# lballore/llimba-3b-instruct

Model from Hugging Face for text generation: lballore/llimba-3b-instruct · transformers

## Facts

- Type: LLM
- Task: text-generation
- Framework: transformers
- License: apache-2.0 — Permissive open license: commercial use, modification, and redistribution are permitted. (https://huggingface.co/lballore/llimba-3b-instruct/blob/main/LICENSE)
- Source: https://huggingface.co/lballore/llimba-3b-instruct
- Hosting: hosted (9 files)
- Files: 9 hosted of 10 listed
- File counts: "listed" is every file in the source repository; only hosted files download from Hugging Bay, and the download plan lists the files it selects for one download, so its count can differ.
- Hosted here: 9 files, 6.18 GB
- Full repo upstream: 6.18 GB
- Recommended file: 6.17 GB · model.safetensors
- Parameters: 3B
- Download: https://huggingbay.xyz/artifact/hf-model-lballore-llimba-3b-instruct/files (download plan: https://huggingbay.xyz/api/artifacts/hf-model-lballore-llimba-3b-instruct/download-plan)
- Per-file SHA-256: available for 9 of 9 hosted files via https://huggingbay.xyz/api/artifacts/hf-model-lballore-llimba-3b-instruct/files?limit=100
- Run-ready: no (Hugging Bay has not certified a run path for this model; hosted bytes are downloadable and hash-checkable only)
- Peer-assisted fallbacks: 0 verified, 0 pending review
- Downloads: 35
- Stars: 1
- Card completeness: 96%

## Description

External Hugging Face catalog record for lballore/llimba-3b-instruct. Task: text generation. Library: transformers. Languages: srd, en, zh, fr, es, pt, de, it. Datasets referenced by card metadata: lballore/llimba-corpus, lballore/llimba-sft, lballore/llimba-flores-srd-eval, facebook/flores. Base model: lballore/llimba-3b-instruct-cpt. Includes 8 model-card evaluation signals. Popularity signal at import: 30 downloads and 1 likes.

## Model Card Facts

- Declared license: apache-2.0
- Pipeline: text-generation
- Library: transformers
- Base model: lballore/llimba-3b-instruct-cpt
- Task categories: text-generation
- Languages: srd, en, zh, fr, es, pt, de, it
- Referenced datasets: lballore/llimba-corpus, lballore/llimba-sft, lballore/llimba-flores-srd-eval, facebook/flores
- Paper ids: 2605.09015, 2411.13453
- Model-index entries: llimba-3b-instruct
- Evaluation signals: 8 model-card eval results
- File hints: 10 upstream file samples

## Evaluation Signals

- FLORES-200 (Sardinian subset, 997 sentences): EN-SC BLEU = 28.47 (translation)
- FLORES-200 (Sardinian subset, 997 sentences): EN-SC chrF = 56.8 (translation)
- FLORES-200 (Sardinian subset, 997 sentences): IT-SC BLEU = 21.25 (translation)
- FLORES-200 (Sardinian subset, 997 sentences): IT-SC chrF = 52.08 (translation)
- FLORES-200 (Sardinian subset, 997 sentences): ES-SC BLEU = 18.57 (translation)
- FLORES-200 (Sardinian subset, 997 sentences): ES-SC chrF = 49.41 (translation)
- FLORES-200 (Sardinian subset, 997 sentences): SC-EN BLEU = 41.28 (translation)
- FLORES-200 (Sardinian subset, 997 sentences): SC-EN chrF = 64.64 (translation)

## Referenced Data

- Dataset: lballore/llimba-corpus
- Dataset: lballore/llimba-sft
- Dataset: lballore/llimba-flores-srd-eval
- Dataset: facebook/flores
- Paper: 2605.09015
- Paper: 2411.13453

## File Hints

- .gitattributes (1.5 KB)
- LICENSE (12 KB)
- README.md (14 KB)
- chat_template.jinja (2.4 KB)
- config.json (1.5 KB)
- generation_config.json (216 B)
- model.safetensors (5.7 GB) LFS
- tokenizer.json (11 MB) LFS
- tokenizer_config.json (664 B)
- training_config.json (641 B)

## Quality Checklist

- Summary: passed
- Description: passed
- Model card metadata: passed
- License: passed
- Provenance: passed
- File inventory: passed
- Hosted bytes: passed
- Scan result: passed
- Manifest signature: passed

## Safety Notes

- External catalog metadata imported from Hugging Face public API on 2026-09-04T18:59:11.859Z.
- Open license detected as apache-2.0.
- Model-index metadata found for llimba-3b-instruct.
- Model card references dataset context: lballore/llimba-corpus, lballore/llimba-sft, lballore/llimba-flores-srd-eval, facebook/flores.
- Not a run kit: hosted, hash-checked bytes only; run readiness is not certified for this model yet.

## Actions

- Web: https://huggingbay.xyz/artifact/hf-model-lballore-llimba-3b-instruct
- Files API: https://huggingbay.xyz/api/artifacts/hf-model-lballore-llimba-3b-instruct/files?limit=100
- Metadata: https://huggingbay.xyz/api/artifacts/hf-model-lballore-llimba-3b-instruct/metadata.json
- Distribution: https://huggingbay.xyz/api/artifacts/hf-model-lballore-llimba-3b-instruct/distribution
- Manifest: https://huggingbay.xyz/api/manifests/hf-model-lballore-llimba-3b-instruct
- Download plan (per-file SHA-256 + signed download URLs): https://huggingbay.xyz/api/artifacts/hf-model-lballore-llimba-3b-instruct/download-plan
- Hosted download rows: https://huggingbay.xyz/api/downloadable-files?artifactId=hf-model-lballore-llimba-3b-instruct&limit=100
- Request mirror: not applicable.
- Submit peer fallback for review: POST https://huggingbay.xyz/api/artifacts/hf-model-lballore-llimba-3b-instruct/decentralized-fallbacks