Spaces:
Running
on
Zero
Running
on
Zero
Roll back interruption changes
Browse files- utils/models.py +13 -69
utils/models.py
CHANGED
@@ -1,5 +1,5 @@
|
|
1 |
import os
|
2 |
-
#
|
3 |
import torch._dynamo
|
4 |
torch._dynamo.config.suppress_errors = True
|
5 |
|
@@ -17,7 +17,8 @@ from transformers import (
|
|
17 |
BitNetForCausalLM
|
18 |
)
|
19 |
from .prompts import format_rag_prompt
|
20 |
-
|
|
|
21 |
|
22 |
models = {
|
23 |
"Qwen2.5-1.5b-Instruct": "qwen/qwen2.5-1.5b-instruct",
|
@@ -47,13 +48,13 @@ tokenizer_cache = {}
|
|
47 |
model_names = list(models.keys())
|
48 |
|
49 |
|
50 |
-
#
|
51 |
-
class InterruptCriteria(StoppingCriteria):
|
52 |
-
|
53 |
-
|
54 |
-
|
55 |
-
|
56 |
-
|
57 |
|
58 |
|
59 |
@spaces.GPU
|
@@ -61,20 +62,12 @@ def generate_summaries(example, model_a_name, model_b_name):
|
|
61 |
"""
|
62 |
Generates summaries for the given example using the assigned models sequentially.
|
63 |
"""
|
64 |
-
|
65 |
-
print("Generation interrupted before starting")
|
66 |
-
return "", ""
|
67 |
-
|
68 |
context_text = ""
|
69 |
context_parts = []
|
70 |
|
71 |
if "full_contexts" in example and example["full_contexts"]:
|
72 |
for i, ctx in enumerate(example["full_contexts"]):
|
73 |
-
# Check interrupt during context processing
|
74 |
-
if generation_interrupt.is_set():
|
75 |
-
print("Generation interrupted during context processing")
|
76 |
-
return "", ""
|
77 |
-
|
78 |
content = ""
|
79 |
|
80 |
# Extract content from either dict or string
|
@@ -97,18 +90,10 @@ def generate_summaries(example, model_a_name, model_b_name):
|
|
97 |
|
98 |
question = example.get("question", "")
|
99 |
|
100 |
-
if generation_interrupt.is_set():
|
101 |
-
print("Generation interrupted before model A")
|
102 |
-
return "", ""
|
103 |
-
|
104 |
print(f"Starting inference for Model A: {model_a_name}")
|
105 |
# Run model A
|
106 |
summary_a = run_inference(models[model_a_name], context_text, question)
|
107 |
|
108 |
-
if generation_interrupt.is_set():
|
109 |
-
print("Generation interrupted after model A, before model B")
|
110 |
-
return summary_a, ""
|
111 |
-
|
112 |
print(f"Starting inference for Model B: {model_b_name}")
|
113 |
# Run model B
|
114 |
summary_b = run_inference(models[model_b_name], context_text, question)
|
@@ -121,13 +106,8 @@ def generate_summaries(example, model_a_name, model_b_name):
|
|
121 |
def run_inference(model_name, context, question):
|
122 |
"""
|
123 |
Run inference using the specified model.
|
124 |
-
Returns the generated text
|
125 |
"""
|
126 |
-
# Check interrupt at the beginning
|
127 |
-
if generation_interrupt.is_set():
|
128 |
-
print(f"Inference interrupted before starting for {model_name}")
|
129 |
-
return ""
|
130 |
-
|
131 |
device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
|
132 |
result = ""
|
133 |
tokenizer_kwargs = {
|
@@ -146,11 +126,6 @@ def run_inference(model_name, context, question):
|
|
146 |
if model_name in tokenizer_cache:
|
147 |
tokenizer = tokenizer_cache[model_name]
|
148 |
else:
|
149 |
-
# Check interrupt before loading tokenizer
|
150 |
-
if generation_interrupt.is_set():
|
151 |
-
print(f"Inference interrupted before loading tokenizer for {model_name}")
|
152 |
-
return ""
|
153 |
-
|
154 |
# Common arguments for tokenizer loading
|
155 |
tokenizer_load_args = {"padding_side": "left", "token": True}
|
156 |
|
@@ -170,21 +145,8 @@ def run_inference(model_name, context, question):
|
|
170 |
if tokenizer.pad_token is None:
|
171 |
tokenizer.pad_token = tokenizer.eos_token
|
172 |
|
173 |
-
# Check interrupt before loading the model
|
174 |
-
if generation_interrupt.is_set():
|
175 |
-
print(f"Inference interrupted before loading model {model_name}")
|
176 |
-
return ""
|
177 |
-
|
178 |
-
# Create interrupt criteria for this generation
|
179 |
-
interrupt_criteria = InterruptCriteria(generation_interrupt)
|
180 |
-
|
181 |
print("REACHED HERE BEFORE pipe")
|
182 |
print(f"Loading model {model_name}...")
