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 +560 -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_hotpot_train8000_eval7405_v1_reciteonly_qa_Qwen_Qwen1.5-4B_3e-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_hotpot_train8000_eval7405_v1_reciteonly_qa_Qwen_Qwen1.5-4B_3e-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: 1.8696
|
| 22 |
- Accuracy: 0.6658
|
|
|
|
| 3 |
base_model: Qwen/Qwen1.5-4B
|
| 4 |
tags:
|
| 5 |
- generated_from_trainer
|
| 6 |
+
datasets:
|
| 7 |
+
- tyzhu/lmind_hotpot_train8000_eval7405_v1_reciteonly_qa
|
| 8 |
metrics:
|
| 9 |
- accuracy
|
| 10 |
model-index:
|
| 11 |
- name: lmind_hotpot_train8000_eval7405_v1_reciteonly_qa_Qwen_Qwen1.5-4B_3e-5_lora2
|
| 12 |
+
results:
|
| 13 |
+
- task:
|
| 14 |
+
name: Causal Language Modeling
|
| 15 |
+
type: text-generation
|
| 16 |
+
dataset:
|
| 17 |
+
name: tyzhu/lmind_hotpot_train8000_eval7405_v1_reciteonly_qa
|
| 18 |
+
type: tyzhu/lmind_hotpot_train8000_eval7405_v1_reciteonly_qa
|
| 19 |
+
metrics:
|
| 20 |
+
- name: Accuracy
|
| 21 |
+
type: accuracy
|
| 22 |
+
value: 0.6657583697234353
|
| 23 |
library_name: peft
|
| 24 |
---
|
| 25 |
|
|
|
|
| 28 |
|
| 29 |
# lmind_hotpot_train8000_eval7405_v1_reciteonly_qa_Qwen_Qwen1.5-4B_3e-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_hotpot_train8000_eval7405_v1_reciteonly_qa dataset.
|
| 32 |
It achieves the following results on the evaluation set:
|
| 33 |
- Loss: 1.8696
|
| 34 |
- Accuracy: 0.6658
|
all_results.json
ADDED
|
@@ -0,0 +1,16 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 20.0,
|
| 3 |
+
"eval_accuracy": 0.6657583697234353,
|
| 4 |
+
"eval_loss": 1.8696147203445435,
|
| 5 |
+
"eval_runtime": 8.0567,
|
| 6 |
+
"eval_samples": 500,
|
| 7 |
+
"eval_samples_per_second": 62.06,
|
| 8 |
+
"eval_steps_per_second": 7.82,
|
| 9 |
+
"perplexity": 6.485797072186353,
|
| 10 |
+
"total_flos": 9.221411586147615e+17,
|
| 11 |
+
"train_loss": 1.163707649230957,
|
| 12 |
+
"train_runtime": 11749.7477,
|
| 13 |
+
"train_samples": 8000,
|
| 14 |
+
"train_samples_per_second": 13.617,
|
| 15 |
+
"train_steps_per_second": 0.426
|
| 16 |
+
}
|
eval_results.json
ADDED
|
@@ -0,0 +1,10 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 20.0,
|
| 3 |
+
"eval_accuracy": 0.6657583697234353,
|
| 4 |
+
"eval_loss": 1.8696147203445435,
|
| 5 |
+
"eval_runtime": 8.0567,
|
| 6 |
+
"eval_samples": 500,
|
| 7 |
+
