RichardErkhov/EmbeddedLLM_-_Mistral-7B-Merge-14-v0.1-gguf

Quantization made by Richard Erkhov.

Github

Discord

Request more models

Mistral-7B-Merge-14-v0.1 - GGUF

• Model creator: https://huggingface.co/EmbeddedLLM/
• Original model: https://huggingface.co/EmbeddedLLM/Mistral-7B-Merge-14-v0.1/

Name	Quant method	Size
Mistral-7B-Merge-14-v0.1.Q2_K.gguf	Q2_K	2.53GB
Mistral-7B-Merge-14-v0.1.IQ3_XS.gguf	IQ3_XS	2.81GB
Mistral-7B-Merge-14-v0.1.IQ3_S.gguf	IQ3_S	2.96GB
Mistral-7B-Merge-14-v0.1.Q3_K_S.gguf	Q3_K_S	2.95GB
Mistral-7B-Merge-14-v0.1.IQ3_M.gguf	IQ3_M	3.06GB
Mistral-7B-Merge-14-v0.1.Q3_K.gguf	Q3_K	3.28GB
Mistral-7B-Merge-14-v0.1.Q3_K_M.gguf	Q3_K_M	3.28GB
Mistral-7B-Merge-14-v0.1.Q3_K_L.gguf	Q3_K_L	3.56GB
Mistral-7B-Merge-14-v0.1.IQ4_XS.gguf	IQ4_XS	3.67GB
Mistral-7B-Merge-14-v0.1.Q4_0.gguf	Q4_0	3.83GB
Mistral-7B-Merge-14-v0.1.IQ4_NL.gguf	IQ4_NL	3.87GB
Mistral-7B-Merge-14-v0.1.Q4_K_S.gguf	Q4_K_S	3.86GB
Mistral-7B-Merge-14-v0.1.Q4_K.gguf	Q4_K	4.07GB
Mistral-7B-Merge-14-v0.1.Q4_K_M.gguf	Q4_K_M	4.07GB
Mistral-7B-Merge-14-v0.1.Q4_1.gguf	Q4_1	4.24GB
Mistral-7B-Merge-14-v0.1.Q5_0.gguf	Q5_0	4.65GB
Mistral-7B-Merge-14-v0.1.Q5_K_S.gguf	Q5_K_S	4.65GB
Mistral-7B-Merge-14-v0.1.Q5_K.gguf	Q5_K	4.78GB
Mistral-7B-Merge-14-v0.1.Q5_K_M.gguf	Q5_K_M	4.78GB
Mistral-7B-Merge-14-v0.1.Q5_1.gguf	Q5_1	5.07GB
Mistral-7B-Merge-14-v0.1.Q6_K.gguf	Q6_K	5.53GB
Mistral-7B-Merge-14-v0.1.Q8_0.gguf	Q8_0	7.17GB

Original model description:

license: cc-by-nc-4.0 language: - en tags: - merge base_model: - EmbeddedLLM/Mistral-7B-Merge-14-v0 - janai-hq/trinity-v1

Update 2023-12-19

In light of dataset contamination issue among the merged models raised by the community in recent days, in particular berkeley-nest/Starling-LM-7B-alpha, Q-bert/MetaMath-Cybertron-Starling, and janai-hq/trinity-v1, we decided to remake another model without the models mentioned. Additionally, their CC-by-NC-4.0 license is restrictive and thus are not suitable for an open model.

Model Description

This is an experiment to test merging 14 models using DARE TIES 🦙

The merged model is then merged again with janai-hq/trinity-v1 using Gradient SLERP. The result is a base model that performs quite well but requires some further instruction fine-tuning.

The 14 models are as follows:

1. mistralai/Mistral-7B-Instruct-v0.2
2. ehartford/dolphin-2.2.1-mistral-7b
3. SciPhi/SciPhi-Mistral-7B-32k
4. ehartford/samantha-1.2-mistral-7b
5. Arc53/docsgpt-7b-mistral
6. berkeley-nest/Starling-LM-7B-alpha
7. Q-bert/MetaMath-Cybertron-Starling
8. Open-Orca/Mistral-7B-OpenOrca
9. v1olet/v1olet_marcoroni-go-bruins-merge-7B
10. beowolx/MistralHermes-CodePro-7B-v1
11. TIGER-Lab/MAmmoTH-7B-Mistral
12. teknium/OpenHermes-2.5-Mistral-7B
13. Weyaxi/OpenHermes-2.5-neural-chat-v3-3-Slerp
14. mlabonne/NeuralHermes-2.5-Mistral-7B

• base model: mistralai/Mistral-7B-v0.1

The yaml config file for this model is here:

slices:
  - sources:
      - model: EmbeddedLLM/Mistral-7B-Merge-14-v0
        layer_range: [0, 32]
      - model: janai-hq/trinity-v1
        layer_range: [0, 32]
merge_method: slerp
base_model: EmbeddedLLM/Mistral-7B-Merge-14-v0
parameters:
  t:
    - filter: self_attn
      value: [0, 0.5, 0.3, 0.7, 1]
    - filter: mlp
      value: [1, 0.5, 0.7, 0.3, 0]
    - value: 0.5
dtype: bfloat16