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_3e-4_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_3e-4_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: 3.0184
|
| 22 |
- Accuracy: 0.5886
|
|
|
|
| 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_3e-4_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: 0.5885829596412556
|
| 23 |
library_name: peft
|
| 24 |
---
|
| 25 |
|
|
|
|
| 28 |
|
| 29 |
# lmind_nq_train6000_eval6489_v1_reciteonly_qa_v3_Qwen_Qwen1.5-4B_3e-4_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: 3.0184
|
| 34 |
- Accuracy: 0.5886
|
all_results.json
ADDED
|
@@ -0,0 +1,16 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 19.946666666666665,
|
| 3 |
+
"eval_accuracy": 0.5885829596412556,
|
| 4 |
+
"eval_loss": 3.0184407234191895,
|
| 5 |
+
"eval_runtime": 7.0977,
|
| 6 |
+
"eval_samples": 500,
|
| 7 |
+
"eval_samples_per_second": 70.445,
|
| 8 |
+
"eval_steps_per_second": 8.876,
|
| 9 |
+
"perplexity": 20.45936499070068,
|
| 10 |
+
"total_flos": 3.767212755417825e+17,
|
| 11 |
+
"train_loss": 0.406913380316872,
|
| 12 |
+
"train_runtime": 8744.2637,
|
| 13 |
+
"train_samples": 6000,
|
| 14 |
+
"train_samples_per_second": 13.723,
|
| 15 |
+
"train_steps_per_second": 0.428
|
| 16 |
+
}
|
eval_results.json
ADDED
|
@@ -0,0 +1,10 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 19.946666666666665,
|
| 3 |
+
"eval_accuracy": 0.5885829596412556,
|
| 4 |
+
"eval_loss": 3.0184407234191895,
|
| 5 |
+
"eval_runtime": 7.0977,
|
| 6 |
+
"eval_samples": 500,
|
| 7 |
+
"eval_samples_per_second": 70.445,
|
| 8 |
+
"eval_steps_per_second": 8.876,
|
| 9 |
+
"perplexity": 20.45936499070068
|
| 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": 0.406913380316872,
|
| 5 |
+
"train_runtime": 8744.2637,
|
| 6 |
+
"train_samples": 6000,
|
| 7 |
+
"train_samples_per_second": 13.723,
|
| 8 |
+
"train_steps_per_second": 0.428
|
| 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": 0.2082328200340271,
|
| 14 |
+
"learning_rate": 0.0003,
|
| 15 |
+
"loss": 1.7516,
|
| 16 |
+
"step": 100
|
| 17 |
+
},
|
| 18 |
+
{
|
| 19 |
+
"epoch": 0.9973333333333333,
|
| 20 |
+
"eval_accuracy": 0.608609865470852,
|
| 21 |
+
"eval_loss": 1.6714181900024414,
|
| 22 |
+
"eval_runtime": 6.6041,
|
| 23 |
+
"eval_samples_per_second": 75.71,
|
| 24 |
+
"eval_steps_per_second": 9.54,
|
| 25 |
+
"step": 187
|
| 26 |
+
},
|
| 27 |
+
{
|
| 28 |
+
"epoch": 1.0666666666666667,
|
| 29 |
+
"grad_norm": 0.2450818121433258,
|
| 30 |
+
"learning_rate": 0.0003,
|
| 31 |
+
"loss": 1.6997,
|
| 32 |
+
"step": 200
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"epoch": 1.6,
|
