |
+
"step": 5350
|
| 1625 |
+
},
|
| 1626 |
+
{
|
| 1627 |
+
"epoch": 9.496466431095406,
|
| 1628 |
+
"grad_norm": 3.1843507289886475,
|
| 1629 |
+
"learning_rate": 2.8563015312131925e-06,
|
| 1630 |
+
"loss": 0.0613,
|
| 1631 |
+
"step": 5375
|
| 1632 |
+
},
|
| 1633 |
+
{
|
| 1634 |
+
"epoch": 9.540636042402827,
|
| 1635 |
+
"grad_norm": 4.932063579559326,
|
| 1636 |
+
"learning_rate": 2.6109148017275226e-06,
|
| 1637 |
+
"loss": 0.0743,
|
| 1638 |
+
"step": 5400
|
| 1639 |
+
},
|
| 1640 |
+
{
|
| 1641 |
+
"epoch": 9.584805653710248,
|
| 1642 |
+
"grad_norm": 7.54895544052124,
|
| 1643 |
+
"learning_rate": 2.365528072241853e-06,
|
| 1644 |
+
"loss": 0.0683,
|
| 1645 |
+
"step": 5425
|
| 1646 |
+
},
|
| 1647 |
+
{
|
| 1648 |
+
"epoch": 9.628975265017669,
|
| 1649 |
+
"grad_norm": 5.509974956512451,
|
| 1650 |
+
"learning_rate": 2.120141342756184e-06,
|
| 1651 |
+
"loss": 0.0736,
|
| 1652 |
+
"step": 5450
|
| 1653 |
+
},
|
| 1654 |
+
{
|
| 1655 |
+
"epoch": 9.673144876325088,
|
| 1656 |
+
"grad_norm": 4.273847579956055,
|
| 1657 |
+
"learning_rate": 1.8747546132705144e-06,
|
| 1658 |
+
"loss": 0.0618,
|
| 1659 |
+
"step": 5475
|
| 1660 |
+
},
|
| 1661 |
+
{
|
| 1662 |
+
"epoch": 9.717314487632509,
|
| 1663 |
+
"grad_norm": 7.0803303718566895,
|
| 1664 |
+
"learning_rate": 1.629367883784845e-06,
|
| 1665 |
+
"loss": 0.0691,
|
| 1666 |
+
"step": 5500
|
| 1667 |
+
},
|
| 1668 |
+
{
|
| 1669 |
+
"epoch": 9.76148409893993,
|
| 1670 |
+
"grad_norm": 6.049960613250732,
|
| 1671 |
+
"learning_rate": 1.3839811542991755e-06,
|
| 1672 |
+
"loss": 0.0995,
|
| 1673 |
+
"step": 5525
|
| 1674 |
+
},
|
| 1675 |
+
{
|
| 1676 |
+
"epoch": 9.80565371024735,
|
| 1677 |
+
"grad_norm": 3.8500442504882812,
|
| 1678 |
+
"learning_rate": 1.1385944248135062e-06,
|
| 1679 |
+
"loss": 0.0732,
|
| 1680 |
+
"step": 5550
|
| 1681 |
+
},
|
| 1682 |
+
{
|
| 1683 |
+
"epoch": 9.849823321554771,
|
| 1684 |
+
"grad_norm": 4.026426792144775,
|
| 1685 |
+
"learning_rate": 8.932076953278367e-07,
|
| 1686 |
+
"loss": 0.0706,
|
| 1687 |
+
"step": 5575
|
| 1688 |
+
},
|
| 1689 |
+
{
|
| 1690 |
+
"epoch": 9.89399293286219,
|
| 1691 |
+
"grad_norm": 2.691366195678711,
|
| 1692 |
+
"learning_rate": 6.478209658421673e-07,
|
| 1693 |
+
"loss": 0.0807,
|
| 1694 |
+
"step": 5600
|
| 1695 |
+
},
|
| 1696 |
+
{
|
| 1697 |
+
"epoch": 9.93816254416961,
|
| 1698 |
+
"grad_norm": 4.666661262512207,
|
| 1699 |
+
"learning_rate": 4.024342363564979e-07,
|
| 1700 |
+
"loss": 0.0701,
|
| 1701 |
+
"step": 5625
|
| 1702 |
+
},
|
| 1703 |
+
{
|
| 1704 |
+
"epoch": 9.982332155477032,
|
| 1705 |
+
"grad_norm": 3.5465140342712402,
|
| 1706 |
+
"learning_rate": 1.5704750687082843e-07,
|
| 1707 |
+
"loss": 0.0763,
|
| 1708 |
+
"step": 5650
|
| 1709 |
+
},
|
| 1710 |
+
{
|
| 1711 |
+
"epoch": 10.0,
|
| 1712 |
+
"eval_explained_variance": 0.631534218788147,
|
| 1713 |
+
"eval_loss": 2.4403305053710938,
|
| 1714 |
+
"eval_mae": 1.1220409870147705,
|
| 1715 |
+
