| { |
| "Llama-3.2-1B-Instruct": { |
| "name": "Llama 3.2 1B", |
| "size": 1, |
| "size_p": 1.24e9, |
| "attention": "GQA (4x)", |
| "hidden_size": 2048, |
| "num_hidden_layers": 16, |
| "num_attention_heads": 32, |
| "num_key_value_heads": 8 |
| }, |
| "Llama-3.2-3B-Instruct": { |
| "name": "Llama 3.2 3B", |
| "size": 3, |
| "size_p": 3.21e9, |
| "attention": "GQA (3x)", |
| "hidden_size": 3072, |
| "num_hidden_layers": 28, |
| "num_attention_heads": 24, |
| "num_key_value_heads": 8 |
| }, |
| "Llama-3.1-8B-Instruct": { |
| "name": "Llama 3.1 8B", |
| "size": 8, |
| "size_p": 8.03e9, |
| "attention": "GQA (4x)", |
| "hidden_size": 4096, |
| "num_hidden_layers": 32, |
| "num_attention_heads": 32, |
| "num_key_value_heads": 8 |
| }, |
|
|
| "gemma-2-2b-it": { |
| "name": "Gemma 2 2B", |
| "size": 2, |
| "size_p": 2.61e9, |
| "attention": "GQA (2x)", |
| "hidden_size": 2304, |
| "num_hidden_layers": 26, |
| "num_attention_heads": 8, |
| "num_key_value_heads": 4 |
| }, |
| "gemma-2-9b-it": { |
| "name": "Gemma 2 9B", |
| "size": 9, |
| "size_p": 9.24e9, |
| "attention": "GQA (2x)", |
| "hidden_size": 3584, |
| "num_hidden_layers": 42, |
| "num_attention_heads": 16, |
| "num_key_value_heads": 8 |
| }, |
| "gemma-2-27b-it": { |
| "name": "Gemma 2 27B", |
| "size": 27, |
| "size_p": 27.24e9, |
| "attention": "GQA (2x)", |
| "hidden_size": 4608, |
| "num_hidden_layers": 46, |
| "num_attention_heads": 32, |
| "num_key_value_heads": 16 |
| }, |
|
|
| "granite-3b-code-base-2k": { |
| "name": "Granite 3B", |
| "size": 3, |
| "size_p": 3.48e9, |
| "attention": "MHA", |
| "hidden_size": 2560, |
| "num_hidden_layers": 32, |
| "num_attention_heads": 32, |
| "num_key_value_heads": 32 |
| }, |
| "granite-8b-code-base-4k": { |
| "name": "Granite-8B", |
| "size": 8, |
| "size_p": 8.05e9, |
| "attention": "GQA (4x)", |
| "hidden_size": 4096, |
| "num_hidden_layers": 36, |
| "num_attention_heads": 32, |
| "num_key_value_heads": 8 |
| }, |
|
|
| "Qwen2-1.5B": { |
| "name": "Qwen2 1.5B", |
| "size": 1.5, |
| "size_p": 1.54e9, |
| "attention": "GQA (6x)", |
| "hidden_size": 1536, |
| "num_hidden_layers": 28, |
| "num_attention_heads": 12, |
| "num_key_value_heads": 2 |
| }, |
| "Qwen2-7B": { |
| "name": "Qwen2 7B", |
| "size": 7, |
| "size_p": 7.62e9, |
| "attention": "GQA (7x)", |
| "hidden_size": 3584, |
| "num_hidden_layers": 28, |
| "num_attention_heads": 28, |
| "num_key_value_heads": 4 |
| }, |
|
|
| "opt-1.3b": { |
| "name": "OPT 1.3B", |
| "size": 1.3, |
| "size_p": 1.32e9, |
| "attention": "MHA", |
| "hidden_size": 2048, |
| "num_hidden_layers": 24, |
| "num_attention_heads": 32, |
| "num_key_value_heads": 32 |
| }, |
| "opt-2.7b": { |
| "name": "OPT 2.7B", |
| "size": 2.7, |
| "size_p": 2.75e9, |
| "attention": "MHA", |
| "hidden_size": 2560, |
| "num_hidden_layers": 32, |
| "num_attention_heads": 32, |
| "num_key_value_heads": 32 |
| }, |
| "opt-6.7b": { |
| "name": "OPT 6.7B", |
| "size": 6.7, |
| "size_p": 6.71e9, |
| "attention": "MHA", |
| "hidden_size": 4096, |
| "num_hidden_layers": 32, |
| "num_attention_heads": 32, |
| "num_key_value_heads": 32 |
| }, |
|
|
| "falcon-rw-7b": { |
| "name": "Falcon 7B", |
| "size": 7, |
| "size_p": 7.78e9, |
| "attention": "MHA", |
| "hidden_size": 4096, |
| "num_hidden_layers": 36, |
| "num_attention_heads": 64, |
| "num_key_value_heads": 64 |
| }, |
|
|
| "Mistral-Small-24B-Instruct-2501": { |
| "name": "Mistral 24B", |
| "size": 24, |
| "size_p": 23.56e9, |
| "attention": "GQA(4x)", |
| "hidden_size": 5120, |
| "num_hidden_layers": 40, |
| "num_attention_heads": 32, |
| "num_key_value_heads": 8 |
| } |
| } |