|
183 |
-
|
184 |
-
# Check interrupt before model loading
|
185 |
-
if generation_interrupt.is_set():
|
186 |
-
print(f"Inference interrupted during model loading for {model_name}")
|
187 |
-
return ""
|
188 |
|
189 |
if "bitnet" in model_name.lower():
|
190 |
bitnet_model = BitNetForCausalLM.from_pretrained(
|
@@ -226,11 +188,6 @@ def run_inference(model_name, context, question):
|
|
226 |
torch_dtype=torch.bfloat16,
|
227 |
)
|
228 |
|
229 |
-
# Final interrupt check before generation
|
230 |
-
if generation_interrupt.is_set():
|
231 |
-
print(f"Inference interrupted before generation for {model_name}")
|
232 |
-
return ""
|
233 |
-
|
234 |
text_input = format_rag_prompt(question, context, accepts_sys)
|
235 |
|
236 |
print(f"Starting generation for {model_name}")
|
@@ -239,7 +196,6 @@ def run_inference(model_name, context, question):
|
|
239 |
result = pipe(
|
240 |
text_input,
|
241 |
max_new_tokens=512,
|
242 |
-
stopping_criteria=[interrupt_criteria],
|
243 |
generation_kwargs={"skip_special_tokens": True}
|
244 |
)[0]["generated_text"]
|
245 |
|
@@ -263,18 +219,12 @@ def run_inference(model_name, context, question):
|
|
263 |
prompt_tokens_length = input_ids.shape[1]
|
264 |
|
265 |
with torch.inference_mode():
|
266 |
-
# Check interrupt before generation
|
267 |
-
if generation_interrupt.is_set():
|
268 |
-
print(f"Inference interrupted before torch generation for {model_name}")
|
269 |
-
return ""
|
270 |
-
|
271 |
output_sequences = model.generate(
|
272 |
input_ids=input_ids,
|
273 |
attention_mask=attention_mask,
|
274 |
max_new_tokens=512,
|
275 |
eos_token_id=tokenizer.eos_token_id,
|
276 |
-
pad_token_id=tokenizer.pad_token_id
|
277 |
-
stopping_criteria=[interrupt_criteria]
|
278 |
)
|
279 |
|
280 |
generated_token_ids = output_sequences[0][prompt_tokens_length:]
|
@@ -288,15 +238,10 @@ def run_inference(model_name, context, question):
|
|
288 |
# **tokenizer_kwargs,
|
289 |
# ).to(bitnet_model.device)
|
290 |
# with torch.inference_mode():
|
291 |
-
# # Check interrupt before generation
|
292 |
-
# if generation_interrupt.is_set():
|
293 |
-
# return ""
|
294 |
# output_sequences = bitnet_model.generate(
|
295 |
# **formatted,
|
296 |
# max_new_tokens=512,
|
297 |
-
# stopping_criteria=[interrupt_criteria]
|
298 |
# )
|
299 |
-
|
300 |
# result = tokenizer.decode(output_sequences[0][formatted['input_ids'].shape[-1]:], skip_special_tokens=True)
|
301 |
else: # For other models
|
302 |
formatted = pipe.tokenizer.apply_chat_template(
|
@@ -310,7 +255,6 @@ def run_inference(model_name, context, question):
|
|
310 |
outputs = pipe(
|
311 |
formatted,
|
312 |
max_new_tokens=512,
|
313 |
-
stopping_criteria=[interrupt_criteria],
|
314 |
generation_kwargs={"skip_special_tokens": True}
|
315 |
)
|
316 |
result = outputs[0]["generated_text"][input_length:]
|
|
|
1 |
import os
|
2 |
+
# Keep Dynamo error suppression
|
3 |
import torch._dynamo
|
4 |
torch._dynamo.config.suppress_errors = True
|
5 |
|
|
|
17 |
BitNetForCausalLM
|
18 |
)
|
19 |
from .prompts import format_rag_prompt
|
20 |
+
# Remove interrupt import
|
21 |
+
# from .shared import generation_interrupt
|
22 |
|
23 |
models = {
|
24 |
"Qwen2.5-1.5b-Instruct": "qwen/qwen2.5-1.5b-instruct",
|
|
|
48 |
model_names = list(models.keys())
|
49 |
|
50 |
|
51 |
+
# Remove interrupt criteria class since we're not using it
|
52 |
+
# class InterruptCriteria(StoppingCriteria):
|
53 |
+
# def __init__(self, interrupt_event):
|
54 |
+
# self.interrupt_event = interrupt_event
|
55 |
+
#
|
56 |
+
# def __call__(self, input_ids, scores, **kwargs):
|
57 |
+
# return self.interrupt_event.is_set()
|
58 |
|
59 |
|
60 |
@spaces.GPU
|
|
|
62 |
"""
|
63 |
Generates summaries for the given example using the assigned models sequentially.
