Beta
Explore
Marketplace
Neural Labs
Chat
Wallet
Docs
llama2-7b-finetuned-finance – AI Model by egupta | AlphaNeural AI
You can deploy this model and start earning money today!
egupta
/
llama2-7b-finetuned-finance
like
0
peft
safetensors
meta-llama/Llama-2-7b-hf
adapter
us
Views
No views yet
Model card
Files and Versions
Community
API
Deploy
This model is fine-tuned using question-answer (QA) pairs scraped from Reddit finance communities.
Reddit QA Collection
Subreddits scraped:
"wallstreetbets"
"trading"
"daytrading"
"cryptocurrency"
"stocks"
"ETFs"
"bogleheads"
"investing"
"personalfinance"
"stockmarket"
"options"
"securityanalysis"
"dividendinvesting"
Fine-Tuning with LoRA + 4-bit Quantization
Used PEFT with LoRA and BitsAndBytes 4-bit quantization.
Training Details
4-bit quantization using
nf4
LoRA:
r=8
,
lora_alpha=16
,
dropout=0.05
Batch: 4, Accumulation: 4, Epochs: 2
