End of training
Browse files- README.md +14 -2
- all_results.json +16 -0
- eval_results.json +10 -0
- tokenizer.json +1 -6
- train_results.json +9 -0
- trainer_state.json +469 -0
README.md
CHANGED
|
@@ -3,11 +3,23 @@ license: other
|
|
| 3 |
base_model: Qwen/Qwen1.5-4B
|
| 4 |
tags:
|
| 5 |
- generated_from_trainer
|
|
|
|
|
|
|
| 6 |
metrics:
|
| 7 |
- accuracy
|
| 8 |
model-index:
|
| 9 |
- name: lmind_nq_train6000_eval6489_v1_reciteonly_qa_v3_Qwen_Qwen1.5-4B_5e-5_lora2
|
| 10 |
-
results:
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 11 |
library_name: peft
|
| 12 |
---
|
| 13 |
|
|
@@ -16,7 +28,7 @@ should probably proofread and complete it, then remove this comment. -->
|
|
| 16 |
|
| 17 |
# lmind_nq_train6000_eval6489_v1_reciteonly_qa_v3_Qwen_Qwen1.5-4B_5e-5_lora2
|
| 18 |
|
| 19 |
-
This model is a fine-tuned version of [Qwen/Qwen1.5-4B](https://huggingface.co/Qwen/Qwen1.5-4B) on
|
| 20 |
It achieves the following results on the evaluation set:
|
| 21 |
- Loss: 2.7806
|
| 22 |
- Accuracy: 0.5793
|
|
|
|
| 3 |
base_model: Qwen/Qwen1.5-4B
|
| 4 |
tags:
|
| 5 |
- generated_from_trainer
|
| 6 |
+
datasets:
|
| 7 |
+
- tyzhu/lmind_nq_train6000_eval6489_v1_reciteonly_qa_v3
|
| 8 |
metrics:
|
| 9 |
- accuracy
|
| 10 |
model-index:
|
| 11 |
- name: lmind_nq_train6000_eval6489_v1_reciteonly_qa_v3_Qwen_Qwen1.5-4B_5e-5_lora2
|
| 12 |
+
results:
|
| 13 |
+
- task:
|
| 14 |
+
name: Causal Language Modeling
|
| 15 |
+
type: text-generation
|
| 16 |
+
dataset:
|
| 17 |
+
name: tyzhu/lmind_nq_train6000_eval6489_v1_reciteonly_qa_v3
|
| 18 |
+
type: tyzhu/lmind_nq_train6000_eval6489_v1_reciteonly_qa_v3
|
| 19 |
+
metrics:
|
| 20 |
+
- name: Accuracy
|
| 21 |
+
type: accuracy
|
| 22 |
+
value: 0.5792914798206278
|
| 23 |
library_name: peft
|
| 24 |
---
|
| 25 |
|
|
|
|
| 28 |
|
| 29 |
# lmind_nq_train6000_eval6489_v1_reciteonly_qa_v3_Qwen_Qwen1.5-4B_5e-5_lora2
|
| 30 |
|
| 31 |
+
This model is a fine-tuned version of [Qwen/Qwen1.5-4B](https://huggingface.co/Qwen/Qwen1.5-4B) on the tyzhu/lmind_nq_train6000_eval6489_v1_reciteonly_qa_v3 dataset.
