Downloads · 30 days
0
bankless-llama3-8b-soft-prompt-v1
FloKrause25/bankless-llama3-8b-soft-prompt-v1
bankless-llama3-8b-soft-prompt-v1 is a machine learning model from FloKrause25. Use it for the machine learning task on the model card, and read the license before you ship it in a product.
This model combines: 1. Base: Meta-Llama-3-8B-Instruct (4-bit quantized) 2. LoRA adapter: Fine-tuned on Bankless content 3. Soft prompts: 20 virtual tokens for enhanced Bankless style
Access
Public
Updated Jun 26, 2025
Repo size
17.5 MB
Likes
0
Public
Repo makeup
Click a slice to open those files.
.json17.3 MB · 98%
At a glance
- Access
- Public
- Created
- Jun 26, 2025
- Updated
- Jun 26, 2025
- SHA
- e1fdfd12
Model card
From the Hugging Face model README
Bankless Llama-3 8B with LoRA + Soft Prompts
This model combines:
- Base: Meta-Llama-3-8B-Instruct (4-bit quantized)
- LoRA adapter: Fine-tuned on Bankless content
- Soft prompts: 20 virtual tokens for enhanced Bankless style
Training Details
- LoRA checkpoint: /content/drive/MyDrive/bankless_training/session_20250626_064641/checkpoint-695
- Soft prompt training: 80 train / 20 test examples
- Final eval loss: 1.2585
- Training date: 2025-06-26
Usage
This model generates Bankless-style newsletter content focusing on crypto and AI topics.
- Created
- Jun 26, 2025
- Updated
- Jun 26, 2025