"eval_samples_per_second": 62.06,
|
| 8 |
+
"eval_steps_per_second": 7.82,
|
| 9 |
+
"perplexity": 6.485797072186353
|
| 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": 20.0,
|
| 3 |
+
"total_flos": 9.221411586147615e+17,
|
| 4 |
+
"train_loss": 1.163707649230957,
|
| 5 |
+
"train_runtime": 11749.7477,
|
| 6 |
+
"train_samples": 8000,
|
| 7 |
+
"train_samples_per_second": 13.617,
|
| 8 |
+
"train_steps_per_second": 0.426
|
| 9 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,560 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 20.0,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 5000,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.4,
|
| 13 |
+
"grad_norm": 0.1675417125225067,
|
| 14 |
+
"learning_rate": 3e-05,
|
| 15 |
+
"loss": 1.5895,
|
| 16 |
+
"step": 100
|
| 17 |
+
},
|
| 18 |
+
{
|
| 19 |
+
"epoch": 0.8,
|
| 20 |
+
"grad_norm": 0.1831749528646469,
|
| 21 |
+
"learning_rate": 3e-05,
|
| 22 |
+
"loss": 1.4742,
|
| 23 |
+
"step": 200
|
| 24 |
+
},
|
| 25 |
+
{
|
| 26 |
+
"epoch": 1.0,
|
| 27 |
+
"eval_accuracy": 0.6748151382823872,
|
| 28 |
+
"eval_loss": 1.531341314315796,
|
| 29 |
+
"eval_runtime": 9.1549,
|
| 30 |
+
"eval_samples_per_second": 54.616,
|
| 31 |
+
"eval_steps_per_second": 6.882,
|
| 32 |
+
"step": 250
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"epoch": 1.2,
|
| 36 |
+
"grad_norm": 0.22951605916023254,
|
| 37 |
+
"learning_rate": 3e-05,
|
| 38 |
+
"loss": 1.4652,
|
| 39 |
+
"step": 300
|
| 40 |
+
},
|
| 41 |
+
{
|
| 42 |
+
"epoch": 1.6,
|
| 43 |
+
"grad_norm": 0.23647047579288483,
|
| 44 |
+
"learning_rate": 3e-05,
|
| 45 |
+
"loss": 1.4601,
|
| 46 |
+
"step": 400
|
| 47 |
+
},
|
| 48 |
+
{
|
| 49 |
+
"epoch": 2.0,
|
| 50 |
+
"grad_norm": 0.2248772829771042,
|
| 51 |
+
"learning_rate": 3e-05,
|
| 52 |
+
"loss": 1.45,
|
| 53 |
+
"step": 500
|
| 54 |
+
},
|
| 55 |
+
{
|
| 56 |
+
"epoch": 2.0,
|
| 57 |
+
"eval_accuracy": 0.6756593886462882,
|
| 58 |
+
"eval_loss": 1.5196012258529663,
|
| 59 |
+
"eval_runtime": 9.0923,
|
| 60 |
+
"eval_samples_per_second": 54.992,
|
| 61 |
+
"eval_steps_per_second": 6.929,
|
| 62 |
+
"step": 500
|
| 63 |
+
},
|
| 64 |
+
{
|
| 65 |
+
"epoch": 2.4,
|
| 66 |
+
"grad_norm": 0.26751402020454407,
|
| 67 |
+
"learning_rate": 3e-05,
|
| 68 |
+
"loss": 1.4361,
|
| 69 |
+
"step": 600
|
| 70 |
+
},
|
| 71 |
+
{
|
| 72 |
+
"epoch": 2.8,
|
| 73 |
+
"grad_norm": 0.29390770196914673,
