shb/ legal-embed:latest

4 Downloads Updated 4 days ago

embedding

ollama pull shb/legal-embed

curl http://localhost:11434/api/embed \
  -d '{
    "model": "shb/legal-embed",
    "input": "Why is the sky blue?"
  }'

import ollama

response = ollama.embed(
    model='shb/legal-embed',
    input='The sky is blue because of Rayleigh scattering',
)
print(response.embeddings)

import ollama from 'ollama'

const response = await ollama.embed({
  model: 'shb/legal-embed',
  input: 'The sky is blue because of Rayleigh scattering',
})
console.log(response.embeddings)

Details

Updated 4 days ago

4 days ago

b66e1c4b96f1 · 274MB ·

model

archnomic-bert

·

parameters137M

·

quantizationF16

274MB

Readme

legal-embed

Fine-tuned nomic-embed-text-v1.5 for retrieval over Indian criminal law — BNS 2023, IPC 1860, BNSS 2023, BSA 2023. Built for RAG: maps a legal question to the correct statutory section and separates it from look-alike sections.

Dataset

11,587 contrastive triplets synthesised from the bare Acts: - anchor — a query (search_query: ...) - positive — the correct section (search_document: ...) - negative — a hard negative: a confusingly similar section from the same chapter (e.g. Theft vs Robbery vs Extortion)

Training

Hyperparameter	Value
Loss	TripletLoss, cosine, margin 0.5
Epochs	5
Batch size	16
Max seq length	512
Learning rate	2e-5 (AdamW)
Warmup ratio	0.1
Precision	fp32
Eval split	5% (580 triplets)

Result

Triplet accuracy: base 0.9552 → fine-tuned 0.9983

Usage

Always prefix inputs: search_query: for questions, search_document: for indexed sections. `bash ollama pull shb/legal-embed curl http://localhost:11434/api/embed -d '{"model":"shb/legal-embed","input":"search_query: What constitutes murder under the new criminal code?"}' \`

# legal-embed

Fine-tuned `nomic-embed-text-v1.5` for **retrieval over Indian criminal law** —
BNS 2023, IPC 1860, BNSS 2023, BSA 2023. Built for RAG: maps a legal question to the
correct statutory section and separates it from look-alike sections.

## Dataset
11,587 contrastive triplets synthesised from the bare Acts:
- **anchor** — a query (`search_query: ...`)
- **positive** — the correct section (`search_document: ...`)
- **negative** — a hard negative: a confusingly similar section from the *same chapter*
  (e.g. Theft vs Robbery vs Extortion)

## Training
| Hyperparameter | Value |
|---|---|
| Loss | TripletLoss, cosine, margin 0.5 |
| Epochs | 5 |
| Batch size | 16 |
| Max seq length | 512 |
| Learning rate | 2e-5 (AdamW) |
| Warmup ratio | 0.1 |
| Precision | fp32 |
| Eval split | 5% (580 triplets) |

## Result
Triplet accuracy: base **0.9552** → fine-tuned **0.9983**

## Usage
Always prefix inputs: `search_query:` for questions, `search_document:` for indexed sections.
\```bash
ollama pull shb/legal-embed
curl http://localhost:11434/api/embed -d '{"model":"shb/legal-embed","input":"search_query: What constitutes murder under the new criminal code?"}'
\```

Paste, drop or click to upload images (.png, .jpeg, .jpg, .svg, .gif)