hanluzhi0429@gmail.com
18 lines
585 B
Python
18 lines
585 B
Python
import transformers
|
|
from transformers import AutoModelForCausalLM, AutoTokenizer
|
|
from transformers import LlamaForCausalLM
|
|
tokenizer = AutoTokenizer.from_pretrained("Qwen/Qwen2-0.5B-Instruct")
|
|
from transformers import LlamaConfig
|
|
config = LlamaConfig(
|
|
vocab_size=32000,
|
|
hidden_size=4096,
|
|
num_attention_heads=32,
|
|
num_hidden_layers=32,
|
|
intermediate_size=11008,
|
|
)
|
|
|
|
model = LlamaForCausalLM(config)
|
|
|
|
model = transformers.Qwen2ForCausalLM(config)
|
|
model_size = sum(t.numel() for t in model.parameters())
|
|
print(f"Model Size: {model_size/1000**2:.1f}M parameters") |