metadata

license: apache-2.0
model-index:
  - name: Chupacabra-7B-v2
    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: 65.19
            name: normalized accuracy
        source:
          url: >-
            https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=perlthoughts/Chupacabra-7B-v2
          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: 83.39
            name: normalized accuracy
        source:
          url: >-
            https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=perlthoughts/Chupacabra-7B-v2
          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: 63.6
            name: accuracy
        source:
          url: >-
            https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=perlthoughts/Chupacabra-7B-v2
          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: 57.17
        source:
          url: >-
            https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=perlthoughts/Chupacabra-7B-v2
          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: 78.14
            name: accuracy
        source:
          url: >-
            https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=perlthoughts/Chupacabra-7B-v2
          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: 54.74
            name: accuracy
        source:
          url: >-
            https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=perlthoughts/Chupacabra-7B-v2
          name: Open LLM Leaderboard

Chupacabra 7B v2

Model Description

This model was made by merging models based on Mistral with the SLERP merge method.

Advantages of SLERP vs averaging weights(common) are as follows:

Spherical Linear Interpolation (SLERP) - Traditionally, model merging often resorts to weight averaging which, although straightforward, might not always capture the intricate features of the models being merged. The SLERP technique addresses this limitation, producing a blended model with characteristics smoothly interpolated from both parent models, ensuring the resultant model captures the essence of both its parents.
Smooth Transitions - SLERP ensures smoother transitions between model parameters. This is especially significant when interpolating between high-dimensional vectors.
Better Preservation of Characteristics - Unlike weight averaging, which might dilute distinct features, SLERP preserves the curvature and characteristics of both models in high-dimensional spaces.
Nuanced Blending - SLERP takes into account the geometric and rotational properties of the models in the vector space, resulting in a blend that is more reflective of both parent models' characteristics.

List of all models and merging path is coming soon.

Purpose

Merging the "thick"est model weights from mistral models using amazing training methods like direct preference optimization (DPO), supervised fine tuning (SFT) and reinforced learning.

I have spent countless hours studying the latest research papers, attending conferences, and networking with experts in the field. I experimented with different algorithms, tactics, fine-tuned hyperparameters, optimizers, and optimized code until I achieved the best possible results.

It has not been without challenges. There were skeptics who doubted my abilities and questioned my approach. My approach can be changed, but a closed mind cannot.

I refused to let their negativity bring me down. Instead, I used their doubts as fuel to push myself even harder. I worked tirelessly (vapenation), day and night, until I finally succeeded in merging with the most performant model weights using SOTA training methods like DPO and other advanced techniques described above.

Thank you openchat 3.5 for showing me the way.

"Hate it or love it, the underdogs on top." - The Game

Here is my contribution.

Prompt Template

Replace {system} with your system prompt, and {prompt} with your prompt instruction.

<|im_start|>system
{system}<|im_end|>
<|im_start|>user
{prompt}<|im_end|>
<|im_start|>assistant

Bug fixes

Fixed issue with generation and the incorrect model weights. Model weights have been corrected and now generation works again. Reuploading GGUF to the GGUF repository as well as the AWQ versions.
Fixed issue with tokenizer not stopping correctly and changed prompt template.
Uploaded new merged model weights.

More info

Developed by: Ray Hernandez
Model type: Mistral
Language(s) (NLP): English
License: Apache 2.0

Model Sources [optional]

Uses

Direct Use

[More Information Needed]

Downstream Use [optional]

[More Information Needed]

Out-of-Scope Use

[More Information Needed]

Bias, Risks, and Limitations

[More Information Needed]

Recommendations

Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.

How to Get Started with the Model

Use the code below to get started with the model.

[More Information Needed]

Training Details

Training Data

[More Information Needed]

Training Procedure

Preprocessing [optional]

[More Information Needed]

Training Hyperparameters

Training regime: [More Information Needed]

Speeds, Sizes, Times [optional]

[More Information Needed]

Evaluation

Testing Data, Factors & Metrics

Testing Data

[More Information Needed]

Factors

[More Information Needed]

Metrics

[More Information Needed]

Results

[More Information Needed]

Summary

Model Examination [optional]

[More Information Needed]

Technical Specifications [optional]

Model Architecture and Objective

[More Information Needed]

Compute Infrastructure

[More Information Needed]

Hardware

[More Information Needed]

Software

[More Information Needed]

Citation [optional]

BibTeX:

[More Information Needed]

APA:

[More Information Needed]

Glossary [optional]

[More Information Needed]

More Information [optional]

[More Information Needed]

Model Card Authors [optional]

[More Information Needed]

Model Card Contact

[More Information Needed]

Open LLM Leaderboard Evaluation Results

Detailed results can be found here

Metric	Value
Avg.	67.04
AI2 Reasoning Challenge (25-Shot)	65.19
HellaSwag (10-Shot)	83.39
MMLU (5-Shot)	63.60
TruthfulQA (0-shot)	57.17
Winogrande (5-shot)	78.14
GSM8k (5-shot)	54.74