|
| 74 |
+
"learning_rate": 3e-05,
|
| 75 |
+
"loss": 1.4269,
|
| 76 |
+
"step": 700
|
| 77 |
+
},
|
| 78 |
+
{
|
| 79 |
+
"epoch": 3.0,
|
| 80 |
+
"eval_accuracy": 0.6760756914119359,
|
| 81 |
+
"eval_loss": 1.5133788585662842,
|
| 82 |
+
"eval_runtime": 9.1109,
|
| 83 |
+
"eval_samples_per_second": 54.879,
|
| 84 |
+
"eval_steps_per_second": 6.915,
|
| 85 |
+
"step": 750
|
| 86 |
+
},
|
| 87 |
+
{
|
| 88 |
+
"epoch": 3.2,
|
| 89 |
+
"grad_norm": 0.3322733938694,
|
| 90 |
+
"learning_rate": 3e-05,
|
| 91 |
+
"loss": 1.4184,
|
| 92 |
+
"step": 800
|
| 93 |
+
},
|
| 94 |
+
{
|
| 95 |
+
"epoch": 3.6,
|
| 96 |
+
"grad_norm": 0.3646068871021271,
|
| 97 |
+
"learning_rate": 3e-05,
|
| 98 |
+
"loss": 1.3973,
|
| 99 |
+
"step": 900
|
| 100 |
+
},
|
| 101 |
+
{
|
| 102 |
+
"epoch": 4.0,
|
| 103 |
+
"grad_norm": 0.4150474965572357,
|
| 104 |
+
"learning_rate": 3e-05,
|
| 105 |
+
"loss": 1.3999,
|
| 106 |
+
"step": 1000
|
| 107 |
+
},
|
| 108 |
+
{
|
| 109 |
+
"epoch": 4.0,
|
| 110 |
+
"eval_accuracy": 0.6762474526928676,
|
| 111 |
+
"eval_loss": 1.511995792388916,
|
| 112 |
+
"eval_runtime": 9.0922,
|
| 113 |
+
"eval_samples_per_second": 54.992,
|
| 114 |
+
"eval_steps_per_second": 6.929,
|
| 115 |
+
"step": 1000
|
| 116 |
+
},
|
| 117 |
+
{
|
| 118 |
+
"epoch": 4.4,
|
| 119 |
+
"grad_norm": 0.44478657841682434,
|
| 120 |
+
"learning_rate": 3e-05,
|
| 121 |
+
"loss": 1.3624,
|
| 122 |
+
"step": 1100
|
| 123 |
+
},
|
| 124 |
+
{
|
| 125 |
+
"epoch": 4.8,
|
| 126 |
+
"grad_norm": 0.48866602778434753,
|
| 127 |
+
"learning_rate": 3e-05,
|
| 128 |
+
"loss": 1.3614,
|
| 129 |
+
"step": 1200
|
| 130 |
+
},
|
| 131 |
+
{
|
| 132 |
+
"epoch": 5.0,
|
| 133 |
+
"eval_accuracy": 0.6759650655021834,
|
| 134 |
+
"eval_loss": 1.5192290544509888,
|
| 135 |
+
"eval_runtime": 9.1143,
|
| 136 |
+
"eval_samples_per_second": 54.859,
|
| 137 |
+
"eval_steps_per_second": 6.912,
|
| 138 |
+
"step": 1250
|
| 139 |
+
},
|
| 140 |
+
{
|
| 141 |
+
"epoch": 5.2,
|
| 142 |
+
"grad_norm": 0.5195505023002625,
|
| 143 |
+
"learning_rate": 3e-05,
|
| 144 |
+
"loss": 1.3431,
|
| 145 |
+
"step": 1300
|
| 146 |
+
},
|
| 147 |
+
{
|
| 148 |
+
"epoch": 5.6,
|
| 149 |
+
"grad_norm": 0.5769343972206116,
|
| 150 |
+
"learning_rate": 3e-05,
|
| 151 |
+
"loss": 1.3264,
|
| 152 |
+
"step": 1400
|
| 153 |
+
},
|
| 154 |
+
{
|
| 155 |
+
"epoch": 6.0,
|
| 156 |
+
"grad_norm": 0.5181849598884583,