| 36 |
+
"grad_norm": 0.36302751302719116,
|
| 37 |
+
"learning_rate": 0.0003,
|
| 38 |
+
"loss": 1.5219,
|
| 39 |
+
"step": 300
|
| 40 |
+
},
|
| 41 |
+
{
|
| 42 |
+
"epoch": 2.0,
|
| 43 |
+
"eval_accuracy": 0.6104035874439462,
|
| 44 |
+
"eval_loss": 1.6736148595809937,
|
| 45 |
+
"eval_runtime": 7.0545,
|
| 46 |
+
"eval_samples_per_second": 70.877,
|
| 47 |
+
"eval_steps_per_second": 8.93,
|
| 48 |
+
"step": 375
|
| 49 |
+
},
|
| 50 |
+
{
|
| 51 |
+
"epoch": 2.1333333333333333,
|
| 52 |
+
"grad_norm": 0.4031146764755249,
|
| 53 |
+
"learning_rate": 0.0003,
|
| 54 |
+
"loss": 1.4473,
|
| 55 |
+
"step": 400
|
| 56 |
+
},
|
| 57 |
+
{
|
| 58 |
+
"epoch": 2.6666666666666665,
|
| 59 |
+
"grad_norm": 0.4989413917064667,
|
| 60 |
+
"learning_rate": 0.0003,
|
| 61 |
+
"loss": 1.2037,
|
| 62 |
+
"step": 500
|
| 63 |
+
},
|
| 64 |
+
{
|
| 65 |
+
"epoch": 2.997333333333333,
|
| 66 |
+
"eval_accuracy": 0.6081076233183856,
|
| 67 |
+
"eval_loss": 1.756110429763794,
|
| 68 |
+
"eval_runtime": 6.5632,
|
| 69 |
+
"eval_samples_per_second": 76.183,
|
| 70 |
+
"eval_steps_per_second": 9.599,
|
| 71 |
+
"step": 562
|
| 72 |
+
},
|
| 73 |
+
{
|
| 74 |
+
"epoch": 3.2,
|
| 75 |
+
"grad_norm": 0.5431676506996155,
|
| 76 |
+
"learning_rate": 0.0003,
|
| 77 |
+
"loss": 1.0868,
|
| 78 |
+
"step": 600
|
| 79 |
+
},
|
| 80 |
+
{
|
| 81 |
+
"epoch": 3.7333333333333334,
|
| 82 |
+
"grad_norm": 0.5269047021865845,
|
| 83 |
+
"learning_rate": 0.0003,
|
| 84 |
+
"loss": 0.8815,
|
| 85 |
+
"step": 700
|
| 86 |
+
},
|
| 87 |
+
{
|
| 88 |
+
"epoch": 4.0,
|
| 89 |
+
"eval_accuracy": 0.6032645739910314,
|
| 90 |
+
"eval_loss": 1.8874716758728027,
|
| 91 |
+
"eval_runtime": 6.1584,
|
| 92 |
+
"eval_samples_per_second": 81.19,
|
| 93 |
+
"eval_steps_per_second": 10.23,
|
| 94 |
+
"step": 750
|
| 95 |
+
},
|
| 96 |
+
{
|
| 97 |
+
"epoch": 4.266666666666667,
|
| 98 |
+
"grad_norm": 0.681224524974823,
|
| 99 |
+
"learning_rate": 0.0003,
|
| 100 |
+
"loss": 0.7337,
|
| 101 |
+
"step": 800
|
| 102 |
+
},
|
| 103 |
+
{
|
| 104 |
+
"epoch": 4.8,
|
| 105 |
+
"grad_norm": 0.7092007994651794,
|
| 106 |
+
"learning_rate": 0.0003,
|
| 107 |
+
"loss": 0.6016,
|
| 108 |
+
"step": 900
|
| 109 |
+
},
|
| 110 |
+
{
|
| 111 |
+
"epoch": 4.997333333333334,
|
| 112 |
+
"eval_accuracy": 0.5979730941704036,
|
| 113 |
+
"eval_loss": 2.076803684234619,
|
| 114 |
+
"eval_runtime": 6.7069,
|
| 115 |
+
"eval_samples_per_second": 74.55,
|
| 116 |
+
"eval_steps_per_second": 9.393,
|
| 117 |
+
"step": 937
|
| 118 |
+
},
|
| 119 |
+
{
|
| 120 |
+
"epoch": 5.333333333333333,
|
| 121 |
+
"grad_norm": 0.6530160903930664,
|
| 122 |
+
"learning_rate": 0.0003,