|
| 32 |
It achieves the following results on the evaluation set:
|
| 33 |
- Loss: 2.7806
|
| 34 |
- Accuracy: 0.5793
|
all_results.json
ADDED
|
@@ -0,0 +1,16 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 19.946666666666665,
|
| 3 |
+
"eval_accuracy": 0.5792914798206278,
|
| 4 |
+
"eval_loss": 2.7805869579315186,
|
| 5 |
+
"eval_runtime": 6.3407,
|
| 6 |
+
"eval_samples": 500,
|
| 7 |
+
"eval_samples_per_second": 78.856,
|
| 8 |
+
"eval_steps_per_second": 9.936,
|
| 9 |
+
"perplexity": 16.128484912424078,
|
| 10 |
+
"total_flos": 3.767212755417825e+17,
|
| 11 |
+
"train_loss": 1.0034164194117254,
|
| 12 |
+
"train_runtime": 8271.9938,
|
| 13 |
+
"train_samples": 6000,
|
| 14 |
+
"train_samples_per_second": 14.507,
|
| 15 |
+
"train_steps_per_second": 0.452
|
| 16 |
+
}
|
eval_results.json
ADDED
|
@@ -0,0 +1,10 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 19.946666666666665,
|
| 3 |
+
"eval_accuracy": 0.5792914798206278,
|
| 4 |
+
"eval_loss": 2.7805869579315186,
|
| 5 |
+
"eval_runtime": 6.3407,
|
| 6 |
+
"eval_samples": 500,
|
| 7 |
+
"eval_samples_per_second": 78.856,
|
| 8 |
+
"eval_steps_per_second": 9.936,
|
| 9 |
+
"perplexity": 16.128484912424078
|
| 10 |
+
}
|
tokenizer.json
CHANGED
|
@@ -1,11 +1,6 @@
|
|
| 1 |
{
|
| 2 |
"version": "1.0",
|
| 3 |
-
"truncation":
|
| 4 |
-
"direction": "Right",
|
| 5 |
-
"max_length": 1024,
|
| 6 |
-
"strategy": "LongestFirst",
|
| 7 |
-
"stride": 0
|
| 8 |
-
},
|
| 9 |
"padding": null,
|
| 10 |
"added_tokens": [
|
| 11 |
{
|
|
|
|
| 1 |
{
|
| 2 |
"version": "1.0",
|
| 3 |
+
"truncation": null,
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 4 |
"padding": null,
|
| 5 |
"added_tokens": [
|
| 6 |
{
|
train_results.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 19.946666666666665,
|
| 3 |
+
"total_flos": 3.767212755417825e+17,
|
| 4 |
+
"train_loss": 1.0034164194117254,
|
| 5 |
+
"train_runtime": 8271.9938,
|
| 6 |
+
"train_samples": 6000,
|
| 7 |
+
"train_samples_per_second": 14.507,
|
| 8 |
+
"train_steps_per_second": 0.452
|
| 9 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,469 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 19.946666666666665,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 3740,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.5333333333333333,
|
| 13 |
+
"grad_norm": 0.26278844475746155,
|
| 14 |
+
"learning_rate": 5e-05,
|
| 15 |
+
"loss": 1.8147,
|
| 16 |
+
"step": 100
|
| 17 |
+
},
|
| 18 |
+
{
|
| 19 |
+
"epoch": 0.9973333333333333,
|
| 20 |
+
"eval_accuracy": 0.6045112107623318,
|
| 21 |
+
"eval_loss": 1.7073266506195068,
|
| 22 |
+
"eval_runtime": 6.5303,
|
| 23 |
+
"eval_samples_per_second": 76.566,
|
| 24 |
+
"eval_steps_per_second": 9.647,
|
| 25 |
+
"step": 187
