808AI150m is an open-weight model from Sadatsami. Its published files total 13.9 GB.
A modern 150M-parameter decoder-only language model. Engineered by Sadat Sami · Released under the Apache License 2.0 808 AI is an open-source language model designed and engineered by Sadat Sami as part of his research into efficient, modern language model…
Model Card
A modern 150M-parameter decoder-only language model. Engineered by Sadat Sami · Released under the Apache License 2.0 808 AI is an open-source language model designed and engineered by Sadat Sami as part of his research into efficient, modern language model architectures. It is a compact decoder-only transformer built for fine-tuning on conversational and instruction-following tasks, and can be trained end-to-end on a single workstation GPU. - Grouped Query Attention (GQA) — 12 query heads, 4 key/value heads. - Query-Key Normalization (QK-Norm) — RMSNorm on q/k heads before RoPE. - Layer-wise Sliding Window Attention — alternating local/global pattern. - Rotary Position Embeddings (RoPE)…
Excerpt from the card by Sadatsami.
Identity and Version
- Repository
- Sadatsami/808AI150m
- Publisher
- Sadatsami
- Task
- Not stated by the source
- Modality
- Other
- Library
- Not stated by the source
- Parameters
- Not stated by the source
- Languages
- Not stated by the source
- Revision
- 9a0c82b8c1709904556ee0212b18c03345cbe21c
- First published
- 2026-09-20
- Last updated
- 2026-10-04
Files and Weights
99 files, 13.9 GB in total. The weights are 23 files totalling 13.8 GB in safetensors.
Every file
| File | Type | Size | SHA-256 |
|---|---|---|---|
| checkpoint-1000/model.safetensors | Weights | 601.8 MB | 0235a0d62d95 |
| checkpoint-10000/model.safetensors | Weights | 601.8 MB | b533cd8c57fd |
| checkpoint-10500/model.safetensors | Weights | 601.8 MB | 3fab2532d98b |
| checkpoint-11000/model.safetensors | Weights | 601.8 MB | 33a52f8b4f09 |
| checkpoint-11500/model.safetensors | Weights | 601.8 MB | df75fd33b347 |
| checkpoint-1500/model.safetensors | Weights | 601.8 MB | fa4a3e1119ad |
| checkpoint-2000/model.safetensors | Weights | 601.8 MB | d25480374a34 |
| checkpoint-2500/model.safetensors | Weights | 601.8 MB | f2fb27043a93 |
| checkpoint-3000/model.safetensors | Weights | 601.8 MB | 39a75fd3f0ff |
| checkpoint-3500/model.safetensors | Weights | 601.8 MB | 9189b5f44100 |
| checkpoint-4000/model.safetensors | Weights | 601.8 MB | 7812f73a0fa5 |
| checkpoint-4500/model.safetensors | Weights | 601.8 MB | 277c7cdb463f |
| checkpoint-500/model.safetensors | Weights | 601.8 MB | c4cc90feebbc |
| checkpoint-5000/model.safetensors | Weights | 601.8 MB | e9e22759c452 |
| checkpoint-5500/model.safetensors | Weights | 601.8 MB | 6fe15bb5302e |
| checkpoint-6000/model.safetensors | Weights | 601.8 MB | fe091938926e |
| checkpoint-6500/model.safetensors | Weights | 601.8 MB | 8e91a7d42e60 |
| checkpoint-7000/model.safetensors | Weights | 601.8 MB | 70072cf250cd |
| checkpoint-7500/model.safetensors | Weights | 601.8 MB | c292dcff92af |
| checkpoint-8000/model.safetensors | Weights | 601.8 MB | ccce03dcdf1f |
| checkpoint-8500/model.safetensors | Weights | 601.8 MB | 6ee69d326822 |
| checkpoint-9000/model.safetensors | Weights | 601.8 MB | f0fd66e1c1ec |
| checkpoint-9500/model.safetensors | Weights | 601.8 MB | 0e3f276d4c16 |
| __init__.py | Configuration | 337 B | — |
| checkpoint-1000/config.json | Configuration | 750 B | — |
| checkpoint-1000/generation_config.json | Configuration | 216 B | — |
| checkpoint-10000/config.json | Configuration | 751 B | — |
| checkpoint-10000/trainer_state.json | Configuration | 343 B | — |
| checkpoint-10500/config.json | Configuration | 751 B | — |
| checkpoint-10500/trainer_state.json | Configuration | 343 B | — |
| checkpoint-11000/config.json | Configuration | 751 B | — |
| checkpoint-11000/trainer_state.json | Configuration | 343 B | — |
| checkpoint-11500/config.json | Configuration | 751 B | — |
| checkpoint-11500/trainer_state.json | Configuration | 343 B | — |
| checkpoint-1500/config.json | Configuration | 750 B | — |
| checkpoint-1500/trainer_state.json | Configuration | 341 B | — |