|
| 157 |
+
"learning_rate": 3e-05,
|
| 158 |
+
"loss": 1.3303,
|
| 159 |
+
"step": 1500
|
| 160 |
+
},
|
| 161 |
+
{
|
| 162 |
+
"epoch": 6.0,
|
| 163 |
+
"eval_accuracy": 0.6755080058224163,
|
| 164 |
+
"eval_loss": 1.5265752077102661,
|
| 165 |
+
"eval_runtime": 9.1234,
|
| 166 |
+
"eval_samples_per_second": 54.804,
|
| 167 |
+
"eval_steps_per_second": 6.905,
|
| 168 |
+
"step": 1500
|
| 169 |
+
},
|
| 170 |
+
{
|
| 171 |
+
"epoch": 6.4,
|
| 172 |
+
"grad_norm": 0.6372528076171875,
|
| 173 |
+
"learning_rate": 3e-05,
|
| 174 |
+
"loss": 1.2883,
|
| 175 |
+
"step": 1600
|
| 176 |
+
},
|
| 177 |
+
{
|
| 178 |
+
"epoch": 6.8,
|
| 179 |
+
"grad_norm": 0.6501044034957886,
|
| 180 |
+
"learning_rate": 3e-05,
|
| 181 |
+
"loss": 1.2946,
|
| 182 |
+
"step": 1700
|
| 183 |
+
},
|
| 184 |
+
{
|
| 185 |
+
"epoch": 7.0,
|
| 186 |
+
"eval_accuracy": 0.6747045123726346,
|
| 187 |
+
"eval_loss": 1.5446096658706665,
|
| 188 |
+
"eval_runtime": 9.0797,
|
| 189 |
+
"eval_samples_per_second": 55.068,
|
| 190 |
+
"eval_steps_per_second": 6.939,
|
| 191 |
+
"step": 1750
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"epoch": 7.2,
|
| 195 |
+
"grad_norm": 0.7209816575050354,
|
| 196 |
+
"learning_rate": 3e-05,
|
| 197 |
+
"loss": 1.2705,
|
| 198 |
+
"step": 1800
|
| 199 |
+
},
|
| 200 |
+
{
|
| 201 |
+
"epoch": 7.6,
|
| 202 |
+
"grad_norm": 0.7495877742767334,
|
| 203 |
+
"learning_rate": 3e-05,
|
| 204 |
+
"loss": 1.2498,
|
| 205 |
+
"step": 1900
|
| 206 |
+
},
|
| 207 |
+
{
|
| 208 |
+
"epoch": 8.0,
|
| 209 |
+
"grad_norm": 0.681526780128479,
|
| 210 |
+
"learning_rate": 3e-05,
|
| 211 |
+
"loss": 1.2518,
|
| 212 |
+
"step": 2000
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 8.0,
|
| 216 |
+
"eval_accuracy": 0.6745036390101893,
|
| 217 |
+
"eval_loss": 1.5590205192565918,
|
| 218 |
+
"eval_runtime": 8.0597,
|
| 219 |
+
"eval_samples_per_second": 62.037,
|
| 220 |
+
"eval_steps_per_second": 7.817,
|
| 221 |
+
"step": 2000
|
| 222 |
+
},
|
| 223 |
+
{
|
| 224 |
+
"epoch": 8.4,
|
| 225 |
+
"grad_norm": 0.7470565438270569,
|
| 226 |
+
"learning_rate": 3e-05,
|
| 227 |
+
"loss": 1.2196,
|
| 228 |
+
"step": 2100
|
| 229 |
+
},
|
| 230 |
+
{
|
| 231 |
+
"epoch": 8.8,
|
| 232 |
+
"grad_norm": 0.7745229005813599,
|
| 233 |
+
"learning_rate": 3e-05,
|
| 234 |
+
"loss": 1.2082,
|
| 235 |
+
"step": 2200
|
| 236 |
+
},
|
| 237 |
+
{
|
| 238 |
+
