|
|
|
|
|
""" |
|
|
Test script to verify the trained model works correctly. |
|
|
""" |
|
|
|
|
|
from transformers import AutoModelForCausalLM, AutoTokenizer |
|
|
import torch |
|
|
import time |
|
|
|
|
|
def test_model_loading(): |
|
|
"""Test that the model loads successfully.""" |
|
|
print("π§ͺ Testing model loading...") |
|
|
|
|
|
model_path = "/home/x/adaptai/experiments/qwen3-8b-elizabeth-simple/" |
|
|
|
|
|
|
|
|
print("π₯ Loading tokenizer...") |
|
|
tokenizer = AutoTokenizer.from_pretrained(model_path, trust_remote_code=True) |
|
|
print("β
Tokenizer loaded successfully!") |
|
|
|
|
|
|
|
|
print("π₯ Loading model...") |
|
|
start_time = time.time() |
|
|
|
|
|
model = AutoModelForCausalLM.from_pretrained( |
|
|
model_path, |
|
|
torch_dtype=torch.bfloat16, |
|
|
device_map="auto", |
|
|
trust_remote_code=True |
|
|
) |
|
|
|
|
|
load_time = time.time() - start_time |
|
|
print(f"β
Model loaded successfully in {load_time:.2f} seconds!") |
|
|
|
|
|
|
|
|
print(f"π Model device: {model.device}") |
|
|
print(f"π Model dtype: {model.dtype}") |
|
|
|
|
|
return model, tokenizer |
|
|
|
|
|
def test_inference(model, tokenizer): |
|
|
"""Test basic inference.""" |
|
|
print("\nπ§ͺ Testing inference...") |
|
|
|
|
|
|
|
|
prompt = "Hello, how are you today?" |
|
|
|
|
|
print(f"π Prompt: {prompt}") |
|
|
|
|
|
|
|
|
inputs = tokenizer(prompt, return_tensors="pt").to(model.device) |
|
|
print(f"π’ Input tokens: {inputs.input_ids.shape}") |
|
|
|
|
|
|
|
|
start_time = time.time() |
|
|
with torch.no_grad(): |
|
|
outputs = model.generate( |
|
|
**inputs, |
|
|
max_new_tokens=50, |
|
|
do_sample=True, |
|
|
temperature=0.7, |
|
|
pad_token_id=tokenizer.eos_token_id |
|
|
) |
|
|
|
|
|
gen_time = time.time() - start_time |
|
|
|
|
|
|
|
|
response = tokenizer.decode(outputs[0], skip_special_tokens=True) |
|
|
|
|
|
print(f"β
Generation completed in {gen_time:.2f} seconds!") |
|
|
print(f"π¬ Response: {response}") |
|
|
|
|
|
return response |
|
|
|
|
|
def test_tool_use_capability(model, tokenizer): |
|
|
"""Test tool use capability.""" |
|
|
print("\nπ§ͺ Testing tool use capability...") |
|
|
|
|
|
tool_prompt = """Please help me calculate the square root of 144 using the calculator tool. |
|
|
|
|
|
Available tools: |
|
|
- calculator: Performs mathematical calculations |
|
|
|
|
|
Please respond with the tool call format.""" |
|
|
|
|
|
print(f"π Tool prompt: {tool_prompt[:100]}...") |
|
|
|
|
|
inputs = tokenizer(tool_prompt, return_tensors="pt").to(model.device) |
|
|
|
|
|
with torch.no_grad(): |
|
|
outputs = model.generate( |
|
|
**inputs, |
|
|
max_new_tokens=100, |
|
|
do_sample=True, |
|
|
temperature=0.3, |
|
|
pad_token_id=tokenizer.eos_token_id |
|
|
) |
|
|
|
|
|
response = tokenizer.decode(outputs[0], skip_special_tokens=True) |
|
|
|
|
|
print(f"π¬ Tool use response: {response}") |
|
|
|
|
|
|
|
|
if "calculator" in response.lower() or "tool" in response.lower(): |
|
|
print("β
Tool use capability detected!") |
|
|
return True |
|
|
else: |
|
|
print("β οΈ Tool use pattern not clearly detected") |
|
|
return False |
|
|
|
|
|
if __name__ == "__main__": |
|
|
print("=" * 60) |
|
|
print("π€ Qwen3-8B-Elizabeth-Simple Model Test") |
|
|
print("=" * 60) |
|
|
|
|
|
try: |
|
|
model, tokenizer = test_model_loading() |
|
|
test_inference(model, tokenizer) |
|
|
tool_use_detected = test_tool_use_capability(model, tokenizer) |
|
|
|
|
|
print("\n" + "=" * 60) |
|
|
print("π Model Test Summary:") |
|
|
print(f" β
Model loading: Successful") |
|
|
print(f" β
Basic inference: Working") |
|
|
print(f" β
Tool use capability: {'Detected' if tool_use_detected else 'Needs verification'}") |
|
|
print(" π Model is ready for deployment!") |
|
|
print("=" * 60) |
|
|
|
|
|
except Exception as e: |
|
|
print(f"\nβ Test failed with error: {e}") |
|
|
raise |