I highly recommend you to use our flagship 3B model instead: https://huggingface.co/Pinkstack/PARM-V2-QwQ-Qwen-2.5-o1-3B-GGUF <- this model has generally better performance and is higher quality.

Uploaded model

Further trained Llama 3.2 3B with enhanced reasoning, aimed to mimic deepseek r1 quality.

Trained with Nvidia Tesla T4

This is the first model we've trained it may have abnormalities, Phi 3.5 soon.

  • Developed by: PinkStack
  • License: apache-2.0
  • Finetuned from model : unsloth/llama-3.2-3b-instruct-bnb-4bit

This llama model was trained using Unsloth and Huggingface's TRL library.

Downloads last month
31
GGUF
Model size
3.21B params
Architecture
llama

8-bit

Inference Providers NEW
This model is not currently available via any of the supported third-party Inference Providers, and the model is not deployed on the HF Inference API.

Dataset used to train Pinkstack/Llama-3.2-3B-o1