imnbharath/milpac-sarvam-full
translation model by imnbharath
Older snapshot: check the original source for current information.
From the publisher
Model card & documentation
Source previewRead the publisher’s intended use, setup instructions, evaluations and limitations. The original model card is the source of truth.
Model: Sarvam-Translate, full fine-tuned on MILPaC (legal domain) Architecture: Causal LM (Gemma3-4B-IT based), 4.3B parameters Checkpoint size: 8.6 GB (bf16) Prepared by: Bharath Chandra Nimmala, NIT Warangal Contact: bharathjrf0326@nitw.ac.in This package contains a fine-tuned translation model plus a standalone inference script that can be used for integration. Use translatebatch() from infersarvam.py as the primary integration point. Translates English legal text into 9 Indic languages. Fine-tuned specifically on legal-domain text (MILPaC dataset — Acts, CCIFAQ, IP categories), so it performs best on legal/formal register text, not general conversational text. Supported languages (pass the exact English name shown): The checkpoint (8.6 GB) is public on Hugging Face: Repo: https://huggingface.co/imnbharath/milpac-sarvam-full No access token needed. Load it directly by repo ID: frompretrained() will download and cache it automatically on first use. Alternatively, clone it locally: and point --modelpath at the local folder. The torch build above is validated on CUDA 12.4. If your environment…
Read the full model card ↗ · Preview checked 2026-10-29T13:34:34.425Z
Inside the original model card — Document outline
- Sarvam-Translate Full FT — MILPaC Legal Translation Model
- 1. What this model does
- 2. Files in this package
- 3. Getting the model
- 4. Setup
- 5. Usage
- As a Python module (recommended for integration)
- From the command line
- 6. Hardware requirements
- 7. Reproducibility note
- 8. Sample input/output pairs (all 9 languages)
- 9. Known limitations
Headings are captured from the source. Links open the publisher’s document, not a locally hosted copy.
Documentation belongs to its respective authors. Reported project/model license: gpl-3.0. A listing is not a grant of reuse or training rights. Confirm the document’s own terms at the source.
Model overview
imnbharath/milpac-sarvam-full is a translation model repository published by imnbharath on Hugging Face. A library was not reported.
This page summarizes Hub metadata. For intended use, training data, evaluation results and limitations, consult the original model card.
Model facts
- Task
- translation
- Library
- Not reported
- Recent downloads (30 days)
- 0
- Cumulative likes
- 0
- Hugging Face trending score
- 0
- Reported safetensors parameters
- 4,300,079,472
- Architecture
- Gemma3ForConditionalGeneration
- License
- gpl-3.0
- Access
- Not gated by Hugging Face
- Created
- 2026-09-15T06:16:55.000Z
- Last modified
- 2026-09-15T06:33:07.000Z
Compatibility and lineage
Reported languages: bn, gu, hi, ml, mr, or, pa, ta, te, en
Reported base models: sarvamai/sarvam-translate
Parameter count is not a RAM/VRAM requirement. Check precision, quantization, context length and runtime compatibility in the model card; no hardware or API-cost claim is inferred here.
Use and evaluate
Check license terms and access requirements first. Review the model card’s documented loading instructions, evaluations and safety limitations. Benchmark scores are not imported by this directory, and popularity is not an accuracy ranking.
No model weights are downloaded or executed by AltAPIs. Never enable remote model code without reviewing it.
Source and freshness
Source: Hugging Face Hub. Metadata observed 2026-09-15T06:34:03.805Z. Daily imports are snapshots, not real-time monitoring.
Popularity and source listings do not establish security, suitability, licensing rights or benchmark performance.