"epoch": 9.0,
|
| 239 |
+
"eval_accuracy": 0.6740349344978166,
|
| 240 |
+
"eval_loss": 1.571682333946228,
|
| 241 |
+
"eval_runtime": 9.0961,
|
| 242 |
+
"eval_samples_per_second": 54.969,
|
| 243 |
+
"eval_steps_per_second": 6.926,
|
| 244 |
+
"step": 2250
|
| 245 |
+
},
|
| 246 |
+
{
|
| 247 |
+
"epoch": 9.2,
|
| 248 |
+
"grad_norm": 0.8478706479072571,
|
| 249 |
+
"learning_rate": 3e-05,
|
| 250 |
+
"loss": 1.2017,
|
| 251 |
+
"step": 2300
|
| 252 |
+
},
|
| 253 |
+
{
|
| 254 |
+
"epoch": 9.6,
|
| 255 |
+
"grad_norm": 0.9340612292289734,
|
| 256 |
+
"learning_rate": 3e-05,
|
| 257 |
+
"loss": 1.1742,
|
| 258 |
+
"step": 2400
|
| 259 |
+
},
|
| 260 |
+
{
|
| 261 |
+
"epoch": 10.0,
|
| 262 |
+
"grad_norm": 0.9207039475440979,
|
| 263 |
+
"learning_rate": 3e-05,
|
| 264 |
+
"loss": 1.19,
|
| 265 |
+
"step": 2500
|
| 266 |
+
},
|
| 267 |
+
{
|
| 268 |
+
"epoch": 10.0,
|
| 269 |
+
"eval_accuracy": 0.6727074235807861,
|
| 270 |
+
"eval_loss": 1.6021865606307983,
|
| 271 |
+
"eval_runtime": 9.0984,
|
| 272 |
+
"eval_samples_per_second": 54.955,
|
| 273 |
+
"eval_steps_per_second": 6.924,
|
| 274 |
+
"step": 2500
|
| 275 |
+
},
|
| 276 |
+
{
|
| 277 |
+
"epoch": 10.4,
|
| 278 |
+
"grad_norm": 0.9244349598884583,
|
| 279 |
+
"learning_rate": 3e-05,
|
| 280 |
+
"loss": 1.1299,
|
| 281 |
+
"step": 2600
|
| 282 |
+
},
|
| 283 |
+
{
|
| 284 |
+
"epoch": 10.8,
|
| 285 |
+
"grad_norm": 0.9110805988311768,
|
| 286 |
+
"learning_rate": 3e-05,
|
| 287 |
+
"loss": 1.1523,
|
| 288 |
+
"step": 2700
|
| 289 |
+
},
|
| 290 |
+
{
|
| 291 |
+
"epoch": 11.0,
|
| 292 |
+
"eval_accuracy": 0.672608442503639,
|
| 293 |
+
"eval_loss": 1.6098225116729736,
|
| 294 |
+
"eval_runtime": 9.117,
|
| 295 |
+
"eval_samples_per_second": 54.843,
|
| 296 |
+
"eval_steps_per_second": 6.91,
|
| 297 |
+
"step": 2750
|
| 298 |
+
},
|
| 299 |
+
{
|
| 300 |
+
"epoch": 11.2,
|
| 301 |
+
"grad_norm": 0.916287899017334,
|
| 302 |
+
"learning_rate": 3e-05,
|
| 303 |
+
"loss": 1.1278,
|
| 304 |
+
"step": 2800
|
| 305 |
+
},
|
| 306 |
+
{
|
| 307 |
+
"epoch": 11.6,
|
| 308 |
+
"grad_norm": 1.0008416175842285,
|
| 309 |
+
"learning_rate": 3e-05,
|
| 310 |
+
"loss": 1.0981,
|
| 311 |
+
"step": 2900
|
| 312 |
+
},
|
| 313 |
+
{
|
| 314 |
+
"epoch": 12.0,
|
| 315 |
+
"grad_norm": 0.9763438701629639,
|
| 316 |
+
"learning_rate": 3e-05,
|
| 317 |
+
"loss": 1.1193,
|
| 318 |
+
"step": 3000
|
| 319 |
+
},
|