|
| 123 |
+
"loss": 0.4625,
|
| 124 |
+
"step": 1000
|
| 125 |
+
},
|
| 126 |
+
{
|
| 127 |
+
"epoch": 5.866666666666667,
|
| 128 |
+
"grad_norm": 0.7848784923553467,
|
| 129 |
+
"learning_rate": 0.0003,
|
| 130 |
+
"loss": 0.3979,
|
| 131 |
+
"step": 1100
|
| 132 |
+
},
|
| 133 |
+
{
|
| 134 |
+
"epoch": 6.0,
|
| 135 |
+
"eval_accuracy": 0.5953094170403588,
|
| 136 |
+
"eval_loss": 2.260585069656372,
|
| 137 |
+
"eval_runtime": 7.0817,
|
| 138 |
+
"eval_samples_per_second": 70.605,
|
| 139 |
+
"eval_steps_per_second": 8.896,
|
| 140 |
+
"step": 1125
|
| 141 |
+
},
|
| 142 |
+
{
|
| 143 |
+
"epoch": 6.4,
|
| 144 |
+
"grad_norm": 0.8100391030311584,
|
| 145 |
+
"learning_rate": 0.0003,
|
| 146 |
+
"loss": 0.2797,
|
| 147 |
+
"step": 1200
|
| 148 |
+
},
|
| 149 |
+
{
|
| 150 |
+
"epoch": 6.933333333333334,
|
| 151 |
+
"grad_norm": 0.723619282245636,
|
| 152 |
+
"learning_rate": 0.0003,
|
| 153 |
+
"loss": 0.2591,
|
| 154 |
+
"step": 1300
|
| 155 |
+
},
|
| 156 |
+
{
|
| 157 |
+
"epoch": 6.997333333333334,
|
| 158 |
+
"eval_accuracy": 0.5932645739910314,
|
| 159 |
+
"eval_loss": 2.4669973850250244,
|
| 160 |
+
"eval_runtime": 7.1955,
|
| 161 |
+
"eval_samples_per_second": 69.488,
|
| 162 |
+
"eval_steps_per_second": 8.755,
|
| 163 |
+
"step": 1312
|
| 164 |
+
},
|
| 165 |
+
{
|
| 166 |
+
"epoch": 7.466666666666667,
|
| 167 |
+
"grad_norm": 0.6629557013511658,
|
| 168 |
+
"learning_rate": 0.0003,
|
| 169 |
+
"loss": 0.1822,
|
| 170 |
+
"step": 1400
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 8.0,
|
| 174 |
+
"grad_norm": 0.6765617728233337,
|
| 175 |
+
"learning_rate": 0.0003,
|
| 176 |
+
"loss": 0.1821,
|
| 177 |
+
"step": 1500
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 8.0,
|
| 181 |
+
"eval_accuracy": 0.592152466367713,
|
| 182 |
+
"eval_loss": 2.6145341396331787,
|
| 183 |
+
"eval_runtime": 7.0099,
|
| 184 |
+
"eval_samples_per_second": 71.327,
|
| 185 |
+
"eval_steps_per_second": 8.987,
|
| 186 |
+
"step": 1500
|
| 187 |
+
},
|
| 188 |
+
{
|
| 189 |
+
"epoch": 8.533333333333333,
|
| 190 |
+
"grad_norm": 0.5296387672424316,
|
| 191 |
+
"learning_rate": 0.0003,
|
| 192 |
+
"loss": 0.1338,
|
| 193 |
+
"step": 1600
|
| 194 |
+
},
|
| 195 |
+
{
|
| 196 |
+
"epoch": 8.997333333333334,
|
| 197 |
+
"eval_accuracy": 0.5910582959641255,
|
| 198 |
+
"eval_loss": 2.739866256713867,
|
| 199 |
+
"eval_runtime": 7.1202,
|
| 200 |
+
"eval_samples_per_second": 70.223,
|
| 201 |
+
"eval_steps_per_second": 8.848,
|
| 202 |
+
"step": 1687
|
| 203 |
+
},
|
| 204 |
+
{
|
| 205 |
+
"epoch": 9.066666666666666,
|
| 206 |
+
"grad_norm": 0.402245432138443,
|
| 207 |
+