|
| 26 |
+
},
|
| 27 |
+
{
|
| 28 |
+
"epoch": 1.0666666666666667,
|
| 29 |
+
"grad_norm": 0.2971494495868683,
|
| 30 |
+
"learning_rate": 5e-05,
|
| 31 |
+
"loss": 1.7413,
|
| 32 |
+
"step": 200
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"epoch": 1.6,
|
| 36 |
+
"grad_norm": 0.36278319358825684,
|
| 37 |
+
"learning_rate": 5e-05,
|
| 38 |
+
"loss": 1.6968,
|
| 39 |
+
"step": 300
|
| 40 |
+
},
|
| 41 |
+
{
|
| 42 |
+
"epoch": 2.0,
|
| 43 |
+
"eval_accuracy": 0.606152466367713,
|
| 44 |
+
"eval_loss": 1.6898142099380493,
|
| 45 |
+
"eval_runtime": 6.5974,
|
| 46 |
+
"eval_samples_per_second": 75.787,
|
| 47 |
+
"eval_steps_per_second": 9.549,
|
| 48 |
+
"step": 375
|
| 49 |
+
},
|
| 50 |
+
{
|
| 51 |
+
"epoch": 2.1333333333333333,
|
| 52 |
+
"grad_norm": 0.4214104115962982,
|
| 53 |
+
"learning_rate": 5e-05,
|
| 54 |
+
"loss": 1.6929,
|
| 55 |
+
"step": 400
|
| 56 |
+
},
|
| 57 |
+
{
|
| 58 |
+
"epoch": 2.6666666666666665,
|
| 59 |
+
"grad_norm": 0.5143699049949646,
|
| 60 |
+
"learning_rate": 5e-05,
|
| 61 |
+
"loss": 1.6305,
|
| 62 |
+
"step": 500
|
| 63 |
+
},
|
| 64 |
+
{
|
| 65 |
+
"epoch": 2.997333333333333,
|
| 66 |
+
"eval_accuracy": 0.6071031390134529,
|
| 67 |
+
"eval_loss": 1.6911797523498535,
|
| 68 |
+
"eval_runtime": 6.4685,
|
| 69 |
+
"eval_samples_per_second": 77.298,
|
| 70 |
+
"eval_steps_per_second": 9.74,
|
| 71 |
+
"step": 562
|
| 72 |
+
},
|
| 73 |
+
{
|
| 74 |
+
"epoch": 3.2,
|
| 75 |
+
"grad_norm": 0.6533758044242859,
|
| 76 |
+
"learning_rate": 5e-05,
|
| 77 |
+
"loss": 1.6037,
|
| 78 |
+
"step": 600
|
| 79 |
+
},
|
| 80 |
+
{
|
| 81 |
+
"epoch": 3.7333333333333334,
|
| 82 |
+
"grad_norm": 0.6560174226760864,
|
| 83 |
+
"learning_rate": 5e-05,
|
| 84 |
+
"loss": 1.5527,
|
| 85 |
+
"step": 700
|
| 86 |
+
},
|
| 87 |
+
{
|
| 88 |
+
"epoch": 4.0,
|
| 89 |
+
"eval_accuracy": 0.606152466367713,
|
| 90 |
+
"eval_loss": 1.7046079635620117,
|
| 91 |
+
"eval_runtime": 6.4983,
|
| 92 |
+
"eval_samples_per_second": 76.943,
|
| 93 |
+
"eval_steps_per_second": 9.695,
|
| 94 |
+
"step": 750
|
| 95 |
+
},
|
| 96 |
+
{
|
| 97 |
+
"epoch": 4.266666666666667,
|
| 98 |
+
"grad_norm": 0.8206190466880798,
|
| 99 |
+
"learning_rate": 5e-05,
|
| 100 |
+
"loss": 1.5014,
|
| 101 |
+
"step": 800
|
| 102 |
+
},
|
| 103 |
+
{
|
| 104 |
+
"epoch": 4.8,
|
| 105 |
+
"grad_norm": 0.8708794116973877,
|
| 106 |
+
"learning_rate": 5e-05,
|
| 107 |
+
"loss": 1.4642,
|
| 108 |
+
"step": 900
|
| 109 |
+
},
|
| 110 |
+
{
|
| 111 |
+
"epoch": 4.997333333333334,
|
| 112 |
+
"eval_accuracy": 0.6055156950672645,
|
| 113 |
+
"eval_loss": 1.7308964729309082,
|
| 114 |
+
"eval_runtime": 6.1199,
|
| 115 |
+
"eval_samples_per_second": 81.701,
|
| 116 |
+
"eval_steps_per_second": 10.294,
|
| 117 |
+