| 320 |
+
{
|
| 321 |
+
"epoch": 12.0,
|
| 322 |
+
"eval_accuracy": 0.671589519650655,
|
| 323 |
+
"eval_loss": 1.6344681978225708,
|
| 324 |
+
"eval_runtime": 9.1836,
|
| 325 |
+
"eval_samples_per_second": 54.445,
|
| 326 |
+
"eval_steps_per_second": 6.86,
|
| 327 |
+
"step": 3000
|
| 328 |
+
},
|
| 329 |
+
{
|
| 330 |
+
"epoch": 12.4,
|
| 331 |
+
"grad_norm": 1.0682412385940552,
|
| 332 |
+
"learning_rate": 3e-05,
|
| 333 |
+
"loss": 1.0604,
|
| 334 |
+
"step": 3100
|
| 335 |
+
},
|
| 336 |
+
{
|
| 337 |
+
"epoch": 12.8,
|
| 338 |
+
"grad_norm": 1.1306017637252808,
|
| 339 |
+
"learning_rate": 3e-05,
|
| 340 |
+
"loss": 1.0736,
|
| 341 |
+
"step": 3200
|
| 342 |
+
},
|
| 343 |
+
{
|
| 344 |
+
"epoch": 13.0,
|
| 345 |
+
"eval_accuracy": 0.6707016011644833,
|
| 346 |
+
"eval_loss": 1.674833059310913,
|
| 347 |
+
"eval_runtime": 8.098,
|
| 348 |
+
"eval_samples_per_second": 61.744,
|
| 349 |
+
"eval_steps_per_second": 7.78,
|
| 350 |
+
"step": 3250
|
| 351 |
+
},
|
| 352 |
+
{
|
| 353 |
+
"epoch": 13.2,
|
| 354 |
+
"grad_norm": 1.154135823249817,
|
| 355 |
+
"learning_rate": 3e-05,
|
| 356 |
+
"loss": 1.054,
|
| 357 |
+
"step": 3300
|
| 358 |
+
},
|
| 359 |
+
{
|
| 360 |
+
"epoch": 13.6,
|
| 361 |
+
"grad_norm": 1.1169854402542114,
|
| 362 |
+
"learning_rate": 3e-05,
|
| 363 |
+
"loss": 1.0253,
|
| 364 |
+
"step": 3400
|
| 365 |
+
},
|
| 366 |
+
{
|
| 367 |
+
"epoch": 14.0,
|
| 368 |
+
"grad_norm": 1.0877137184143066,
|
| 369 |
+
"learning_rate": 3e-05,
|
| 370 |
+
"loss": 1.0414,
|
| 371 |
+
"step": 3500
|
| 372 |
+
},
|
| 373 |
+
{
|
| 374 |
+
"epoch": 14.0,
|
| 375 |
+
"eval_accuracy": 0.6701280931586608,
|
| 376 |
+
"eval_loss": 1.688016653060913,
|
| 377 |
+
"eval_runtime": 9.1712,
|
| 378 |
+
"eval_samples_per_second": 54.519,
|
| 379 |
+
"eval_steps_per_second": 6.869,
|
| 380 |
+
"step": 3500
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 14.4,
|
| 384 |
+
"grad_norm": 1.2245118618011475,
|
| 385 |
+
"learning_rate": 3e-05,
|
| 386 |
+
"loss": 0.9823,
|
| 387 |
+
"step": 3600
|
| 388 |
+
},
|
| 389 |
+
{
|
| 390 |
+
"epoch": 14.8,
|
| 391 |
+
"grad_norm": 1.2784464359283447,
|
| 392 |
+
"learning_rate": 3e-05,
|
| 393 |
+
"loss": 1.0069,
|
| 394 |
+
"step": 3700
|
| 395 |
+
},
|
| 396 |
+
{
|
| 397 |
+
"epoch": 15.0,
|
| 398 |
+
"eval_accuracy": 0.6693682678311499,
|
| 399 |
+
"eval_loss": 1.7182435989379883,
|
| 400 |
+