"learning_rate": 0.0003,
|
| 208 |
+
"loss": 0.1407,
|
| 209 |
+
"step": 1700
|
| 210 |
+
},
|
| 211 |
+
{
|
| 212 |
+
"epoch": 9.6,
|
| 213 |
+
"grad_norm": 0.44858765602111816,
|
| 214 |
+
"learning_rate": 0.0003,
|
| 215 |
+
"loss": 0.1172,
|
| 216 |
+
"step": 1800
|
| 217 |
+
},
|
| 218 |
+
{
|
| 219 |
+
"epoch": 10.0,
|
| 220 |
+
"eval_accuracy": 0.5914618834080717,
|
| 221 |
+
"eval_loss": 2.833005666732788,
|
| 222 |
+
"eval_runtime": 7.0274,
|
| 223 |
+
"eval_samples_per_second": 71.15,
|
| 224 |
+
"eval_steps_per_second": 8.965,
|
| 225 |
+
"step": 1875
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"epoch": 10.133333333333333,
|
| 229 |
+
"grad_norm": 0.4508216381072998,
|
| 230 |
+
"learning_rate": 0.0003,
|
| 231 |
+
"loss": 0.1195,
|
| 232 |
+
"step": 1900
|
| 233 |
+
},
|
| 234 |
+
{
|
| 235 |
+
"epoch": 10.666666666666666,
|
| 236 |
+
"grad_norm": 0.37036266922950745,
|
| 237 |
+
"learning_rate": 0.0003,
|
| 238 |
+
"loss": 0.1102,
|
| 239 |
+
"step": 2000
|
| 240 |
+
},
|
| 241 |
+
{
|
| 242 |
+
"epoch": 10.997333333333334,
|
| 243 |
+
"eval_accuracy": 0.5914349775784753,
|
| 244 |
+
"eval_loss": 2.8673934936523438,
|
| 245 |
+
"eval_runtime": 6.5903,
|
| 246 |
+
"eval_samples_per_second": 75.869,
|
| 247 |
+
"eval_steps_per_second": 9.56,
|
| 248 |
+
"step": 2062
|
| 249 |
+
},
|
| 250 |
+
{
|
| 251 |
+
"epoch": 11.2,
|
| 252 |
+
"grad_norm": 0.3041936457157135,
|
| 253 |
+
"learning_rate": 0.0003,
|
| 254 |
+
"loss": 0.1131,
|
| 255 |
+
"step": 2100
|
| 256 |
+
},
|
| 257 |
+
{
|
| 258 |
+
"epoch": 11.733333333333333,
|
| 259 |
+
"grad_norm": 0.5073165893554688,
|
| 260 |
+
"learning_rate": 0.0003,
|
| 261 |
+
"loss": 0.1079,
|
| 262 |
+
"step": 2200
|
| 263 |
+
},
|
| 264 |
+
{
|
| 265 |
+
"epoch": 12.0,
|
| 266 |
+
"eval_accuracy": 0.5903497757847533,
|
| 267 |
+
"eval_loss": 2.894710063934326,
|
| 268 |
+
"eval_runtime": 7.0771,
|
| 269 |
+
"eval_samples_per_second": 70.65,
|
| 270 |
+
"eval_steps_per_second": 8.902,
|
| 271 |
+
"step": 2250
|
| 272 |
+
},
|
| 273 |
+
{
|
| 274 |
+
"epoch": 12.266666666666667,
|
| 275 |
+
"grad_norm": 0.4313170909881592,
|
| 276 |
+
"learning_rate": 0.0003,
|
| 277 |
+
"loss": 0.1083,
|
| 278 |
+
"step": 2300
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 12.8,
|
| 282 |
+
"grad_norm": 0.4307994544506073,
|
| 283 |
+
"learning_rate": 0.0003,
|
| 284 |
+
"loss": 0.11,
|
| 285 |
+
"step": 2400
|
| 286 |
+
},
|
| 287 |
+
{
|
| 288 |
+
"epoch": 12.997333333333334,
|
| 289 |
+
"eval_accuracy": 0.589354260089686,
|
| 290 |
+
"eval_loss": 2.9230430126190186,
|
| 291 |
+
"eval_runtime": 7.1832,
|
| 292 |
+
"eval_samples_per_second": 69.607,