"step": 937
|
| 118 |
+
},
|
| 119 |
+
{
|
| 120 |
+
"epoch": 5.333333333333333,
|
| 121 |
+
"grad_norm": 0.9805185198783875,
|
| 122 |
+
"learning_rate": 5e-05,
|
| 123 |
+
"loss": 1.379,
|
| 124 |
+
"step": 1000
|
| 125 |
+
},
|
| 126 |
+
{
|
| 127 |
+
"epoch": 5.866666666666667,
|
| 128 |
+
"grad_norm": 0.9485296010971069,
|
| 129 |
+
"learning_rate": 5e-05,
|
| 130 |
+
"loss": 1.3868,
|
| 131 |
+
"step": 1100
|
| 132 |
+
},
|
| 133 |
+
{
|
| 134 |
+
"epoch": 6.0,
|
| 135 |
+
"eval_accuracy": 0.6036502242152466,
|
| 136 |
+
"eval_loss": 1.770605206489563,
|
| 137 |
+
"eval_runtime": 5.6622,
|
| 138 |
+
"eval_samples_per_second": 88.305,
|
| 139 |
+
"eval_steps_per_second": 11.126,
|
| 140 |
+
"step": 1125
|
| 141 |
+
},
|
| 142 |
+
{
|
| 143 |
+
"epoch": 6.4,
|
| 144 |
+
"grad_norm": 1.2094453573226929,
|
| 145 |
+
"learning_rate": 5e-05,
|
| 146 |
+
"loss": 1.2855,
|
| 147 |
+
"step": 1200
|
| 148 |
+
},
|
| 149 |
+
{
|
| 150 |
+
"epoch": 6.933333333333334,
|
| 151 |
+
"grad_norm": 1.1076228618621826,
|
| 152 |
+
"learning_rate": 5e-05,
|
| 153 |
+
"loss": 1.278,
|
| 154 |
+
"step": 1300
|
| 155 |
+
},
|
| 156 |
+
{
|
| 157 |
+
"epoch": 6.997333333333334,
|
| 158 |
+
"eval_accuracy": 0.6015605381165919,
|
| 159 |
+
"eval_loss": 1.8113144636154175,
|
| 160 |
+
"eval_runtime": 6.0735,
|
| 161 |
+
"eval_samples_per_second": 82.325,
|
| 162 |
+
"eval_steps_per_second": 10.373,
|
| 163 |
+
"step": 1312
|
| 164 |
+
},
|
| 165 |
+
{
|
| 166 |
+
"epoch": 7.466666666666667,
|
| 167 |
+
"grad_norm": 1.1597787141799927,
|
| 168 |
+
"learning_rate": 5e-05,
|
| 169 |
+
"loss": 1.1879,
|
| 170 |
+
"step": 1400
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 8.0,
|
| 174 |
+
"grad_norm": 1.139780879020691,
|
| 175 |
+
"learning_rate": 5e-05,
|
| 176 |
+
"loss": 1.1816,
|
| 177 |
+
"step": 1500
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 8.0,
|
| 181 |
+
"eval_accuracy": 0.6003766816143498,
|
| 182 |
+
"eval_loss": 1.8514131307601929,
|
| 183 |
+
"eval_runtime": 6.0462,
|
| 184 |
+
"eval_samples_per_second": 82.696,
|
| 185 |
+
"eval_steps_per_second": 10.42,
|
| 186 |
+
"step": 1500
|
| 187 |
+
},
|
| 188 |
+
{
|
| 189 |
+
"epoch": 8.533333333333333,
|
| 190 |
+
"grad_norm": 1.4597896337509155,
|
| 191 |
+
"learning_rate": 5e-05,
|
| 192 |
+
"loss": 1.0781,
|
| 193 |
+
"step": 1600
|
| 194 |
+
},
|
| 195 |
+
{
|
| 196 |
+
"epoch": 8.997333333333334,
|
| 197 |
+
"eval_accuracy": 0.5972645739910314,
|
| 198 |
+
"eval_loss": 1.9348456859588623,
|
| 199 |
+
"eval_runtime": 7.2059,
|
| 200 |
+
"eval_samples_per_second": 69.387,
|
| 201 |
+
"eval_steps_per_second": 8.743,
|
| 202 |
+
"step": 1687
|
| 203 |
+
},
|
| 204 |
+
{
|
| 205 |
+
"epoch": 9.066666666666666,
|
| 206 |
+
