RichardErkhov/Heng666_-_StableMedZephyr-Merged-3b-gguf

Quantization made by Richard Erkhov.

Github

Discord

Request more models

StableMedZephyr-Merged-3b - GGUF

• Model creator: https://huggingface.co/Heng666/
• Original model: https://huggingface.co/Heng666/StableMedZephyr-Merged-3b/

Name	Quant method	Size
StableMedZephyr-Merged-3b.Q2_K.gguf	Q2_K	1.01GB
StableMedZephyr-Merged-3b.Q3_K_S.gguf	Q3_K_S	1.17GB
StableMedZephyr-Merged-3b.Q3_K.gguf	Q3_K	1.3GB
StableMedZephyr-Merged-3b.Q3_K_M.gguf	Q3_K_M	1.3GB
StableMedZephyr-Merged-3b.Q3_K_L.gguf	Q3_K_L	1.4GB
StableMedZephyr-Merged-3b.IQ4_XS.gguf	IQ4_XS	1.43GB
StableMedZephyr-Merged-3b.Q4_0.gguf	Q4_0	1.5GB
StableMedZephyr-Merged-3b.IQ4_NL.gguf	IQ4_NL	1.51GB
StableMedZephyr-Merged-3b.Q4_K_S.gguf	Q4_K_S	1.51GB
StableMedZephyr-Merged-3b.Q4_K.gguf	Q4_K	1.59GB
StableMedZephyr-Merged-3b.Q4_K_M.gguf	Q4_K_M	1.59GB
StableMedZephyr-Merged-3b.Q4_1.gguf	Q4_1	1.65GB
StableMedZephyr-Merged-3b.Q5_0.gguf	Q5_0	1.81GB
StableMedZephyr-Merged-3b.Q5_K_S.gguf	Q5_K_S	1.81GB
StableMedZephyr-Merged-3b.Q5_K.gguf	Q5_K	1.86GB
StableMedZephyr-Merged-3b.Q5_K_M.gguf	Q5_K_M	1.86GB
StableMedZephyr-Merged-3b.Q5_1.gguf	Q5_1	1.96GB
StableMedZephyr-Merged-3b.Q6_K.gguf	Q6_K	2.14GB
StableMedZephyr-Merged-3b.Q8_0.gguf	Q8_0	2.77GB

Original model description:

license: other language: - en library_name: transformers pipeline_tag: text-generation tags: - causal-lm - text-generation-inference - merge

FOR EXPERIMENT

Description

stabilityai/stablelm-zephyr-3b, StableMed-3b merged with a new, experimental implementation of "dare ties" via mergekit. See:

Language Models are Super Mario: Absorbing Abilities from Homologous Models as a Free Lunch

https://github.com/cg123/mergekit/tree/dare

Usage

StableLM Zephyr 3B uses the following instruction format:

<|user|>
List 3 synonyms for the word "tiny"<|endoftext|>
<|assistant|>
1. Dwarf
2. Little
3. Petite<|endoftext|>

---

Testing Notes

Merged in mergekit with the following config, and the tokenizer from chargoddard's Yi-Llama:

models:
  - model: stabilityai/stablelm-zephyr-3b
    # no parameters necessary for base model
  - model: cxllin/StableMed-3b
    parameters:
      weight: 0.08
      density: 0.5
merge_method: dare_ties
base_model: stabilityai/stablelm-zephyr-3b
parameters:
  int8_mask: true
dtype: bfloat16

Model Details

• License: StabilityAI Non-Commercial Research Community License