charlottemeyer/qwen3-0.6b-quantized-phase1-full-precision

CS-552 Phase 1 Full-Precision Model

This model is part of the CS-552 course project for quantized language models.

Model Details

• Base Model: Qwen/Qwen3-0.6B-Base
• Training Phase: Phase 1 Full-Precision
• Dataset: simplescaling/s1K-1.1_tokenized
• Training Method: Full-precision SFT

Usage

from transformers import AutoModelForCausalLM, AutoTokenizer

model = AutoModelForCausalLM.from_pretrained("charlottemeyer/qwen3-0.6b-quantized-phase1-full-precision")
tokenizer = AutoTokenizer.from_pretrained("charlottemeyer/qwen3-0.6b-quantized-phase1-full-precision")

# Generate response
inputs = tokenizer("Your question here", return_tensors="pt")
outputs = model.generate(**inputs, max_new_tokens=100)
response = tokenizer.decode(outputs[0], skip_special_tokens=True)

Training Details

• Trained for CS-552 course evaluation
• Optimized for reasoning and MCQA tasks
• Uses answer-first format for logit extraction