Text Generation
Shadow
English
small-language-model
cpu
retrieval
memory-on-disk
ternary
offline
edge
Instructions to use QLNI/shadow-50m-instruct with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Shadow
How to use QLNI/shadow-50m-instruct with Shadow:
# No code snippets available yet for this library. # To use this model, check the repository files and the library's documentation. # Want to help? PRs adding snippets are welcome at: # https://github.com/huggingface/huggingface.js
- Notebooks
- Google Colab
- Kaggle
| { | |
| "model_type": "shadow", | |
| "architectures": ["SHADOW50MInstruct"], | |
| "model_name": "SHADOW 50M Instruct", | |
| "version": "1.1", | |
| "num_parameters": 44100000, | |
| "hidden_size": 896, | |
| "num_hidden_layers": 6, | |
| "num_attention_heads": 14, | |
| "num_key_value_heads": 2, | |
| "head_dim": 64, | |
| "intermediate_size": 2464, | |
| "hidden_act": "silu", | |
| "vocab_size": 73880, | |
| "vocab_code_bits": 512, | |
| "vocab_trainable": false, | |
| "weight_format": "ternary", | |
| "bits_per_weight": 1.58, | |
| "max_position_embeddings": 8192, | |
| "archive_max_tokens": 100000000, | |
| "archive_bytes_per_token": 288, | |
| "deployment_file": "shadow50_instruct.shdw", | |
| "deployment_bytes": 19799680, | |
| "runtime": "shadow50 kernel, C, AVX2 FMA F16C, CPU", | |
| "license": "mit", | |
| "homepage": "https://github.com/QLNI/SHADOW-50M-Instruct" | |
| } | |