Downloads · 30 days
9
17% of all-time downloads
viper-llama31-8b
mx003/viper-llama31-8b
viper-llama31-8b is a machine learning model from mx003. Use it for the machine learning task on the model card, and read the license before you ship it in a product. The card lists the license as llama3.1.
Fine-tune of unsloth/Meta-Llama-3.1-8B-Instruct that converts raw MITRE CVE records into the VIPER (Vulnerability Intelligence Parsed Extraction Record) JSON schema. The schema is not given in the prompt — the model l…
All-time downloads
54
Public
Parameters
8B
16.1 GB on disk
Likes
1
Public
Repo makeup
Click a slice to open those files.
.safetensors16.1 GB · 100%
At a glance
- License
- llama3.1
- Model type
- llama
- Access
- Public
- Created
- Jun 27, 2026
- Updated
- Jun 27, 2026
- SHA
- 816deb38
Base models
Datasets
Model card
From the Hugging Face model README
VIPER — CVE → VIPER Schema Extractor (Llama 3.1 8B, LoRA)
Fine-tune of unsloth/Meta-Llama-3.1-8B-Instruct that converts raw MITRE CVE records into the VIPER (Vulnerability Intelligence Parsed Extraction Record) JSON schema. The schema is not given in the prompt — the model learned it from the examples.
Training summary
| Base model | unsloth/Meta-Llama-3.1-8B-Instruct |
| Method | LoRA (bf16),r=32, α=64, dropout=0.05 |
| Trainable params | 83,886,080 / 8,114,147,328 (1.03%) |
| Dataset | mx003/viper — 989 records |
| Epochs | 2 |
| Sequence length | 4096 |
| Learning rate | 2e-4 (cosine) |
| Hardware | RTX 5090 (32 GB) |
Full hyperparameters are in config.yaml; machine-readable
metrics in training_metadata.json.
- Type
- llama
- License
- llama3.1
- Created
- Jun 27, 2026
- Updated
- Jun 27, 2026