|
| 293 |
+
"eval_steps_per_second": 8.771,
|
| 294 |
+
"step": 2437
|
| 295 |
+
},
|
| 296 |
+
{
|
| 297 |
+
"epoch": 13.333333333333334,
|
| 298 |
+
"grad_norm": 0.38372334837913513,
|
| 299 |
+
"learning_rate": 0.0003,
|
| 300 |
+
"loss": 0.1082,
|
| 301 |
+
"step": 2500
|
| 302 |
+
},
|
| 303 |
+
{
|
| 304 |
+
"epoch": 13.866666666666667,
|
| 305 |
+
"grad_norm": 0.49263113737106323,
|
| 306 |
+
"learning_rate": 0.0003,
|
| 307 |
+
"loss": 0.1136,
|
| 308 |
+
"step": 2600
|
| 309 |
+
},
|
| 310 |
+
{
|
| 311 |
+
"epoch": 14.0,
|
| 312 |
+
"eval_accuracy": 0.5888340807174888,
|
| 313 |
+
"eval_loss": 2.9049084186553955,
|
| 314 |
+
"eval_runtime": 6.7163,
|
| 315 |
+
"eval_samples_per_second": 74.446,
|
| 316 |
+
"eval_steps_per_second": 9.38,
|
| 317 |
+
"step": 2625
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 14.4,
|
| 321 |
+
"grad_norm": 0.42505690455436707,
|
| 322 |
+
"learning_rate": 0.0003,
|
| 323 |
+
"loss": 0.1086,
|
| 324 |
+
"step": 2700
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 14.933333333333334,
|
| 328 |
+
"grad_norm": 0.6179661750793457,
|
| 329 |
+
"learning_rate": 0.0003,
|
| 330 |
+
"loss": 0.1173,
|
| 331 |
+
"step": 2800
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 14.997333333333334,
|
| 335 |
+
"eval_accuracy": 0.5882511210762332,
|
| 336 |
+
"eval_loss": 2.8788018226623535,
|
| 337 |
+
"eval_runtime": 6.217,
|
| 338 |
+
"eval_samples_per_second": 80.425,
|
| 339 |
+
"eval_steps_per_second": 10.133,
|
| 340 |
+
"step": 2812
|
| 341 |
+
},
|
| 342 |
+
{
|
| 343 |
+
"epoch": 15.466666666666667,
|
| 344 |
+
"grad_norm": 0.45964017510414124,
|
| 345 |
+
"learning_rate": 0.0003,
|
| 346 |
+
"loss": 0.109,
|
| 347 |
+
"step": 2900
|
| 348 |
+
},
|
| 349 |
+
{
|
| 350 |
+
"epoch": 16.0,
|
| 351 |
+
"grad_norm": 0.5801168084144592,
|
| 352 |
+
"learning_rate": 0.0003,
|
| 353 |
+
"loss": 0.1163,
|
| 354 |
+
"step": 3000
|
| 355 |
+
},
|
| 356 |
+
{
|
| 357 |
+
"epoch": 16.0,
|
| 358 |
+
"eval_accuracy": 0.589237668161435,
|
| 359 |
+
"eval_loss": 2.9582040309906006,
|
| 360 |
+
"eval_runtime": 7.1226,
|
| 361 |
+
"eval_samples_per_second": 70.199,
|
| 362 |
+
"eval_steps_per_second": 8.845,
|
| 363 |
+
"step": 3000
|
| 364 |
+
},
|
| 365 |
+
{
|
| 366 |
+
"epoch": 16.533333333333335,
|
| 367 |
+
"grad_norm": 0.4587346315383911,
|
| 368 |
+
"learning_rate": 0.0003,
|
| 369 |
+
"loss": 0.1047,
|
| 370 |
+
"step": 3100
|
| 371 |
+
},
|
| 372 |
+
{
|
| 373 |
+
"epoch": 16.997333333333334,
|
| 374 |
+
"eval_accuracy": 0.5885650224215246,
|
| 375 |
+
"eval_loss": 2.9484808444976807,
|
| 376 |
+
"eval_runtime": 6.6834,
|
| 377 |
