MOBIRIZER-TECHNOLOGY/AnuLM

Space: dropdown to switch between the four checkpoints 4 days ago

help wanted good first issue

0

AnuLM: a small, trainable Sarvam-shaped MoE language model for Hindi, English and Python, trained from scratch on one GPU. Code, recipes, tutorial, and four checkpoints.

Python

#code-generation#hindi#indic-languages#machine-translation#mixture-of-experts#pytorch#small-language-model

Diagrams for docs/ARCHITECTURE.md 4 days ago

help wanted good first issue

MOBIRIZER-TECHNOLOGY/AnuLM

0

AnuLM: a small, trainable Sarvam-shaped MoE language model for Hindi, English and Python, trained from scratch on one GPU. Code, recipes, tutorial, and four checkpoints.

Python

#code-generation#hindi#indic-languages#machine-translation#mixture-of-experts#pytorch#small-language-model

Hindi translation of docs/TUTORIAL.md 4 days ago

help wanted good first issue

MOBIRIZER-TECHNOLOGY/AnuLM

0

AnuLM: a small, trainable Sarvam-shaped MoE language model for Hindi, English and Python, trained from scratch on one GPU. Code, recipes, tutorial, and four checkpoints.

Python

#code-generation#hindi#indic-languages#machine-translation#mixture-of-experts#pytorch#small-language-model

GGUF conversion / llama.cpp support 4 days ago

help wanted

MOBIRIZER-TECHNOLOGY/AnuLM

0

AnuLM: a small, trainable Sarvam-shaped MoE language model for Hindi, English and Python, trained from scratch on one GPU. Code, recipes, tutorial, and four checkpoints.

Python

#code-generation#hindi#indic-languages#machine-translation#mixture-of-experts#pytorch#small-language-model

Emit a tokenizers-library tokenizer.json 4 days ago

help wanted good first issue

MOBIRIZER-TECHNOLOGY/AnuLM

0

AnuLM: a small, trainable Sarvam-shaped MoE language model for Hindi, English and Python, trained from scratch on one GPU. Code, recipes, tutorial, and four checkpoints.

Python

#code-generation#hindi#indic-languages#machine-translation#mixture-of-experts#pytorch#small-language-model

transformers-loadable export (trust_remote_code) 4 days ago

help wanted good first issue

MOBIRIZER-TECHNOLOGY/AnuLM

0

AnuLM: a small, trainable Sarvam-shaped MoE language model for Hindi, English and Python, trained from scratch on one GPU. Code, recipes, tutorial, and four checkpoints.

Python

#code-generation#hindi#indic-languages#machine-translation#mixture-of-experts#pytorch#small-language-model

Less noisy validation loss 4 days ago

help wanted

MOBIRIZER-TECHNOLOGY/AnuLM

0

AnuLM: a small, trainable Sarvam-shaped MoE language model for Hindi, English and Python, trained from scratch on one GPU. Code, recipes, tutorial, and four checkpoints.

Python

#code-generation#hindi#indic-languages#machine-translation#mixture-of-experts#pytorch#small-language-model

Harden eval_code.py sandboxing 4 days ago

help wanted good first issue

MOBIRIZER-TECHNOLOGY/AnuLM

0

AnuLM: a small, trainable Sarvam-shaped MoE language model for Hindi, English and Python, trained from scratch on one GPU. Code, recipes, tutorial, and four checkpoints.

Python

#code-generation#hindi#indic-languages#machine-translation#mixture-of-experts#pytorch#small-language-model

Add python-edu to the coder mix and re-run the 250k probe 4 days ago

help wanted needs GPU

MOBIRIZER-TECHNOLOGY/AnuLM

0

AnuLM: a small, trainable Sarvam-shaped MoE language model for Hindi, English and Python, trained from scratch on one GPU. Code, recipes, tutorial, and four checkpoints.

Python

#code-generation#hindi#indic-languages#machine-translation#mixture-of-experts#pytorch#small-language-model

Long-context continued pretraining with YaRN at 2,048 tokens 4 days ago

help wanted needs GPU

MOBIRIZER-TECHNOLOGY/AnuLM

0

AnuLM: a small, trainable Sarvam-shaped MoE language model for Hindi, English and Python, trained from scratch on one GPU. Code, recipes, tutorial, and four checkpoints.

Python

#code-generation#hindi#indic-languages#machine-translation#mixture-of-experts#pytorch#small-language-model

Chat fine-tune on the three-language base 4 days ago

help wanted needs GPU

MOBIRIZER-TECHNOLOGY/AnuLM

0

AnuLM: a small, trainable Sarvam-shaped MoE language model for Hindi, English and Python, trained from scratch on one GPU. Code, recipes, tutorial, and four checkpoints.

Python

#code-generation#hindi#indic-languages#machine-translation#mixture-of-experts#pytorch#small-language-model

Linux end-to-end reproduction report 4 days ago

help wanted good first issue

MOBIRIZER-TECHNOLOGY/AnuLM

0

AnuLM: a small, trainable Sarvam-shaped MoE language model for Hindi, English and Python, trained from scratch on one GPU. Code, recipes, tutorial, and four checkpoints.

Python

#code-generation#hindi#indic-languages#machine-translation#mixture-of-experts#pytorch#small-language-model

CI: run test_model.py and a 200-step smoke run on Ubuntu and Windows 4 days ago

help wanted good first issue

MOBIRIZER-TECHNOLOGY/AnuLM

0

AnuLM: a small, trainable Sarvam-shaped MoE language model for Hindi, English and Python, trained from scratch on one GPU. Code, recipes, tutorial, and four checkpoints.

Python

#code-generation#hindi#indic-languages#machine-translation#mixture-of-experts#pytorch#small-language-model

Zenodo DOI for v0.1.x 4 days ago

good first issue release

MOBIRIZER-TECHNOLOGY/AnuLM

0

AnuLM: a small, trainable Sarvam-shaped MoE language model for Hindi, English and Python, trained from scratch on one GPU. Code, recipes, tutorial, and four checkpoints.

Python

#code-generation#hindi#indic-languages#machine-translation#mixture-of-experts#pytorch#small-language-model

Open Issues Need Help