File size: 3,079 Bytes
fb6c5ee |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 |
---
library_name: transformers
pipeline_tag: text-generation
inference: true
widget:
- text: Hello!
example_title: Hello world
group: Python
---
This tiny model is for debugging. It is randomly initialized with the config adapted from [Qwen/Qwen3-235B-A22B](https://huggingface.co/Qwen/Qwen3-235B-A22B).
### Example usage:
```python
from transformers import pipeline
model_id = "yujiepan/qwen3-moe-tiny-random"
pipe = pipeline(
"text-generation", model=model_id, device="cuda",
trust_remote_code=True, max_new_tokens=3,
)
print(pipe("Hello World!"))
from transformers import AutoModelForCausalLM, AutoTokenizer
tokenizer = AutoTokenizer.from_pretrained(model_id)
model = AutoModelForCausalLM.from_pretrained(
model_id,
torch_dtype="auto",
device_map="auto"
)
prompt = "Give me a short introduction to large language model."
messages = [
{"role": "user", "content": prompt}
]
text = tokenizer.apply_chat_template(
messages,
tokenize=False,
add_generation_prompt=True,
enable_thinking=True # Switches between thinking and non-thinking modes. Default is True.
)
print(text)
model_inputs = tokenizer([text], return_tensors="pt").to(model.device)
generated_ids = model.generate(
**model_inputs,
max_new_tokens=128
)
output_ids = generated_ids[0][len(model_inputs.input_ids[0]):].tolist()
try:
# rindex finding 151668 (</think>)
index = len(output_ids) - output_ids[::-1].index(151668)
except ValueError:
index = 0
thinking_content = tokenizer.decode(output_ids[:index], skip_special_tokens=True).strip("\n")
content = tokenizer.decode(output_ids[index:], skip_special_tokens=True).strip("\n")
print("thinking content:", thinking_content)
print("content:", content)
```
### Codes to create this repo:
```python
import torch
from transformers import (
AutoConfig,
AutoModelForCausalLM,
AutoTokenizer,
GenerationConfig,
pipeline,
set_seed,
)
source_model_id = "Qwen/Qwen3-235B-A22B"
save_folder = "/tmp/yujiepan/qwen3-moe-tiny-random"
tokenizer = AutoTokenizer.from_pretrained(
source_model_id, trust_remote_code=True,
)
tokenizer.save_pretrained(save_folder)
config = AutoConfig.from_pretrained(
source_model_id, trust_remote_code=True,
)
config._name_or_path = source_model_id
config.hidden_size = 64
config.intermediate_size = 128
config.moe_intermediate_size = 128
config.head_dim = 32
config.decoder_sparse_step = 2 # layer0=mlp, layer1=moe
config.num_experts = 8
config.num_experts_per_tok = 2
config.num_key_value_heads = 1
config.num_attention_heads = 2
config.num_hidden_layers = 2
config.max_window_layers = 1
config.tie_word_embeddings = True
model = AutoModelForCausalLM.from_config(
config,
torch_dtype=torch.bfloat16,
trust_remote_code=True,
)
model.generation_config = GenerationConfig.from_pretrained(
source_model_id, trust_remote_code=True,
)
set_seed(42)
with torch.no_grad():
for name, p in sorted(model.named_parameters()):
torch.nn.init.normal_(p, 0, 0.5)
print(name, p.shape)
model.save_pretrained(save_folder)
``` |