Coperniq Sales SFT v3
Fine-tuned Qwen2.5-3B-Instruct for MEP contractor sales emails.
Training
- Base Model: unsloth/Qwen2.5-3B-Instruct (3B params)
- Method: LoRA (r=16, alpha=16)
- Examples: 8,558 (Close CRM + Gong + synthetic)
- Steps: 2,000
- Final Loss: 0.43
Data Sources
- 7,879 Close CRM emails
- 575 Gong call transcripts
- 104 synthetic examples
Usage
from unsloth import FastLanguageModel
model, tokenizer = FastLanguageModel.from_pretrained(
"scientiacapital/coperniq-sales-sft",
max_seq_length=2048,
load_in_4bit=True
)
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support