"grad_norm": 1.6389895677566528,
|
| 207 |
+
"learning_rate": 5e-05,
|
| 208 |
+
"loss": 1.0849,
|
| 209 |
+
"step": 1700
|
| 210 |
+
},
|
| 211 |
+
{
|
| 212 |
+
"epoch": 9.6,
|
| 213 |
+
"grad_norm": 1.656604290008545,
|
| 214 |
+
"learning_rate": 5e-05,
|
| 215 |
+
"loss": 1.0042,
|
| 216 |
+
"step": 1800
|
| 217 |
+
},
|
| 218 |
+
{
|
| 219 |
+
"epoch": 10.0,
|
| 220 |
+
"eval_accuracy": 0.596932735426009,
|
| 221 |
+
"eval_loss": 1.961929440498352,
|
| 222 |
+
"eval_runtime": 6.6418,
|
| 223 |
+
"eval_samples_per_second": 75.28,
|
| 224 |
+
"eval_steps_per_second": 9.485,
|
| 225 |
+
"step": 1875
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"epoch": 10.133333333333333,
|
| 229 |
+
"grad_norm": 1.7063632011413574,
|
| 230 |
+
"learning_rate": 5e-05,
|
| 231 |
+
"loss": 0.971,
|
| 232 |
+
"step": 1900
|
| 233 |
+
},
|
| 234 |
+
{
|
| 235 |
+
"epoch": 10.666666666666666,
|
| 236 |
+
"grad_norm": 1.6970856189727783,
|
| 237 |
+
"learning_rate": 5e-05,
|
| 238 |
+
"loss": 0.9175,
|
| 239 |
+
"step": 2000
|
| 240 |
+
},
|
| 241 |
+
{
|
| 242 |
+
"epoch": 10.997333333333334,
|
| 243 |
+
"eval_accuracy": 0.5937937219730942,
|
| 244 |
+
"eval_loss": 2.043671131134033,
|
| 245 |
+
"eval_runtime": 6.266,
|
| 246 |
+
"eval_samples_per_second": 79.796,
|
| 247 |
+
"eval_steps_per_second": 10.054,
|
| 248 |
+
"step": 2062
|
| 249 |
+
},
|
| 250 |
+
{
|
| 251 |
+
"epoch": 11.2,
|
| 252 |
+
"grad_norm": 1.8025792837142944,
|
| 253 |
+
"learning_rate": 5e-05,
|
| 254 |
+
"loss": 0.8789,
|
| 255 |
+
"step": 2100
|
| 256 |
+
},
|
| 257 |
+
{
|
| 258 |
+
"epoch": 11.733333333333333,
|
| 259 |
+
"grad_norm": 1.6678303480148315,
|
| 260 |
+
"learning_rate": 5e-05,
|
| 261 |
+
"loss": 0.8259,
|
| 262 |
+
"step": 2200
|
| 263 |
+
},
|
| 264 |
+
{
|
| 265 |
+
"epoch": 12.0,
|
| 266 |
+
"eval_accuracy": 0.5909955156950673,
|
| 267 |
+
"eval_loss": 2.1339964866638184,
|
| 268 |
+
"eval_runtime": 5.6215,
|
| 269 |
+
"eval_samples_per_second": 88.945,
|
| 270 |
+
"eval_steps_per_second": 11.207,
|
| 271 |
+
"step": 2250
|
| 272 |
+
},
|
| 273 |
+
{
|
| 274 |
+
"epoch": 12.266666666666667,
|
| 275 |
+
"grad_norm": 1.7218997478485107,
|
| 276 |
+
"learning_rate": 5e-05,
|
| 277 |
+
"loss": 0.7956,
|
| 278 |
+
"step": 2300
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 12.8,
|
| 282 |
+
"grad_norm": 1.8202147483825684,
|
| 283 |
+
"learning_rate": 5e-05,
|
| 284 |
+
"loss": 0.7665,
|
| 285 |
+
"step": 2400
|
| 286 |
+
},
|
| 287 |
+
{
|
| 288 |
+
"epoch": 12.997333333333334,
|
| 289 |
+
"eval_accuracy": 0.5899013452914799,
|
| 290 |
+
"eval_loss": 2.198331832885742,
|
| 291 |
+
"eval_runtime": 6.0868,
|
| 292 |
+
"eval_samples_per_second": 82.145,
|
| 293 |
+
"eval_steps_per_second": 10.35,
|
| 294 |
+
