MutionHydra/GPT-3.0-YDR-Mution

TEXT GENERATIONPricing:Input $0.37 / Cached $0.074 / Output $0.38Concurrent Unit Cost:1Model Size:8BQuant:FP8Context Size:8kTool Calling:SupportedPublished:Jul 24, 2026License:apache-2.0Architecture:Transformer Open Weights Featherless Exclusive Cold

MutionHydra/GPT-3.0-YDR-Mution is an 8 billion parameter language model developed by MutionHydra, finetuned from unsloth/llama-3-8b-Instruct-bnb-4bit. This model was trained using Unsloth and Huggingface's TRL library, enabling faster training. It is designed for general language tasks, leveraging its Llama-3 base and 8192 token context length.

Loading preview...

Model Overview

MutionHydra/GPT-3.0-YDR-Mution is an 8 billion parameter language model, developed by MutionHydra. It is a finetuned version of the unsloth/llama-3-8b-Instruct-bnb-4bit model, indicating its foundation in the Llama-3 architecture. The training process utilized Unsloth and Huggingface's TRL library, which is noted for enabling faster training.

Key Characteristics

  • Base Model: Finetuned from unsloth/llama-3-8b-Instruct-bnb-4bit, inheriting the capabilities of the Llama-3 family.
  • Parameter Count: 8 billion parameters, offering a balance between performance and computational efficiency.
  • Training Efficiency: Benefits from Unsloth's optimizations, leading to a 2x faster training time.
  • Context Length: Supports an 8192 token context window, suitable for handling moderately long inputs and generating coherent responses.

Use Cases

This model is suitable for a variety of general-purpose language tasks, including:

  • Text generation and completion
  • Instruction following
  • Summarization
  • Question answering

Its Apache-2.0 license allows for flexible use in various applications.

Finetunes
1 models