eganwo/mistral7b-distilled-from-deepseek-r1-qwen32b

TEXT GENERATIONPricing:Input $0.4 / Cached $0.08 / Output $0.8Concurrent Unit Cost:1Model Size:7BQuant:FP8Context Size:4kTool Calling:SupportedPublished:Mar 25, 2025License:apache-2.0Architecture:Transformer Open Weights Featherless Exclusive Cold

The eganwo/mistral7b-distilled-from-deepseek-r1-qwen32b model is a 7 billion parameter language model fine-tuned from Mistral-7B-Instruct-v0.3. This model was distilled from a larger DeepSeek-R1-Qwen32B model, aiming to retain performance in a smaller footprint. It is designed for general language generation tasks, leveraging the Mistral architecture for efficient inference.

Loading preview...

Model Overview

This model, eganwo/mistral7b-distilled-from-deepseek-r1-qwen32b, is a 7 billion parameter language model. It is a fine-tuned version of the mistralai/Mistral-7B-Instruct-v0.3 base model. A key characteristic of this model is its distillation process, where it was trained to mimic the behavior of a larger DeepSeek-R1-Qwen32B model.

Training Details

The model was fine-tuned using the following hyperparameters:

  • Learning Rate: 5e-06
  • Batch Size: 1 (train), 8 (eval)
  • Gradient Accumulation Steps: 4
  • Optimizer: AdamW with default betas and epsilon
  • LR Scheduler: Linear
  • Epochs: 1

Framework Versions

The training utilized:

  • Transformers 4.50.0
  • Pytorch 2.6.0
  • Datasets 3.4.1
  • Tokenizers 0.21.1

Intended Use

While specific intended uses and limitations are not detailed in the provided information, as a distilled Mistral-7B variant, it is generally suitable for a range of natural language processing tasks where a balance between performance and computational efficiency is desired. Its distillation from a larger model suggests an aim to achieve strong performance relative to its size.