---
license: mit
tags:
- moe
- DPO
- RL-TUNED
---
# Truthful_DPO_TomGrc_FusionNet_7Bx2_MoE_13B

## Description
This repo contains GGUF format model files for Truthful_DPO_TomGrc_FusionNet_7Bx2_MoE_13B.

## Files Provided
|                           Name                          |  Quant  | Bits | File Size |              Remark              |
| ------------------------------------------------------- | ------- | ---- | --------- | -------------------------------- |
| truthful_dpo_tomgrc_fusionnet_7bx2_moe_13b.IQ3_XXS.gguf | IQ3_XXS |  3   |   5.30 GB | 3.06 bpw quantization            |
| truthful_dpo_tomgrc_fusionnet_7bx2_moe_13b.IQ3_S.gguf   | IQ3_S   |  3   |   5.60 GB | 3.44 bpw quantization            |
| truthful_dpo_tomgrc_fusionnet_7bx2_moe_13b.IQ3_M.gguf   | IQ3_M   |  3   |   5.74 GB | 3.66 bpw quantization mix        |
| truthful_dpo_tomgrc_fusionnet_7bx2_moe_13b.Q4_0.gguf    | Q4_0    |  4   |   7.28 GB | 3.56G, +0.2166 ppl               |
| truthful_dpo_tomgrc_fusionnet_7bx2_moe_13b.IQ4_NL.gguf  | IQ4_NL  |  4   |   7.36 GB | 4.25 bpw non-linear quantization |
| truthful_dpo_tomgrc_fusionnet_7bx2_moe_13b.Q4_K_M.gguf  | Q4_K_M  |  4   |   7.78 GB | 3.80G, +0.0532 ppl               |
| truthful_dpo_tomgrc_fusionnet_7bx2_moe_13b.Q5_K_M.gguf  | Q5_K_M  |  5   |   9.13 GB | 4.45G, +0.0122 ppl               |
| truthful_dpo_tomgrc_fusionnet_7bx2_moe_13b.Q6_K.gguf    | Q6_K    |  6   |  10.57 GB | 5.15G, +0.0008 ppl               |
| truthful_dpo_tomgrc_fusionnet_7bx2_moe_13b.Q8_0.gguf    | Q8_0    |  8   |  13.69 GB | 6.70G, +0.0004 ppl               |

## Parameters
| path                                                   | type    | architecture       | rope_theta | sliding_win | max_pos_embed |
| ------------------------------------------------------ | ------- | ------------------ | ---------- | ----------- | ------------- |
| yunconglong/Truthful_DPO_TomGrc_FusionNet_7Bx2_MoE_13B | mixtral | MixtralForCausalLM | 10000.0    | null        | 32768         |

## Benchmarks
![](https://i.ibb.co/QNW0WJr/Truthful-DPO-Tom-Grc-Fusion-Net-7-Bx2-Mo-E-13-B.png)

# Original Model Card

 * [DPO Trainer](https://huggingface.co/docs/trl/main/en/dpo_trainer) with dataset jondurbin/truthy-dpo-v0.1 to improve [TomGrc/FusionNet_7Bx2_MoE_14B]
   ```
DPO Trainer
TRL supports the DPO Trainer for training language models from preference data, as described in the paper Direct Preference Optimization: Your Language Model is Secretly a Reward Model by Rafailov et al., 2023. 
   ```