| checkpoint-2000/config.json | Configuration | 750 B | — |
| checkpoint-2000/trainer_state.json | Configuration | 342 B | — |
| checkpoint-2500/config.json | Configuration | 750 B | — |
| checkpoint-2500/trainer_state.json | Configuration | 342 B | — |
| checkpoint-3000/config.json | Configuration | 750 B | — |
| checkpoint-3000/trainer_state.json | Configuration | 342 B | — |
| checkpoint-3500/config.json | Configuration | 750 B | — |
| checkpoint-3500/trainer_state.json | Configuration | 342 B | — |
| checkpoint-4000/config.json | Configuration | 750 B | — |
| checkpoint-4000/trainer_state.json | Configuration | 342 B | — |
| checkpoint-4500/config.json | Configuration | 750 B | — |
| checkpoint-4500/trainer_state.json | Configuration | 342 B | — |
| checkpoint-500/config.json | Configuration | 750 B | — |
| checkpoint-500/generation_config.json | Configuration | 216 B | — |
| checkpoint-5000/config.json | Configuration | 750 B | — |
| checkpoint-5000/trainer_state.json | Configuration | 342 B | — |
| checkpoint-5500/config.json | Configuration | 750 B | — |
| checkpoint-5500/trainer_state.json | Configuration | 342 B | — |
| checkpoint-6000/config.json | Configuration | 750 B | — |
| checkpoint-6000/trainer_state.json | Configuration | 342 B | — |
| checkpoint-6500/config.json | Configuration | 750 B | — |
| checkpoint-6500/trainer_state.json | Configuration | 342 B | — |
| checkpoint-7000/config.json | Configuration | 750 B | — |
| checkpoint-7000/trainer_state.json | Configuration | 342 B | — |
| checkpoint-7500/config.json | Configuration | 750 B | — |
| checkpoint-7500/trainer_state.json | Configuration | 342 B | — |
| checkpoint-8000/config.json | Configuration | 751 B | — |
| checkpoint-8000/trainer_state.json | Configuration | 342 B | — |
| checkpoint-8500/config.json | Configuration | 751 B | — |
| checkpoint-8500/trainer_state.json | Configuration | 342 B | — |
| checkpoint-9000/config.json | Configuration | 751 B | — |
| checkpoint-9000/trainer_state.json | Configuration | 342 B | — |
| checkpoint-9500/config.json | Configuration | 751 B | — |
| checkpoint-9500/trainer_state.json | Configuration | 342 B | — |
| config_808ai.py | Configuration | 2.4 KB | — |
| modeling_808ai.py | Configuration | 21.7 KB | — |
| train_808ai_resume.py | Configuration | 22.3 KB | — |
| LICENSE | Documentation | 601 B | — |
| README.md | Documentation | 1.5 KB | — |
| .gitattributes | Repository | 1.6 KB | — |
| checkpoint-10000/tokenizer.json | Tokenizer | 2.3 MB | — |
| checkpoint-10500/tokenizer.json | Tokenizer | 2.3 MB | — |
| checkpoint-11000/tokenizer.json | Tokenizer | 2.3 MB | — |
| checkpoint-11500/tokenizer.json | Tokenizer | 2.3 MB | — |
| checkpoint-1500/tokenizer.json | Tokenizer | 2.3 MB | — |
| checkpoint-2000/tokenizer.json | Tokenizer | 2.3 MB | — |
| checkpoint-2500/tokenizer.json | Tokenizer | 2.3 MB | — |
| checkpoint-3000/tokenizer.json | Tokenizer | 2.3 MB | — |
| checkpoint-3500/tokenizer.json | Tokenizer | 2.3 MB | — |
| checkpoint-4000/tokenizer.json | Tokenizer | 2.3 MB | — |
| checkpoint-4500/tokenizer.json | Tokenizer | 2.3 MB | — |
| checkpoint-5000/tokenizer.json | Tokenizer | 2.3 MB | — |
| checkpoint-5500/tokenizer.json | Tokenizer | 2.3 MB | — |
| checkpoint-6000/tokenizer.json | Tokenizer | 2.3 MB | — |
| checkpoint-6500/tokenizer.json | Tokenizer | 2.3 MB | — |
| checkpoint-7000/tokenizer.json | Tokenizer | 2.3 MB | — |
| checkpoint-7500/tokenizer.json | Tokenizer | 2.3 MB | — |
| checkpoint-8000/tokenizer.json | Tokenizer | 2.3 MB | — |
| checkpoint-8500/tokenizer.json | Tokenizer | 2.3 MB | — |
| checkpoint-9000/tokenizer.json | Tokenizer | 2.3 MB | — |
| checkpoint-9500/tokenizer.json | Tokenizer | 2.3 MB | — |
| tokenizer.json | Tokenizer | 2.3 MB | — |
| tokenizer_config.json | Tokenizer | 257 B | — |
License and Download
- License
- Not stated by the source
- Access
- Open weights, no gate
- Download size
- 13.8 GB
Released by Sadatsami through its official repository on Hugging Face.
Memory Requirements
| Precision | Weights in memory |
|---|---|
| As published | 13.8 GB |
Weights only, from the published parameter count; the key-value cache and runtime add to this.