"eval_mse": 2.4293971061706543,
|
| 1716 |
+
"eval_r2": 0.6057301759719849,
|
| 1717 |
+
"eval_rmse": 1.5586522817611694,
|
| 1718 |
+
"eval_runtime": 0.9929,
|
| 1719 |
+
"eval_samples_per_second": 2279.266,
|
| 1720 |
+
"eval_steps_per_second": 71.51,
|
| 1721 |
+
"step": 5660
|
| 1722 |
+
}
|
| 1723 |
+
],
|
| 1724 |
+
"logging_steps": 25,
|
| 1725 |
+
"max_steps": 5660,
|
| 1726 |
+
"num_input_tokens_seen": 0,
|
| 1727 |
+
"num_train_epochs": 10,
|
| 1728 |
+
"save_steps": 500,
|
| 1729 |
+
"stateful_callbacks": {
|
| 1730 |
+
"EarlyStoppingCallback": {
|
| 1731 |
+
"args": {
|
| 1732 |
+
"early_stopping_patience": 5,
|
| 1733 |
+
"early_stopping_threshold": 0.01
|
| 1734 |
+
},
|
| 1735 |
+
"attributes": {
|
| 1736 |
+
"early_stopping_patience_counter": 1
|
| 1737 |
+
}
|
| 1738 |
+
},
|
| 1739 |
+
"TrainerControl": {
|
| 1740 |
+
"args": {
|
| 1741 |
+
"should_epoch_stop": false,
|
| 1742 |
+
"should_evaluate": false,
|
| 1743 |
+
"should_log": false,
|
| 1744 |
+
"should_save": true,
|
| 1745 |
+
"should_training_stop": true
|
| 1746 |
+
},
|
| 1747 |
+
"attributes": {}
|
| 1748 |
+
}
|
| 1749 |
+
},
|
| 1750 |
+
"total_flos": 5956780739788800.0,
|
| 1751 |
+
"train_batch_size": 8,
|
| 1752 |
+
"trial_name": null,
|
| 1753 |
+
"trial_params": null
|
| 1754 |
+
}
|
checkpoint-5660/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a3c322a4ba2e6b8a429d05aaa09f515dabcedf761b82c967962fa5d87b18b463
|
| 3 |
+
size 5240
|
config.json
ADDED
|
@@ -0,0 +1,34 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "google-bert/bert-base-uncased",
|
| 3 |
+
"_num_labels": 1,
|
| 4 |
+
"architectures": [
|
| 5 |
+
"BertForSequenceClassification"
|
| 6 |
+
],
|
| 7 |
+
"attention_probs_dropout_prob": 0.1,
|
| 8 |
+
"classifier_dropout": null,
|
| 9 |
+
"gradient_checkpointing": false,
|
| 10 |
+
"hidden_act": "gelu",
|
| 11 |
+
"hidden_dropout_prob": 0.1,
|
| 12 |
+
"hidden_size": 768,
|
| 13 |
+
"id2label": {
|
| 14 |
+
"0": "target"
|
| 15 |
+
},
|
| 16 |
+
"initializer_range": 0.02,
|
| 17 |
+
"intermediate_size": 3072,
|
| 18 |
+
"label2id": {
|
| 19 |
+
"target": 0
|
| 20 |
+
},
|
| 21 |
+
"layer_norm_eps": 1e-12,
|
| 22 |
+
"max_position_embeddings": 512,
|
| 23 |
+
"model_type": "bert",
|
| 24 |
+
"num_attention_heads": 12,
|
| 25 |
+
"num_hidden_layers": 12,
|
| 26 |
+
"pad_token_id": 0,
|
| 27 |
+
"position_embedding_type": "absolute",
|
| 28 |
+
"problem_type": "regression",
|
| 29 |
+
"torch_dtype": "float32",
|
| 30 |
+
"transformers_version": "4.45.0",
|
| 31 |
+
"type_vocab_size": 2,
|
| 32 |
+
"use_cache": true,
|
| 33 |
+
"vocab_size": 30522
|
| 34 |
+
}
|
model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f221ffc1102181fde09e11c8c654b2bd8537a14f574df593c45fc36c48e44544
|
| 3 |
+
size 437955572
|
runs/Oct23_07-17-42_bora-Lambda-Vector/events.out.tfevents.1729693064.bora-Lambda-Vector.288568.0
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 sha256:9f7f41066c57a0c4b43836c8dfb85cd2a0fd74c3d3d94401ed7f2b0037fd9241
