Instructions to use kiddobellamy/Llama_Vision with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use kiddobellamy/Llama_Vision with Transformers:
# pip install -U transformers accelerate # Load model directly from transformers import AutoTokenizer, AutoModelForCausalLM tokenizer = AutoTokenizer.from_pretrained("kiddobellamy/Llama_Vision") model = AutoModelForCausalLM.from_pretrained("kiddobellamy/Llama_Vision", device_map="auto") - Notebooks
- Google Colab
- Kaggle
Download params.json from kiddobellamy/Llama_Vision: direct link, hf CLI and curl.
- Browser
- Download file 322 Bytes
-
https://huggingface.co/kiddobellamy/Llama_Vision/resolve/main/params.json
- Command line
-
hf download hf://kiddobellamy/Llama_Vision/params.json
-
curl -L -o params.json https://huggingface.co/kiddobellamy/Llama_Vision/resolve/main/params.json
322 Bytes
| { | |
| "dim": 8192, | |
| "ffn_dim_multiplier": 1.3, | |
| "multiple_of": 4096, | |
| "n_heads": 64, | |
| "n_kv_heads": 8, | |
| "n_layers": 80, | |
| "norm_eps": 1e-05, | |
| "rope_theta": 500000.0, | |
| "use_scaled_rope": true, | |
| "vision_chunk_size": 560, | |
| "vision_max_num_chunks": 4, | |
| "vocab_size": 128256, | |
| "vision_num_cross_attention_layers": 20 | |
| } |