End of training
Browse files- README.md +12 -9
- all_results.json +13 -0
- config.json +30 -0
- eval_results.json +16 -0
README.md
CHANGED
@@ -1,10 +1,13 @@
|
|
1 |
---
|
2 |
library_name: peft
|
3 |
tags:
|
|
|
4 |
- trl
|
5 |
- dpo
|
6 |
- generated_from_trainer
|
7 |
base_model: data/Meta-Llama-3-8B-Instruct-Merged
|
|
|
|
|
8 |
model-index:
|
9 |
- name: Meta-Llama-3-8B-Instruct-DPO-QLoRA
|
10 |
results: []
|
@@ -16,17 +19,17 @@ should probably proofread and complete it, then remove this comment. -->
|
|
16 |
[<img src="https://raw.githubusercontent.com/wandb/assets/main/wandb-github-badge-28.svg" alt="Visualize in Weights & Biases" width="200" height="32"/>](https://wandb.ai/statking/huggingface/runs/ficbttt2)
|
17 |
# Meta-Llama-3-8B-Instruct-DPO-QLoRA
|
18 |
|
19 |
-
This model
|
20 |
It achieves the following results on the evaluation set:
|
21 |
- Loss: 0.4785
|
22 |
-
- Rewards/chosen: -2.
|
23 |
-
- Rewards/rejected: -3.
|
24 |
-
- Rewards/accuracies: 0.
|
25 |
-
- Rewards/margins: 1.
|
26 |
-
- Logps/rejected: -604.
|
27 |
-
- Logps/chosen: -507.
|
28 |
-
- Logits/rejected: -0.
|
29 |
-
- Logits/chosen: -0.
|
30 |
|
31 |
## Model description
|
32 |
|
|
|
1 |
---
|
2 |
library_name: peft
|
3 |
tags:
|
4 |
+
- alignment-handbook
|
5 |
- trl
|
6 |
- dpo
|
7 |
- generated_from_trainer
|
8 |
base_model: data/Meta-Llama-3-8B-Instruct-Merged
|
9 |
+
datasets:
|
10 |
+
- HuggingFaceH4/ultrafeedback_binarized
|
11 |
model-index:
|
12 |
- name: Meta-Llama-3-8B-Instruct-DPO-QLoRA
|
13 |
results: []
|
|
|
19 |
[<img src="https://raw.githubusercontent.com/wandb/assets/main/wandb-github-badge-28.svg" alt="Visualize in Weights & Biases" width="200" height="32"/>](https://wandb.ai/statking/huggingface/runs/ficbttt2)
|
20 |
# Meta-Llama-3-8B-Instruct-DPO-QLoRA
|
21 |
|
22 |
+
This model is a fine-tuned version of [data/Meta-Llama-3-8B-Instruct-Merged](https://huggingface.co/data/Meta-Llama-3-8B-Instruct-Merged) on the HuggingFaceH4/ultrafeedback_binarized dataset.
|
23 |
It achieves the following results on the evaluation set:
|
24 |
- Loss: 0.4785
|
25 |
+
- Rewards/chosen: -2.3087
|
26 |
+
- Rewards/rejected: -3.5097
|
27 |
+
- Rewards/accuracies: 0.7760
|
28 |
+
- Rewards/margins: 1.2010
|
29 |
+
- Logps/rejected: -604.2300
|
30 |
+
- Logps/chosen: -507.2661
|
31 |
+
- Logits/rejected: -0.8568
|
32 |
+
- Logits/chosen: -0.8381
|
33 |
|
34 |
## Model description
|
35 |
|
all_results.json
CHANGED
@@ -1,5 +1,18 @@
|
|
1 |
{
|
2 |
"epoch": 0.9997382884061764,
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
3 |
"total_flos": 0.0,
|
4 |
"train_loss": 0.504705511212973,
|
5 |
"train_runtime": 53071.3067,
|
|
|
1 |
{
|
2 |
"epoch": 0.9997382884061764,
|
3 |
+
"eval_logits/chosen": -0.8381435871124268,
|
4 |
+
"eval_logits/rejected": -0.8568249344825745,
|
5 |
+
"eval_logps/chosen": -507.26611328125,
|
6 |
+
"eval_logps/rejected": -604.2300415039062,
|
7 |
+
"eval_loss": 0.4785204529762268,
|
8 |
+
"eval_rewards/accuracies": 0.7760000228881836,
|
9 |
+
"eval_rewards/chosen": -2.3087284564971924,
|
10 |
+
"eval_rewards/margins": 1.200958251953125,
|
11 |
+
"eval_rewards/rejected": -3.5096867084503174,
|
12 |
+
"eval_runtime": 490.7419,
|
13 |
+
"eval_samples": 2000,
|
14 |
+
"eval_samples_per_second": 4.075,
|
15 |
+
"eval_steps_per_second": 0.255,
|
16 |
"total_flos": 0.0,
|
17 |
"train_loss": 0.504705511212973,
|
18 |
"train_runtime": 53071.3067,
|
config.json
ADDED
@@ -0,0 +1,30 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_name_or_path": "data/Meta-Llama-3-8B-Instruct-Merged",
|
3 |
+
"architectures": [
|
4 |
+
"LlamaForCausalLM"
|
5 |
+
],
|
6 |
+
"attention_bias": false,
|
7 |
+
"attention_dropout": 0.0,
|
8 |
+
"bos_token_id": 128256,
|
9 |
+
"eos_token_id": 128257,
|
10 |
+
"hidden_act": "silu",
|
11 |
+
"hidden_size": 4096,
|
12 |
+
"initializer_range": 0.02,
|
13 |
+
"intermediate_size": 14336,
|
14 |
+
"max_position_embeddings": 8192,
|
15 |
+
"mlp_bias": false,
|
16 |
+
"model_type": "llama",
|
17 |
+
"num_attention_heads": 32,
|
18 |
+
"num_hidden_layers": 32,
|
19 |
+
"num_key_value_heads": 8,
|
20 |
+
"pad_token_id": 128257,
|
21 |
+
"pretraining_tp": 1,
|
22 |
+
"rms_norm_eps": 1e-05,
|
23 |
+
"rope_scaling": null,
|
24 |
+
"rope_theta": 500000.0,
|
25 |
+
"tie_word_embeddings": false,
|
26 |
+
"torch_dtype": "bfloat16",
|
27 |
+
"transformers_version": "4.41.0",
|
28 |
+
"use_cache": true,
|
29 |
+
"vocab_size": 128258
|
30 |
+
}
|
eval_results.json
ADDED
@@ -0,0 +1,16 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"epoch": 0.9997382884061764,
|
3 |
+
"eval_logits/chosen": -0.8381435871124268,
|
4 |
+
"eval_logits/rejected": -0.8568249344825745,
|
5 |
+
"eval_logps/chosen": -507.26611328125,
|
6 |
+
"eval_logps/rejected": -604.2300415039062,
|
7 |
+
"eval_loss": 0.4785204529762268,
|
8 |
+
"eval_rewards/accuracies": 0.7760000228881836,
|
9 |
+
"eval_rewards/chosen": -2.3087284564971924,
|
10 |
+
"eval_rewards/margins": 1.200958251953125,
|
11 |
+
"eval_rewards/rejected": -3.5096867084503174,
|
12 |
+
"eval_runtime": 490.7419,
|
13 |
+
"eval_samples": 2000,
|
14 |
+
"eval_samples_per_second": 4.075,
|
15 |
+
"eval_steps_per_second": 0.255
|
16 |
+
}
|