|
| 3 |
+
size 58253
|
runs/Oct23_07-17-42_bora-Lambda-Vector/events.out.tfevents.1729693410.bora-Lambda-Vector.288568.1
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:d158b182d7555360b5d3f3c692472dd1d343c51a9a8852f19c19188498ccc405
|
| 3 |
+
size 609
|
special_tokens_map.json
ADDED
|
@@ -0,0 +1,7 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"cls_token": "[CLS]",
|
| 3 |
+
"mask_token": "[MASK]",
|
| 4 |
+
"pad_token": "[PAD]",
|
| 5 |
+
"sep_token": "[SEP]",
|
| 6 |
+
"unk_token": "[UNK]"
|
| 7 |
+
}
|
tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
tokenizer_config.json
ADDED
|
@@ -0,0 +1,55 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 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 |
+
"100": {
|
| 12 |
+
"content": "[UNK]",
|
| 13 |
+
"lstrip": false,
|
| 14 |
+
"normalized": false,
|
| 15 |
+
"rstrip": false,
|
| 16 |
+
"single_word": false,
|
| 17 |
+
"special": true
|
| 18 |
+
},
|
| 19 |
+
"101": {
|
| 20 |
+
"content": "[CLS]",
|
| 21 |
+
"lstrip": false,
|
| 22 |
+
"normalized": false,
|
| 23 |
+
"rstrip": false,
|
| 24 |
+
"single_word": false,
|
| 25 |
+
"special": true
|
| 26 |
+
},
|
| 27 |
+
"102": {
|
| 28 |
+
"content": "[SEP]",
|
| 29 |
+
"lstrip": false,
|
| 30 |
+
"normalized": false,
|
| 31 |
+
"rstrip": false,
|
| 32 |
+
"single_word": false,
|
| 33 |
+
"special": true
|
| 34 |
+
},
|
| 35 |
+
"103": {
|
| 36 |
+
"content": "[MASK]",
|
| 37 |
+
"lstrip": false,
|
| 38 |
+
"normalized": false,
|
| 39 |
+
"rstrip": false,
|
| 40 |
+
"single_word": false,
|
| 41 |
+
"special": true
|
| 42 |
+
}
|
| 43 |
+
},
|
| 44 |
+
"clean_up_tokenization_spaces": true,
|
| 45 |
+
"cls_token": "[CLS]",
|
| 46 |
+
"do_lower_case": true,
|
| 47 |
+
"mask_token": "[MASK]",
|
| 48 |
+
"model_max_length": 512,
|
| 49 |
+
"pad_token": "[PAD]",
|
| 50 |
+
"sep_token": "[SEP]",
|
| 51 |
+
"strip_accents": null,
|
| 52 |
+
"tokenize_chinese_chars": true,
|
| 53 |
+
"tokenizer_class": "BertTokenizer",
|
| 54 |
+
"unk_token": "[UNK]"
|
| 55 |
+
}
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a3c322a4ba2e6b8a429d05aaa09f515dabcedf761b82c967962fa5d87b18b463
|
| 3 |
+
size 5240
|
training_params.json
ADDED
|
@@ -0,0 +1,30 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"data_path": "QaagiFilter5/autotrain-data",
|
| 3 |
+
"model": "google-bert/bert-base-uncased",
|
| 4 |
+
"lr": 5e-05,
|
| 5 |
+
"epochs": 10,
|
| 6 |
+
"max_seq_length": 128,
|
| 7 |
+
"batch_size": 8,
|
| 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": "QaagiFilter5",
|
| 21 |
+
"auto_find_batch_size": false,
|
| 22 |
+
"mixed_precision": "fp16",
|
| 23 |
+
"save_total_limit": 1,
|
| 24 |
+
"push_to_hub": true,
|
| 25 |
+
"eval_strategy": "epoch",
|
| 26 |
+
"username": "vuk123",
|
| 27 |
+
"log": "tensorboard",
|
| 28 |
+
"early_stopping_patience": 5,
|
| 29 |
+
"early_stopping_threshold": 0.01
|
| 30 |
+
}
|
vocab.txt
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|