README.md

metadata

language:
  - en
license: apache-2.0
library_name: transformers
tags:
  - merge
  - mergekit
  - lazymergekit
  - mistral
  - roleplay
  - ResplendentAI/Datura_7B
  - Epiculous/Mika-7B
base_model:
  - ResplendentAI/Datura_7B
  - Epiculous/Mika-7B
model-index:
  - name: Foxglove_7B
    results:
      - task:
          type: text-generation
          name: Text Generation
        dataset:
          name: AI2 Reasoning Challenge (25-Shot)
          type: ai2_arc
          config: ARC-Challenge
          split: test
          args:
            num_few_shot: 25
        metrics:
          - type: acc_norm
            value: 67.83
            name: normalized accuracy
        source:
          url: >-
            https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=aridoverrun/Foxglove_7B
          name: Open LLM Leaderboard
      - task:
          type: text-generation
          name: Text Generation
        dataset:
          name: HellaSwag (10-Shot)
          type: hellaswag
          split: validation
          args:
            num_few_shot: 10
        metrics:
          - type: acc_norm
            value: 86.57
            name: normalized accuracy
        source:
          url: >-
            https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=aridoverrun/Foxglove_7B
          name: Open LLM Leaderboard
      - task:
          type: text-generation
          name: Text Generation
        dataset:
          name: MMLU (5-Shot)
          type: cais/mmlu
          config: all
          split: test
          args:
            num_few_shot: 5
        metrics:
          - type: acc
            value: 62.89
            name: accuracy
        source:
          url: >-
            https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=aridoverrun/Foxglove_7B
          name: Open LLM Leaderboard
      - task:
          type: text-generation
          name: Text Generation
        dataset:
          name: TruthfulQA (0-shot)
          type: truthful_qa
          config: multiple_choice
          split: validation
          args:
            num_few_shot: 0
        metrics:
          - type: mc2
            value: 69.64
        source:
          url: >-
            https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=aridoverrun/Foxglove_7B
          name: Open LLM Leaderboard
      - task:
          type: text-generation
          name: Text Generation
        dataset:
          name: Winogrande (5-shot)
          type: winogrande
          config: winogrande_xl
          split: validation
          args:
            num_few_shot: 5
        metrics:
          - type: acc
            value: 80.74
            name: accuracy
        source:
          url: >-
            https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=aridoverrun/Foxglove_7B
          name: Open LLM Leaderboard
      - task:
          type: text-generation
          name: Text Generation
        dataset:
          name: GSM8k (5-shot)
          type: gsm8k
          config: main
          split: test
          args:
            num_few_shot: 5
        metrics:
          - type: acc
            value: 44.96
            name: accuracy
        source:
          url: >-
            https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=aridoverrun/Foxglove_7B
          name: Open LLM Leaderboard

🌸 Foxglove_7B

Foxglove is a well-rounded RP model. It is smart, does a great job of sticking to character card, and is proficient at following desired markdown.

Foxglove_7B is a merge of the following models using LazyMergekit:

• ResplendentAI/Datura_7B
• Epiculous/Mika-7B

Quantizations

Thanks to mradermacher, static GGUF quants are available here.

Formatting/Preset

Alpaca works best, but Mistral provides good outputs as well.

Configuration

slices:
  - sources:
      - model: ResplendentAI/Datura_7B
        layer_range: [0, 32]
      - model: Epiculous/Mika-7B
        layer_range: [0, 32]
merge_method: slerp
base_model: ResplendentAI/Datura_7B
parameters:
  t:
    - filter: self_attn
      value: [0, 0.7, 0.4, 0.6, 1]  
    - filter: mlp
      value: [0.8, 0.5, 0.7, 0.3, 0]  
    - value: 0.6  
dtype: bfloat16

Usage

!pip install -qU transformers accelerate

from transformers import AutoTokenizer
import transformers
import torch

model = "rmdhirr/Foxglove_7B"
messages = [{"role": "user", "content": "What is a large language model?"}]

tokenizer = AutoTokenizer.from_pretrained(model)
prompt = tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)
pipeline = transformers.pipeline(
    "text-generation",
    model=model,
    torch_dtype=torch.float16,
    device_map="auto",
)

outputs = pipeline(prompt, max_new_tokens=256, do_sample=True, temperature=0.7, top_k=50, top_p=0.95)
print(outputs[0]["generated_text"])

Open LLM Leaderboard Evaluation Results

Detailed results can be found here

Metric	Value
Avg.	68.77
AI2 Reasoning Challenge (25-Shot)	67.83
HellaSwag (10-Shot)	86.57
MMLU (5-Shot)	62.89
TruthfulQA (0-shot)	69.64
Winogrande (5-shot)	80.74
GSM8k (5-shot)	44.96