"eval_runtime": 9.1433,
|
| 401 |
+
"eval_samples_per_second": 54.685,
|
| 402 |
+
"eval_steps_per_second": 6.89,
|
| 403 |
+
"step": 3750
|
| 404 |
+
},
|
| 405 |
+
{
|
| 406 |
+
"epoch": 15.2,
|
| 407 |
+
"grad_norm": 1.182626724243164,
|
| 408 |
+
"learning_rate": 3e-05,
|
| 409 |
+
"loss": 0.9834,
|
| 410 |
+
"step": 3800
|
| 411 |
+
},
|
| 412 |
+
{
|
| 413 |
+
"epoch": 15.6,
|
| 414 |
+
"grad_norm": 1.3419315814971924,
|
| 415 |
+
"learning_rate": 3e-05,
|
| 416 |
+
"loss": 0.9608,
|
| 417 |
+
"step": 3900
|
| 418 |
+
},
|
| 419 |
+
{
|
| 420 |
+
"epoch": 16.0,
|
| 421 |
+
"grad_norm": 1.2352997064590454,
|
| 422 |
+
"learning_rate": 3e-05,
|
| 423 |
+
"loss": 0.9654,
|
| 424 |
+
"step": 4000
|
| 425 |
+
},
|
| 426 |
+
{
|
| 427 |
+
"epoch": 16.0,
|
| 428 |
+
"eval_accuracy": 0.6685036390101893,
|
| 429 |
+
"eval_loss": 1.7521902322769165,
|
| 430 |
+
"eval_runtime": 9.1045,
|
| 431 |
+
"eval_samples_per_second": 54.918,
|
| 432 |
+
"eval_steps_per_second": 6.92,
|
| 433 |
+
"step": 4000
|
| 434 |
+
},
|
| 435 |
+
{
|
| 436 |
+
"epoch": 16.4,
|
| 437 |
+
"grad_norm": 1.439382791519165,
|
| 438 |
+
"learning_rate": 3e-05,
|
| 439 |
+
"loss": 0.9236,
|
| 440 |
+
"step": 4100
|
| 441 |
+
},
|
| 442 |
+
{
|
| 443 |
+
"epoch": 16.8,
|
| 444 |
+
"grad_norm": 1.3681950569152832,
|
| 445 |
+
"learning_rate": 3e-05,
|
| 446 |
+
"loss": 0.9337,
|
| 447 |
+
"step": 4200
|
| 448 |
+
},
|
| 449 |
+
{
|
| 450 |
+
"epoch": 17.0,
|
| 451 |
+
"eval_accuracy": 0.6677409024745269,
|
| 452 |
+
"eval_loss": 1.7825894355773926,
|
| 453 |
+
"eval_runtime": 9.1188,
|
| 454 |
+
"eval_samples_per_second": 54.832,
|
| 455 |
+
"eval_steps_per_second": 6.909,
|
| 456 |
+
"step": 4250
|
| 457 |
+
},
|
| 458 |
+
{
|
| 459 |
+
"epoch": 17.2,
|
| 460 |
+
"grad_norm": 1.417385458946228,
|
| 461 |
+
"learning_rate": 3e-05,
|
| 462 |
+
"loss": 0.9135,
|
| 463 |
+
"step": 4300
|
| 464 |
+
},
|
| 465 |
+
{
|
| 466 |
+
"epoch": 17.6,
|
| 467 |
+
"grad_norm": 1.5673705339431763,
|
| 468 |
+
"learning_rate": 3e-05,
|
| 469 |
+
"loss": 0.8941,
|
| 470 |
+
"step": 4400
|
| 471 |
+
},
|
| 472 |
+
{
|
| 473 |
+
"epoch": 18.0,
|
| 474 |
+
"grad_norm": 1.5578211545944214,
|
| 475 |
+
"learning_rate": 3e-05,
|
| 476 |
+
"loss": 0.9,
|
| 477 |
+
"step": 4500
|
| 478 |
+
},
|
| 479 |
+
{
|
| 480 |
+
"epoch": 18.0,
|
| 481 |
+
"eval_accuracy": 0.6671615720524018,
|
| 482 |
+
"eval_loss": 1.8080195188522339,