"step": 2437
|
| 295 |
+
},
|
| 296 |
+
{
|
| 297 |
+
"epoch": 13.333333333333334,
|
| 298 |
+
"grad_norm": 2.0472333431243896,
|
| 299 |
+
"learning_rate": 5e-05,
|
| 300 |
+
"loss": 0.7085,
|
| 301 |
+
"step": 2500
|
| 302 |
+
},
|
| 303 |
+
{
|
| 304 |
+
"epoch": 13.866666666666667,
|
| 305 |
+
"grad_norm": 1.9246987104415894,
|
| 306 |
+
"learning_rate": 5e-05,
|
| 307 |
+
"loss": 0.6822,
|
| 308 |
+
"step": 2600
|
| 309 |
+
},
|
| 310 |
+
{
|
| 311 |
+
"epoch": 14.0,
|
| 312 |
+
"eval_accuracy": 0.5863856502242153,
|
| 313 |
+
"eval_loss": 2.2949180603027344,
|
| 314 |
+
"eval_runtime": 6.4688,
|
| 315 |
+
"eval_samples_per_second": 77.294,
|
| 316 |
+
"eval_steps_per_second": 9.739,
|
| 317 |
+
"step": 2625
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 14.4,
|
| 321 |
+
"grad_norm": 2.3168394565582275,
|
| 322 |
+
"learning_rate": 5e-05,
|
| 323 |
+
"loss": 0.625,
|
| 324 |
+
"step": 2700
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 14.933333333333334,
|
| 328 |
+
"grad_norm": 2.171243906021118,
|
| 329 |
+
"learning_rate": 5e-05,
|
| 330 |
+
"loss": 0.6189,
|
| 331 |
+
"step": 2800
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 14.997333333333334,
|
| 335 |
+
"eval_accuracy": 0.585390134529148,
|
| 336 |
+
"eval_loss": 2.346409559249878,
|
| 337 |
+
"eval_runtime": 6.485,
|
| 338 |
+
"eval_samples_per_second": 77.101,
|
| 339 |
+
"eval_steps_per_second": 9.715,
|
| 340 |
+
"step": 2812
|
| 341 |
+
},
|
| 342 |
+
{
|
| 343 |
+
"epoch": 15.466666666666667,
|
| 344 |
+
"grad_norm": 1.8979991674423218,
|
| 345 |
+
"learning_rate": 5e-05,
|
| 346 |
+
"loss": 0.5499,
|
| 347 |
+
"step": 2900
|
| 348 |
+
},
|
| 349 |
+
{
|
| 350 |
+
"epoch": 16.0,
|
| 351 |
+
"grad_norm": 1.9712458848953247,
|
| 352 |
+
"learning_rate": 5e-05,
|
| 353 |
+
"loss": 0.5588,
|
| 354 |
+
"step": 3000
|
| 355 |
+
},
|
| 356 |
+
{
|
| 357 |
+
"epoch": 16.0,
|
| 358 |
+
"eval_accuracy": 0.5835426008968609,
|
| 359 |
+
"eval_loss": 2.4582157135009766,
|
| 360 |
+
"eval_runtime": 6.6057,
|
| 361 |
+
"eval_samples_per_second": 75.692,
|
| 362 |
+
"eval_steps_per_second": 9.537,
|
| 363 |
+
"step": 3000
|
| 364 |
+
},
|
| 365 |
+
{
|
| 366 |
+
"epoch": 16.533333333333335,
|
| 367 |
+
"grad_norm": 2.132955312728882,
|
| 368 |
+
"learning_rate": 5e-05,
|
| 369 |
+
"loss": 0.4806,
|
| 370 |
+
"step": 3100
|
| 371 |
+
},
|
| 372 |
+
{
|
| 373 |
+
"epoch": 16.997333333333334,
|
| 374 |
+
"eval_accuracy": 0.5828609865470852,
|
| 375 |
+
"eval_loss": 2.53446102142334,
|
| 376 |
+
"eval_runtime": 5.6263,
|
| 377 |
+
"eval_samples_per_second": 88.868,
|
| 378 |
+
"eval_steps_per_second": 11.197,
|
| 379 |
+
"step": 3187
|
| 380 |
+
},
|
| 381 |
+
{
|
| 382 |
+
"epoch": 17.066666666666666,
|
| 383 |
+