+
"eval_samples_per_second": 74.812,
|
| 378 |
+
"eval_steps_per_second": 9.426,
|
| 379 |
+
"step": 3187
|
| 380 |
+
},
|
| 381 |
+
{
|
| 382 |
+
"epoch": 17.066666666666666,
|
| 383 |
+
"grad_norm": 0.4009888470172882,
|
| 384 |
+
"learning_rate": 0.0003,
|
| 385 |
+
"loss": 0.1145,
|
| 386 |
+
"step": 3200
|
| 387 |
+
},
|
| 388 |
+
{
|
| 389 |
+
"epoch": 17.6,
|
| 390 |
+
"grad_norm": 1.4840149879455566,
|
| 391 |
+
"learning_rate": 0.0003,
|
| 392 |
+
"loss": 0.1044,
|
| 393 |
+
"step": 3300
|
| 394 |
+
},
|
| 395 |
+
{
|
| 396 |
+
"epoch": 18.0,
|
| 397 |
+
"eval_accuracy": 0.5894080717488789,
|
| 398 |
+
"eval_loss": 2.9815316200256348,
|
| 399 |
+
"eval_runtime": 7.2254,
|
| 400 |
+
"eval_samples_per_second": 69.2,
|
| 401 |
+
"eval_steps_per_second": 8.719,
|
| 402 |
+
"step": 3375
|
| 403 |
+
},
|
| 404 |
+
{
|
| 405 |
+
"epoch": 18.133333333333333,
|
| 406 |
+
"grad_norm": 0.4150511920452118,
|
| 407 |
+
"learning_rate": 0.0003,
|
| 408 |
+
"loss": 0.1111,
|
| 409 |
+
"step": 3400
|
| 410 |
+
},
|
| 411 |
+
{
|
| 412 |
+
"epoch": 18.666666666666668,
|
| 413 |
+
"grad_norm": 0.47382423281669617,
|
| 414 |
+
"learning_rate": 0.0003,
|
| 415 |
+
"loss": 0.105,
|
| 416 |
+
"step": 3500
|
| 417 |
+
},
|
| 418 |
+
{
|
| 419 |
+
"epoch": 18.997333333333334,
|
| 420 |
+
"eval_accuracy": 0.5881076233183856,
|
| 421 |
+
"eval_loss": 2.987971544265747,
|
| 422 |
+
"eval_runtime": 7.3831,
|
| 423 |
+
"eval_samples_per_second": 67.722,
|
| 424 |
+
"eval_steps_per_second": 8.533,
|
| 425 |
+
"step": 3562
|
| 426 |
+
},
|
| 427 |
+
{
|
| 428 |
+
"epoch": 19.2,
|
| 429 |
+
"grad_norm": 0.4355124533176422,
|
| 430 |
+
"learning_rate": 0.0003,
|
| 431 |
+
"loss": 0.1068,
|
| 432 |
+
"step": 3600
|
| 433 |
+
},
|
| 434 |
+
{
|
| 435 |
+
"epoch": 19.733333333333334,
|
| 436 |
+
"grad_norm": 0.43823131918907166,
|
| 437 |
+
"learning_rate": 0.0003,
|
| 438 |
+
"loss": 0.1036,
|
| 439 |
+
"step": 3700
|
| 440 |
+
},
|
| 441 |
+
{
|
| 442 |
+
"epoch": 19.946666666666665,
|
| 443 |
+
"eval_accuracy": 0.5885829596412556,
|
| 444 |
+
"eval_loss": 3.0184407234191895,
|
| 445 |
+
"eval_runtime": 6.2011,
|
| 446 |
+
"eval_samples_per_second": 80.631,
|
| 447 |
+
"eval_steps_per_second": 10.159,
|
| 448 |
+
"step": 3740
|
| 449 |
+
},
|
| 450 |
+
{
|
| 451 |
+
"epoch": 19.946666666666665,
|
| 452 |
+
"step": 3740,
|
| 453 |
+
"total_flos": 3.767212755417825e+17,
|
| 454 |
+
"train_loss": 0.406913380316872,
|
| 455 |
+
"train_runtime": 8744.2637,
|
| 456 |
+
"train_samples_per_second": 13.723,
|
| 457 |
+
"train_steps_per_second": 0.428
|
| 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 |
+
}
|