|
| 483 |
+
"eval_runtime": 8.0689,
|
| 484 |
+
"eval_samples_per_second": 61.967,
|
| 485 |
+
"eval_steps_per_second": 7.808,
|
| 486 |
+
"step": 4500
|
| 487 |
+
},
|
| 488 |
+
{
|
| 489 |
+
"epoch": 18.4,
|
| 490 |
+
"grad_norm": 1.671644687652588,
|
| 491 |
+
"learning_rate": 3e-05,
|
| 492 |
+
"loss": 0.8522,
|
| 493 |
+
"step": 4600
|
| 494 |
+
},
|
| 495 |
+
{
|
| 496 |
+
"epoch": 18.8,
|
| 497 |
+
"grad_norm": 1.5266335010528564,
|
| 498 |
+
"learning_rate": 3e-05,
|
| 499 |
+
"loss": 0.8704,
|
| 500 |
+
"step": 4700
|
| 501 |
+
},
|
| 502 |
+
{
|
| 503 |
+
"epoch": 19.0,
|
| 504 |
+
"eval_accuracy": 0.6663289665211063,
|
| 505 |
+
"eval_loss": 1.8349848985671997,
|
| 506 |
+
"eval_runtime": 9.1076,
|
| 507 |
+
"eval_samples_per_second": 54.899,
|
| 508 |
+
"eval_steps_per_second": 6.917,
|
| 509 |
+
"step": 4750
|
| 510 |
+
},
|
| 511 |
+
{
|
| 512 |
+
"epoch": 19.2,
|
| 513 |
+
"grad_norm": 1.5969411134719849,
|
| 514 |
+
"learning_rate": 3e-05,
|
| 515 |
+
"loss": 0.8469,
|
| 516 |
+
"step": 4800
|
| 517 |
+
},
|
| 518 |
+
{
|
| 519 |
+
"epoch": 19.6,
|
| 520 |
+
"grad_norm": 1.5390815734863281,
|
| 521 |
+
"learning_rate": 3e-05,
|
| 522 |
+
"loss": 0.8398,
|
| 523 |
+
"step": 4900
|
| 524 |
+
},
|
| 525 |
+
{
|
| 526 |
+
"epoch": 20.0,
|
| 527 |
+
"grad_norm": 1.5651904344558716,
|
| 528 |
+
"learning_rate": 3e-05,
|
| 529 |
+
"loss": 0.8407,
|
| 530 |
+
"step": 5000
|
| 531 |
+
},
|
| 532 |
+
{
|
| 533 |
+
"epoch": 20.0,
|
| 534 |
+
"eval_accuracy": 0.6657583697234353,
|
| 535 |
+
"eval_loss": 1.8696147203445435,
|
| 536 |
+
"eval_runtime": 9.1023,
|
| 537 |
+
"eval_samples_per_second": 54.931,
|
| 538 |
+
"eval_steps_per_second": 6.921,
|
| 539 |
+
"step": 5000
|
| 540 |
+
},
|
| 541 |
+
{
|
| 542 |
+
"epoch": 20.0,
|
| 543 |
+
"step": 5000,
|
| 544 |
+
"total_flos": 9.221411586147615e+17,
|
| 545 |
+
"train_loss": 1.163707649230957,
|
| 546 |
+
"train_runtime": 11749.7477,
|
| 547 |
+
"train_samples_per_second": 13.617,
|
| 548 |
+
"train_steps_per_second": 0.426
|
| 549 |
+
}
|
| 550 |
+
],
|
| 551 |
+
"logging_steps": 100,
|
| 552 |
+
"max_steps": 5000,
|
| 553 |
+
"num_input_tokens_seen": 0,
|
| 554 |
+
"num_train_epochs": 20,
|
| 555 |
+
"save_steps": 500,
|
| 556 |
+
"total_flos": 9.221411586147615e+17,
|
| 557 |
+
"train_batch_size": 1,
|
| 558 |
+
"trial_name": null,
|
| 559 |
+
"trial_params": null
|
| 560 |
+
}
|