End of training
Browse files- README.md +15 -9
- all_results.json +13 -0
- config.json +1 -1
- eval_results.json +16 -0
README.md
CHANGED
@@ -3,9 +3,15 @@ library_name: transformers
|
|
3 |
license: llama3
|
4 |
base_model: meta-llama/Meta-Llama-3-8B-Instruct
|
5 |
tags:
|
|
|
6 |
- trl
|
7 |
- simpo
|
8 |
- generated_from_trainer
|
|
|
|
|
|
|
|
|
|
|
9 |
model-index:
|
10 |
- name: llama3instruct_-qfUNL-10-0_3-1e-6-1_best
|
11 |
results: []
|
@@ -16,17 +22,17 @@ should probably proofread and complete it, then remove this comment. -->
|
|
16 |
|
17 |
# llama3instruct_-qfUNL-10-0_3-1e-6-1_best
|
18 |
|
19 |
-
This model is a fine-tuned version of [meta-llama/Meta-Llama-3-8B-Instruct](https://huggingface.co/meta-llama/Meta-Llama-3-8B-Instruct) on
|
20 |
It achieves the following results on the evaluation set:
|
21 |
-
- Loss: 1.
|
22 |
-
- Rewards/chosen: -5.
|
23 |
-
- Rewards/rejected: -7.
|
24 |
- Rewards/accuracies: 0.7711
|
25 |
-
- Rewards/margins: 1.
|
26 |
-
- Logps/rejected: -0.
|
27 |
-
- Logps/chosen: -0.
|
28 |
-
- Logits/rejected: -1.
|
29 |
-
- Logits/chosen: -1.
|
30 |
|
31 |
## Model description
|
32 |
|
|
|
3 |
license: llama3
|
4 |
base_model: meta-llama/Meta-Llama-3-8B-Instruct
|
5 |
tags:
|
6 |
+
- alignment-handbook
|
7 |
- trl
|
8 |
- simpo
|
9 |
- generated_from_trainer
|
10 |
+
- trl
|
11 |
+
- simpo
|
12 |
+
- generated_from_trainer
|
13 |
+
datasets:
|
14 |
+
- yakazimir/llama3-ultrafeedback-armorm
|
15 |
model-index:
|
16 |
- name: llama3instruct_-qfUNL-10-0_3-1e-6-1_best
|
17 |
results: []
|
|
|
22 |
|
23 |
# llama3instruct_-qfUNL-10-0_3-1e-6-1_best
|
24 |
|
25 |
+
This model is a fine-tuned version of [meta-llama/Meta-Llama-3-8B-Instruct](https://huggingface.co/meta-llama/Meta-Llama-3-8B-Instruct) on the yakazimir/llama3-ultrafeedback-armorm dataset.
|
26 |
It achieves the following results on the evaluation set:
|
27 |
+
- Loss: 1.8781
|
28 |
+
- Rewards/chosen: -5.4315
|
29 |
+
- Rewards/rejected: -7.2730
|
30 |
- Rewards/accuracies: 0.7711
|
31 |
+
- Rewards/margins: 1.8415
|
32 |
+
- Logps/rejected: -0.7273
|
33 |
+
- Logps/chosen: -0.5432
|
34 |
+
- Logits/rejected: -1.3495
|
35 |
+
- Logits/chosen: -1.3907
|
36 |
|
37 |
## Model description
|
38 |
|
all_results.json
CHANGED
@@ -1,5 +1,18 @@
|
|
1 |
{
|
2 |
"epoch": 0.9989071038251366,
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
3 |
"total_flos": 0.0,
|
4 |
"train_loss": 2.2740620700744363,
|
5 |
"train_runtime": 5957.0658,
|
|
|
1 |
{
|
2 |
"epoch": 0.9989071038251366,
|
3 |
+
"eval_logits/chosen": -1.3906669616699219,
|
4 |
+
"eval_logits/rejected": -1.3494510650634766,
|
5 |
+
"eval_logps/chosen": -0.5431549549102783,
|
6 |
+
"eval_logps/rejected": -0.7273014187812805,
|
7 |
+
"eval_loss": 1.8781323432922363,
|
8 |
+
"eval_rewards/accuracies": 0.7710843086242676,
|
9 |
+
"eval_rewards/chosen": -5.431549549102783,
|
10 |
+
"eval_rewards/margins": 1.841464638710022,
|
11 |
+
"eval_rewards/rejected": -7.273014068603516,
|
12 |
+
"eval_runtime": 33.4862,
|
13 |
+
"eval_samples": 1318,
|
14 |
+
"eval_samples_per_second": 39.359,
|
15 |
+
"eval_steps_per_second": 2.479,
|
16 |
"total_flos": 0.0,
|
17 |
"train_loss": 2.2740620700744363,
|
18 |
"train_runtime": 5957.0658,
|
config.json
CHANGED
@@ -24,6 +24,6 @@
|
|
24 |
"tie_word_embeddings": false,
|
25 |
"torch_dtype": "bfloat16",
|
26 |
"transformers_version": "4.44.2",
|
27 |
-
"use_cache":
|
28 |
"vocab_size": 128256
|
29 |
}
|
|
|
24 |
"tie_word_embeddings": false,
|
25 |
"torch_dtype": "bfloat16",
|
26 |
"transformers_version": "4.44.2",
|
27 |
+
"use_cache": true,
|
28 |
"vocab_size": 128256
|
29 |
}
|
eval_results.json
ADDED
@@ -0,0 +1,16 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"epoch": 0.9989071038251366,
|
3 |
+
"eval_logits/chosen": -1.3906669616699219,
|
4 |
+
"eval_logits/rejected": -1.3494510650634766,
|
5 |
+
"eval_logps/chosen": -0.5431549549102783,
|
6 |
+
"eval_logps/rejected": -0.7273014187812805,
|
7 |
+
"eval_loss": 1.8781323432922363,
|
8 |
+
"eval_rewards/accuracies": 0.7710843086242676,
|
9 |
+
"eval_rewards/chosen": -5.431549549102783,
|
10 |
+
"eval_rewards/margins": 1.841464638710022,
|
11 |
+
"eval_rewards/rejected": -7.273014068603516,
|
12 |
+
"eval_runtime": 33.4862,
|
13 |
+
"eval_samples": 1318,
|
14 |
+
"eval_samples_per_second": 39.359,
|
15 |
+
"eval_steps_per_second": 2.479
|
16 |
+
}
|