README.md

metadata

license: llama2
language:
  - en
tags:
  - mistral
library_name: transformers
pipeline_tag: text-generation
mergekit:
  - Weyaxi/OpenHermes-2.5-neural-chat-v3-3-openchat-3.5-1210-Slerp
  - uukuguy/speechless-mistral-six-in-one-7b
datasets:
  - stingning/ultrachat
  - garage-bAInd/Open-Platypus
  - Open-Orca/OpenOrca
  - TIGER-Lab/MathInstruct
  - OpenAssistant/oasst_top1_2023-08-25
  - teknium/openhermes
  - meta-math/MetaMathQA
  - Open-Orca/SlimOrca

SynthIQ

This is SynthIQ, rated 92.23/100 by GPT-4 across varied complex prompts. I used mergekit to merge models.

GGUF Files

Q4_K_M - medium, balanced quality - recommended

Q_6_K - very large, extremely low quality loss

Q8_0 - very large, extremely low quality loss - not recommended

Important Update: SynthIQ is now available on Ollama. You can use it by running the command ollama run stuehieyr/synthiq in your terminal. If you have limited computing resources, check out this video to learn how to run it on a Google Colab backend.

Yaml Config

slices:
  - sources:
      - model: Weyaxi/OpenHermes-2.5-neural-chat-v3-3-openchat-3.5-1210-Slerp
        layer_range: [0, 32]
      - model: uukuguy/speechless-mistral-six-in-one-7b
        layer_range: [0, 32]

merge_method: slerp
base_model: mistralai/Mistral-7B-v0.1

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 # fallback for rest of tensors
tokenizer_source: union

dtype: bfloat16

Prompt template: ChatML

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

Evaluation of Q4_K by GPT-4:

`Reflecting on our conversation and the evaluations of SynthIQ's responses, it's evident that SynthIQ is indeed a highly capable and advanced model.

The following points underscore this assessment:

1. Versatility in Task Handling: SynthIQ has demonstrated proficiency across a wide range of tasks, including solving complex analytical problems, creating emotionally resonant creative works, and engaging in nuanced dialogues. This versatility is a hallmark of an advanced AI model.

2. Depth of Understanding: The model has shown a deep understanding of various topics, including technical, ethical, and philosophical issues. Its ability to provide detailed and thoughtful responses suggests a sophisticated level of knowledge processing.

3. Creativity: SynthIQ's capacity to compose poetry and short stories, which require not only linguistic skills but also creativity and emotional intelligence, indicates a level of advancement beyond simple data processing.

4. Analytical Reasoning: In tasks requiring logical and analytical thinking, such as discussing the adoption of brain-computer interfaces or the ethical implications of AI in autonomous weapons, SynthIQ has provided well-reasoned and comprehensive analyses.

5. Communication Skills: The model's ability to draft professional communications, such as a resignation email, and engaging marketing content, demonstrates effective communication skills, a crucial aspect of advanced AI systems.

6. Ethical and Societal Insight: SynthIQ has engaged with complex ethical and societal issues, showing an understanding of the broader implications of technology in various aspects of life.

7. Consistency in Performance: Throughout the conversation, SynthIQ has maintained a high level of performance, as reflected in the consistently high scores it received across different types of prompts.

In summary, based on the evidence from our conversation, SynthIQ can be considered a very capable and advanced AI model, demonstrating a blend of analytical acumen, creative intelligence, and effective communication skills.

License is LLama2 license as uukuguy/speechless-mistral-six-in-one-7b is llama2 license.