deepseeksft202507281606

This is a LoRA fine-tuned model based on deepseek-ai/deepseek-vl-7b-chat.

Model Details

• Base Model: deepseek-ai/deepseek-vl-7b-chat
• Fine-tuning Method: LoRA (Low-Rank Adaptation)
• Training Data: Custom vision-language dataset
• Model Size: large
• Parameters: ~37,478,400 trainable parameters

Usage

from peft import PeftModel
from deepseek_vl.models import VLChatProcessor, MultiModalityCausalLM

# Load base model
base_model = MultiModalityCausalLM.from_pretrained("deepseek-ai/deepseek-vl-7b-chat")

# Load LoRA adapter
model = PeftModel.from_pretrained(base_model, "takashiyamamoto/deepseeksft202507281606")

# Load processor
processor = VLChatProcessor.from_pretrained("deepseek-ai/deepseek-vl-7b-chat")

Training Configuration

• LoRA Rank (r): 16
• LoRA Alpha: 32
• Learning Rate: 0.0002
• Epochs: 1