You can deploy this model and start earning money today!

RichardErkhov/
ChenWeiLi_-_Med-ChimeraLlama-3-8B_SHERP-gguf

Name: ChenWeiLi_-_Med-ChimeraLlama-3-8B_SHERP-gguf
Availability: InStock

endpoints_compatible

gguf

template

Views

No views yet

Quantization made by Richard Erkhov.

Github

Discord

Request more models

Med-ChimeraLlama-3-8B_SHERP - GGUF

Model creator: https://huggingface.co/ChenWeiLi/
Original model: https://huggingface.co/ChenWeiLi/Med-ChimeraLlama-3-8B_SHERP/

Name	Quant method	Size
Med-ChimeraLlama-3-8B_SHERP.Q2_K.gguf	Q2_K	2.96GB
Med-ChimeraLlama-3-8B_SHERP.IQ3_XS.gguf	IQ3_XS	3.28GB
Med-ChimeraLlama-3-8B_SHERP.IQ3_S.gguf	IQ3_S	3.43GB
Med-ChimeraLlama-3-8B_SHERP.Q3_K_S.gguf	Q3_K_S	3.41GB
Med-ChimeraLlama-3-8B_SHERP.IQ3_M.gguf	IQ3_M	3.52GB
Med-ChimeraLlama-3-8B_SHERP.Q3_K.gguf	Q3_K	3.74GB
Med-ChimeraLlama-3-8B_SHERP.Q3_K_M.gguf	Q3_K_M	3.74GB
Med-ChimeraLlama-3-8B_SHERP.Q3_K_L.gguf	Q3_K_L	4.03GB
Med-ChimeraLlama-3-8B_SHERP.IQ4_XS.gguf	IQ4_XS	4.18GB
Med-ChimeraLlama-3-8B_SHERP.Q4_0.gguf	Q4_0	4.34GB
Med-ChimeraLlama-3-8B_SHERP.IQ4_NL.gguf	IQ4_NL	4.38GB
Med-ChimeraLlama-3-8B_SHERP.Q4_K_S.gguf	Q4_K_S	4.37GB
Med-ChimeraLlama-3-8B_SHERP.Q4_K.gguf	Q4_K	4.58GB
Med-ChimeraLlama-3-8B_SHERP.Q4_K_M.gguf	Q4_K_M	4.58GB
Med-ChimeraLlama-3-8B_SHERP.Q4_1.gguf	Q4_1	4.78GB
Med-ChimeraLlama-3-8B_SHERP.Q5_0.gguf	Q5_0	5.21GB
Med-ChimeraLlama-3-8B_SHERP.Q5_K_S.gguf	Q5_K_S	5.21GB
Med-ChimeraLlama-3-8B_SHERP.Q5_K.gguf	Q5_K	5.34GB
Med-ChimeraLlama-3-8B_SHERP.Q5_K_M.gguf	Q5_K_M	5.34GB
Med-ChimeraLlama-3-8B_SHERP.Q5_1.gguf	Q5_1	5.65GB
Med-ChimeraLlama-3-8B_SHERP.Q6_K.gguf	Q6_K	6.14GB
Med-ChimeraLlama-3-8B_SHERP.Q8_0.gguf	Q8_0	7.95GB

Original model description:

base_model:

mlabonne/ChimeraLlama-3-8B-v3
johnsnowlabs/JSL-MedLlama-3-8B-v2.0 library_name: transformers tags:
mergekit
merge license: llama3

Chimera_MedLlama-3-8B

This is a merge of pre-trained language models created using mergekit.

Merge Details

Merge Method

This model was merged using the SLERP merge method.

Models Merged

The following models were included in the merge:

Evaluation

multimedqa (0 shot)

Tasks	Version	Filter	Metric	Value		Stderr
- medmcqa	Yaml	none	acc	0.6087	±	0.0075
		none	acc_norm	0.6087	±	0.0075
- medqa_4options	Yaml	none	acc	0.6269	±	0.0136
		none	acc_norm	0.6269	±	0.0136
- anatomy (mmlu)	0	none	acc	0.6963	±	0.0397
- clinical_knowledge (mmlu)	0	none	acc	0.7585	±	0.0263
- college_biology (mmlu)	0	none	acc	0.7847	±	0.0344
- college_medicine (mmlu)	0	none	acc	0.6936	±	0.0351
- medical_genetics (mmlu)	0	none	acc	0.8200	±	0.0386
- professional_medicine (mmlu)	0	none	acc	0.7684	±	0.0256
stem	N/A	none	acc_norm	0.6129	±	0.0066
		none	acc	0.6440	±	0.0057
- pubmedqa	1	none	acc	0.7480	±	0.0194

Groups	Version	Filter	n-shot	Metric	Value		Stderr
stem	N/A	none	0	acc_norm	0.6129	±	0.0066
		none	0	acc	0.6440	±	0.0057

Configuration

The following YAML configuration was used to produce this model:

yaml

slices:
  - sources:
      - model: mlabonne/ChimeraLlama-3-8B-v3
        layer_range: [0, 32]
      - model: johnsnowlabs/JSL-MedLlama-3-8B-v2.0
        layer_range: [0, 32]
merge_method: slerp
base_model:  mlabonne/ChimeraLlama-3-8B-v3
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

RichardErkhov/ChenWeiLi_-_Med-ChimeraLlama-3-8B_SHERP-gguf