codelion (Asankhaya Sharma)

reacted to their post with 👀🤗🚀🔥 13 minutes ago

Post

Inspired by the Nemotron Diffusion recipe, check out dhara-250m: a 250M experimental language model that supports three decoding modes from one set of weights: autoregressive, block-diffusion, and self-speculation.

It is small, easy to try, and meant for exploring diffusion-style decoding and latency tradeoffs in compact LMs.

Model: codelion/dhara-250m

Try the chat demo here: codelion/dhara-chat

posted an update 13 minutes ago

Post

Inspired by the Nemotron Diffusion recipe, check out dhara-250m: a 250M experimental language model that supports three decoding modes from one set of weights: autoregressive, block-diffusion, and self-speculation.

It is small, easy to try, and meant for exploring diffusion-style decoding and latency tradeoffs in compact LMs.

Model: codelion/dhara-250m

Try the chat demo here: codelion/dhara-chat

replied to their post 3 months ago

Thank you 🫡 !

reacted to their post with 🤗👀🚀🔥 3 months ago

Post

3400

Scaling Pedagogical Pre-training to 10 Billion Tokens

New blog post exploring what happens when you take optimal data mixing insights and scale up the data generation itself.

We built Sutra, a multi-stage framework for generating pedagogical pre-training data guided by a knowledge graph of ~2,000 concepts across 9 domains. The pipeline includes structured content generation, six-dimension quality evaluation, diversity management across 20 content styles, and a cleaning stage to prevent collapse.

The result is codelion/sutra-10B, a 10.2 billion token pedagogical dataset with rich metadata (domain, complexity, prerequisites, quality scores) on every entry.

We trained codelion/SmolLM2-70M on it for 3 full epochs (30.6B tokens) on a single A10 GPU in ~78 hours.

Key finding: perplexity kept improving across epochs, but benchmark gains plateaued fast. At 70M parameters, the model hits a representational ceiling that more data alone can't break through.

Full writeup with comparisons against 7 other datasets, detailed benchmark breakdowns, and connections to recent work on synthetic data scaling, curriculum learning, and data mixing laws: https://huggingface.co/blog/codelion/scaling-pedagogical-pretraining-10-billion-tokens

All datasets at multiple scales (10M, 100M, 1B, 10B) plus seed concepts and an SFT variant are in the Sutra Pedagogical Datasets collection.

2 replies

·

posted an update 3 months ago

Post

3400

Scaling Pedagogical Pre-training to 10 Billion Tokens

New blog post exploring what happens when you take optimal data mixing insights and scale up the data generation itself.

We built Sutra, a multi-stage framework for generating pedagogical pre-training data guided by a knowledge graph of ~2,000 concepts across 9 domains. The pipeline includes structured content generation, six-dimension quality evaluation, diversity management across 20 content styles, and a cleaning stage to prevent collapse.

The result is codelion/sutra-10B, a 10.2 billion token pedagogical dataset with rich metadata (domain, complexity, prerequisites, quality scores) on every entry.

We trained codelion/SmolLM2-70M on it for 3 full epochs (30.6B tokens) on a single A10 GPU in ~78 hours.

Key finding: perplexity kept improving across epochs, but benchmark gains plateaued fast. At 70M parameters, the model hits a representational ceiling that more data alone can't break through.

Full writeup with comparisons against 7 other datasets, detailed benchmark breakdowns, and connections to recent work on synthetic data scaling, curriculum learning, and data mixing laws: https://huggingface.co/blog/codelion/scaling-pedagogical-pretraining-10-billion-tokens

All datasets at multiple scales (10M, 100M, 1B, 10B) plus seed concepts and an SFT variant are in the Sutra Pedagogical Datasets collection.

2 replies

·

reacted to their post with ❤️👍🚀🔥 4 months ago

Post

3268

Reverse Engineering a $500M Mystery: From HashHop to Memory-Augmented Language Models

I wrote a deep dive into how Magic AI's 100M token context window might work, starting from their HashHop benchmark and building up to MALM - a Memory-Augmented Language Model.

Key insight: treating each key as a single token enables perfect retrieval at unlimited context lengths.

The article covers:

- How HashHop works and why its perfect accuracy is suspicious
- Building a tokenized solver that achieves 100% accuracy
- Scaling to MALM for real code search tasks
- Why this approach could handle 100M+ tokens

Read the full article: https://huggingface.co/blog/codelion/reverse-engineering-magic-hashhop

Try the model: codelion/malm-165m

Code: https://github.com/codelion/hash-hop

1 reply

·

posted an update 4 months ago

Post

3268

Reverse Engineering a $500M Mystery: From HashHop to Memory-Augmented Language Models

I wrote a deep dive into how Magic AI's 100M token context window might work, starting from their HashHop benchmark and building up to MALM - a Memory-Augmented Language Model.

Key insight: treating each key as a single token enables perfect retrieval at unlimited context lengths.

The article covers:

- How HashHop works and why its perfect accuracy is suspicious
- Building a tokenized solver that achieves 100% accuracy
- Scaling to MALM for real code search tasks
- Why this approach could handle 100M+ tokens

Read the full article: https://huggingface.co/blog/codelion/reverse-engineering-magic-hashhop

Try the model: codelion/malm-165m

Code: https://github.com/codelion/hash-hop

1 reply

·

reacted to their post with 👍🤗🚀🔥 5 months ago

Post

6179

Introducing Dhara-70M: A diffusion language model that achieves 3.8x higher throughput than autoregressive models!

Key findings from our research on optimal architectures for small language models:

→ Depth beats width: 32 layers outperforms 12 layers at the same parameter count
→ Best-in-class factuality: 47.5% on TruthfulQA
→ 10x training efficiency using WSD (Warmup-Stable-Decay) conversion
→ Canon layers add only 0.13% parameters but improve reasoning

We trained on 1B tokens using the optimal 50-30-20 dataset mix (PDFs + filtered web + educational content), then converted to diffusion with just 100M additional tokens.

Blog: https://huggingface.co/blog/codelion/optimal-model-architecture
Model: codelion/dhara-70m

1 reply

·

Asankhaya Sharma

AI & ML interests

Recent Activity

Organizations

Asankhaya Sharma

AI & ML interests

Recent Activity

Organizations

codelion's activity