|
64 |
"""
|
65 |
+
# Remove interrupt checks
|
|
|
|
|
|
|
66 |
context_text = ""
|
67 |
context_parts = []
|
68 |
|
69 |
if "full_contexts" in example and example["full_contexts"]:
|
70 |
for i, ctx in enumerate(example["full_contexts"]):
|
|
|
|
|
|
|
|
|
|
|
71 |
content = ""
|
72 |
|
73 |
# Extract content from either dict or string
|
|
|
90 |
|
91 |
question = example.get("question", "")
|
92 |
|
|
|
|
|
|
|
|
|
93 |
print(f"Starting inference for Model A: {model_a_name}")
|
94 |
# Run model A
|
95 |
summary_a = run_inference(models[model_a_name], context_text, question)
|
96 |
|
|
|
|
|
|
|
|
|
97 |
print(f"Starting inference for Model B: {model_b_name}")
|
98 |
# Run model B
|
99 |
summary_b = run_inference(models[model_b_name], context_text, question)
|
|
|
106 |
def run_inference(model_name, context, question):
|
107 |
"""
|
108 |
Run inference using the specified model.
|
109 |
+
Returns the generated text.
|
110 |
"""
|
|
|
|
|
|
|
|
|
|
|
111 |
device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
|
112 |
result = ""
|
113 |
tokenizer_kwargs = {
|
|
|
126 |
if model_name in tokenizer_cache:
|
127 |
tokenizer = tokenizer_cache[model_name]
|
128 |
else:
|
|
|
|
|
|
|
|
|
|
|
129 |
# Common arguments for tokenizer loading
|
130 |
tokenizer_load_args = {"padding_side": "left", "token": True}
|
131 |
|
|
|
145 |
if tokenizer.pad_token is None:
|
146 |
tokenizer.pad_token = tokenizer.eos_token
|
147 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
148 |
print("REACHED HERE BEFORE pipe")
|
149 |
print(f"Loading model {model_name}...")
|
|
|
|
|
|
|
|
|
|
|
150 |
|
151 |
if "bitnet" in model_name.lower():
|
152 |
bitnet_model = BitNetForCausalLM.from_pretrained(
|
|
|
188 |
torch_dtype=torch.bfloat16,
|
189 |
)
|
190 |
|
|
|
|
|
|
|
|
|
|
|
191 |
text_input = format_rag_prompt(question, context, accepts_sys)
|
192 |
|
193 |
print(f"Starting generation for {model_name}")
|
|
|
196 |
result = pipe(
|
197 |
text_input,
|
198 |
max_new_tokens=512,
|
|
|
199 |
generation_kwargs={"skip_special_tokens": True}
|
200 |
)[0]["generated_text"]
|
201 |
|
|
|
219 |
prompt_tokens_length = input_ids.shape[1]
|
220 |
|
221 |
with torch.inference_mode():
|
|
|
|
|
|
|
|
|
|
|
222 |
output_sequences = model.generate(
|
223 |
input_ids=input_ids,
|
224 |
attention_mask=attention_mask,
|
225 |
max_new_tokens=512,
|
226 |
eos_token_id=tokenizer.eos_token_id,
|
227 |
+
pad_token_id=tokenizer.pad_token_id
|
|
|
228 |
)
|
229 |
|
230 |
generated_token_ids = output_sequences[0][prompt_tokens_length:]
|
|
|
238 |
# **tokenizer_kwargs,
|
239 |
# ).to(bitnet_model.device)
|
240 |
# with torch.inference_mode():
|
|
|
|
|
|
|
241 |
# output_sequences = bitnet_model.generate(
|
242 |
# **formatted,
|
243 |
# max_new_tokens=512,
|
|
|
244 |
# )
|
|
|
245 |
# result = tokenizer.decode(output_sequences[0][formatted['input_ids'].shape[-1]:], skip_special_tokens=True)
|
246 |
else: # For other models
|
247 |
formatted = pipe.tokenizer.apply_chat_template(
|
|
|
255 |
outputs = pipe(
|
256 |
formatted,
|
257 |
max_new_tokens=512,
|
|
|
258 |
generation_kwargs={"skip_special_tokens": True}
|
259 |
)
|
260 |
result = outputs[0]["generated_text"][input_length:]
|