"grad_norm": 2.198476552963257,
|
| 384 |
+
"learning_rate": 5e-05,
|
| 385 |
+
"loss": 0.4894,
|
| 386 |
+
"step": 3200
|
| 387 |
+
},
|
| 388 |
+
{
|
| 389 |
+
"epoch": 17.6,
|
| 390 |
+
"grad_norm": 2.3147218227386475,
|
| 391 |
+
"learning_rate": 5e-05,
|
| 392 |
+
"loss": 0.4272,
|
| 393 |
+
"step": 3300
|
| 394 |
+
},
|
| 395 |
+
{
|
| 396 |
+
"epoch": 18.0,
|
| 397 |
+
"eval_accuracy": 0.5805112107623318,
|
| 398 |
+
"eval_loss": 2.630317211151123,
|
| 399 |
+
"eval_runtime": 5.7146,
|
| 400 |
+
"eval_samples_per_second": 87.495,
|
| 401 |
+
"eval_steps_per_second": 11.024,
|
| 402 |
+
"step": 3375
|
| 403 |
+
},
|
| 404 |
+
{
|
| 405 |
+
"epoch": 18.133333333333333,
|
| 406 |
+
"grad_norm": 2.428415298461914,
|
| 407 |
+
"learning_rate": 5e-05,
|
| 408 |
+
"loss": 0.4341,
|
| 409 |
+
"step": 3400
|
| 410 |
+
},
|
| 411 |
+
{
|
| 412 |
+
"epoch": 18.666666666666668,
|
| 413 |
+
"grad_norm": 2.2315878868103027,
|
| 414 |
+
"learning_rate": 5e-05,
|
| 415 |
+
"loss": 0.3805,
|
| 416 |
+
"step": 3500
|
| 417 |
+
},
|
| 418 |
+
{
|
| 419 |
+
"epoch": 18.997333333333334,
|
| 420 |
+
"eval_accuracy": 0.5805829596412556,
|
| 421 |
+
"eval_loss": 2.65861177444458,
|
| 422 |
+
"eval_runtime": 6.2726,
|
| 423 |
+
"eval_samples_per_second": 79.712,
|
| 424 |
+
"eval_steps_per_second": 10.044,
|
| 425 |
+
"step": 3562
|
| 426 |
+
},
|
| 427 |
+
{
|
| 428 |
+
"epoch": 19.2,
|
| 429 |
+
"grad_norm": 2.0963428020477295,
|
| 430 |
+
"learning_rate": 5e-05,
|
| 431 |
+
"loss": 0.3742,
|
| 432 |
+
"step": 3600
|
| 433 |
+
},
|
| 434 |
+
{
|
| 435 |
+
"epoch": 19.733333333333334,
|
| 436 |
+
"grad_norm": 2.4846127033233643,
|
| 437 |
+
"learning_rate": 5e-05,
|
| 438 |
+
"loss": 0.3421,
|
| 439 |
+
"step": 3700
|
| 440 |
+
},
|
| 441 |
+
{
|
| 442 |
+
"epoch": 19.946666666666665,
|
| 443 |
+
"eval_accuracy": 0.5792914798206278,
|
| 444 |
+
"eval_loss": 2.7805869579315186,
|
| 445 |
+
"eval_runtime": 6.1628,
|
| 446 |
+
"eval_samples_per_second": 81.132,
|
| 447 |
+
"eval_steps_per_second": 10.223,
|
| 448 |
+
"step": 3740
|
| 449 |
+
},
|
| 450 |
+
{
|
| 451 |
+
"epoch": 19.946666666666665,
|
| 452 |
+
"step": 3740,
|
| 453 |
+
"total_flos": 3.767212755417825e+17,
|
| 454 |
+
"train_loss": 1.0034164194117254,
|
| 455 |
+
"train_runtime": 8271.9938,
|
| 456 |
+
"train_samples_per_second": 14.507,
|
| 457 |
+
"train_steps_per_second": 0.452
|
| 458 |
+
}
|
| 459 |
+
],
|
| 460 |
+
"logging_steps": 100,
|
| 461 |
+
"max_steps": 3740,
|
| 462 |
+
"num_input_tokens_seen": 0,
|
| 463 |
+
"num_train_epochs": 20,
|
| 464 |
+
"save_steps": 500,
|
| 465 |
+
"total_flos": 3.767212755417825e+17,
|
| 466 |
+
"train_batch_size": 1,
|
| 467 |
+
"trial_name": null,
|
| 468 |
+
"trial_params": null
|
| 469 |
+
}
|