End of training
Browse files- README.md +9 -8
- all_results.json +13 -0
- config.json +35 -0
- eval_results.json +16 -0
README.md
CHANGED
@@ -2,6 +2,7 @@
|
|
2 |
license: mit
|
3 |
library_name: peft
|
4 |
tags:
|
|
|
5 |
- trl
|
6 |
- dpo
|
7 |
- generated_from_trainer
|
@@ -19,15 +20,15 @@ should probably proofread and complete it, then remove this comment. -->
|
|
19 |
|
20 |
This model is a fine-tuned version of [microsoft/Phi-3-mini-4k-instruct](https://huggingface.co/microsoft/Phi-3-mini-4k-instruct) on the None dataset.
|
21 |
It achieves the following results on the evaluation set:
|
22 |
-
- Loss: 0.
|
23 |
-
- Rewards/chosen: -0.
|
24 |
-
- Rewards/rejected: -0.
|
25 |
- Rewards/accuracies: 0.7540
|
26 |
-
- Rewards/margins: 0.
|
27 |
-
- Logps/rejected: -459.
|
28 |
-
- Logps/chosen: -
|
29 |
-
- Logits/rejected: 12.
|
30 |
-
- Logits/chosen: 13.
|
31 |
|
32 |
## Model description
|
33 |
|
|
|
2 |
license: mit
|
3 |
library_name: peft
|
4 |
tags:
|
5 |
+
- alignment-handbook
|
6 |
- trl
|
7 |
- dpo
|
8 |
- generated_from_trainer
|
|
|
20 |
|
21 |
This model is a fine-tuned version of [microsoft/Phi-3-mini-4k-instruct](https://huggingface.co/microsoft/Phi-3-mini-4k-instruct) on the None dataset.
|
22 |
It achieves the following results on the evaluation set:
|
23 |
+
- Loss: 0.6130
|
24 |
+
- Rewards/chosen: -0.4194
|
25 |
+
- Rewards/rejected: -0.5933
|
26 |
- Rewards/accuracies: 0.7540
|
27 |
+
- Rewards/margins: 0.1739
|
28 |
+
- Logps/rejected: -459.6432
|
29 |
+
- Logps/chosen: -448.1436
|
30 |
+
- Logits/rejected: 12.5287
|
31 |
+
- Logits/chosen: 13.8414
|
32 |
|
33 |
## Model description
|
34 |
|
all_results.json
CHANGED
@@ -1,5 +1,18 @@
|
|
1 |
{
|
2 |
"epoch": 0.9991111111111111,
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
3 |
"total_flos": 0.0,
|
4 |
"train_loss": 0.6475430355801701,
|
5 |
"train_runtime": 7558.4306,
|
|
|
1 |
{
|
2 |
"epoch": 0.9991111111111111,
|
3 |
+
"eval_logits/chosen": 13.841368675231934,
|
4 |
+
"eval_logits/rejected": 12.528715133666992,
|
5 |
+
"eval_logps/chosen": -448.14361572265625,
|
6 |
+
"eval_logps/rejected": -459.6432189941406,
|
7 |
+
"eval_loss": 0.6129743456840515,
|
8 |
+
"eval_rewards/accuracies": 0.7539682388305664,
|
9 |
+
"eval_rewards/chosen": -0.41938650608062744,
|
10 |
+
"eval_rewards/margins": 0.1739136427640915,
|
11 |
+
"eval_rewards/rejected": -0.5933001041412354,
|
12 |
+
"eval_runtime": 89.7858,
|
13 |
+
"eval_samples": 1000,
|
14 |
+
"eval_samples_per_second": 11.138,
|
15 |
+
"eval_steps_per_second": 0.702,
|
16 |
"total_flos": 0.0,
|
17 |
"train_loss": 0.6475430355801701,
|
18 |
"train_runtime": 7558.4306,
|
config.json
ADDED
@@ -0,0 +1,35 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_name_or_path": "microsoft/Phi-3-mini-4k-instruct",
|
3 |
+
"architectures": [
|
4 |
+
"Phi3ForCausalLM"
|
5 |
+
],
|
6 |
+
"attention_bias": false,
|
7 |
+
"attention_dropout": 0.0,
|
8 |
+
"auto_map": {
|
9 |
+
"AutoConfig": "microsoft/Phi-3-mini-4k-instruct--configuration_phi3.Phi3Config",
|
10 |
+
"AutoModelForCausalLM": "microsoft/Phi-3-mini-4k-instruct--modeling_phi3.Phi3ForCausalLM"
|
11 |
+
},
|
12 |
+
"bos_token_id": 1,
|
13 |
+
"embd_pdrop": 0.0,
|
14 |
+
"eos_token_id": 32000,
|
15 |
+
"hidden_act": "silu",
|
16 |
+
"hidden_size": 3072,
|
17 |
+
"initializer_range": 0.02,
|
18 |
+
"intermediate_size": 8192,
|
19 |
+
"max_position_embeddings": 4096,
|
20 |
+
"model_type": "phi3",
|
21 |
+
"num_attention_heads": 32,
|
22 |
+
"num_hidden_layers": 32,
|
23 |
+
"num_key_value_heads": 32,
|
24 |
+
"original_max_position_embeddings": 4096,
|
25 |
+
"pad_token_id": 32000,
|
26 |
+
"resid_pdrop": 0.0,
|
27 |
+
"rms_norm_eps": 1e-05,
|
28 |
+
"rope_scaling": null,
|
29 |
+
"rope_theta": 10000.0,
|
30 |
+
"sliding_window": 2047,
|
31 |
+
"tie_word_embeddings": false,
|
32 |
+
"transformers_version": "4.42.3",
|
33 |
+
"use_cache": true,
|
34 |
+
"vocab_size": 32064
|
35 |
+
}
|
eval_results.json
ADDED
@@ -0,0 +1,16 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"epoch": 0.9991111111111111,
|
3 |
+
"eval_logits/chosen": 13.841368675231934,
|
4 |
+
"eval_logits/rejected": 12.528715133666992,
|
5 |
+
"eval_logps/chosen": -448.14361572265625,
|
6 |
+
"eval_logps/rejected": -459.6432189941406,
|
7 |
+
"eval_loss": 0.6129743456840515,
|
8 |
+
"eval_rewards/accuracies": 0.7539682388305664,
|
9 |
+
"eval_rewards/chosen": -0.41938650608062744,
|
10 |
+
"eval_rewards/margins": 0.1739136427640915,
|
11 |
+
"eval_rewards/rejected": -0.5933001041412354,
|
12 |
+
"eval_runtime": 89.7858,
|
13 |
+
"eval_samples": 1000,
|
14 |
+
"eval_samples_per_second": 11.138,
|
15 |
+
"eval_steps_per_second": 0.702
|
16 |
+
}
|