Update handler.py
Browse files- handler.py +22 -17
handler.py
CHANGED
@@ -1,12 +1,10 @@
|
|
1 |
-
from typing import Any, Dict
|
2 |
-
|
3 |
import torch
|
4 |
import transformers
|
5 |
from transformers import AutoModelForCausalLM, AutoTokenizer
|
6 |
-
|
7 |
dtype = torch.bfloat16 if torch.cuda.get_device_capability()[0] == 8 else torch.float16
|
8 |
-
|
9 |
-
|
10 |
class EndpointHandler:
|
11 |
def __init__(self, path=""):
|
12 |
tokenizer = AutoTokenizer.from_pretrained(path, trust_remote_code=True)
|
@@ -18,20 +16,27 @@ class EndpointHandler:
|
|
18 |
torch_dtype=dtype,
|
19 |
trust_remote_code=True,
|
20 |
)
|
21 |
-
|
22 |
-
generation_config = model.generation_config
|
23 |
-
generation_config.max_new_tokens = 1000
|
24 |
-
generation_config.temperature = 0
|
25 |
-
generation_config.num_return_sequences = 1
|
26 |
-
generation_config.pad_token_id = tokenizer.eos_token_id
|
27 |
-
generation_config.eos_token_id = tokenizer.eos_token_id
|
28 |
-
|
29 |
-
|
30 |
self.pipeline = transformers.pipeline(
|
31 |
"text-generation", model=model, tokenizer=tokenizer
|
32 |
)
|
33 |
-
|
34 |
def __call__(self, data: Dict[str, Any]) -> Dict[str, Any]:
|
35 |
prompt = data.pop("inputs", data)
|
36 |
-
result = self.pipeline(
|
37 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
from typing import Any, Dict
|
|
|
2 |
import torch
|
3 |
import transformers
|
4 |
from transformers import AutoModelForCausalLM, AutoTokenizer
|
5 |
+
|
6 |
dtype = torch.bfloat16 if torch.cuda.get_device_capability()[0] == 8 else torch.float16
|
7 |
+
|
|
|
8 |
class EndpointHandler:
|
9 |
def __init__(self, path=""):
|
10 |
tokenizer = AutoTokenizer.from_pretrained(path, trust_remote_code=True)
|
|
|
16 |
torch_dtype=dtype,
|
17 |
trust_remote_code=True,
|
18 |
)
|
19 |
+
|
20 |
+
self.generation_config = model.generation_config
|
21 |
+
self.generation_config.max_new_tokens = 1000
|
22 |
+
self.generation_config.temperature = 0.7 # Changed from 0 to 0.7
|
23 |
+
self.generation_config.num_return_sequences = 1
|
24 |
+
self.generation_config.pad_token_id = tokenizer.eos_token_id
|
25 |
+
self.generation_config.eos_token_id = tokenizer.eos_token_id
|
26 |
+
|
|
|
27 |
self.pipeline = transformers.pipeline(
|
28 |
"text-generation", model=model, tokenizer=tokenizer
|
29 |
)
|
30 |
+
|
31 |
def __call__(self, data: Dict[str, Any]) -> Dict[str, Any]:
|
32 |
prompt = data.pop("inputs", data)
|
33 |
+
result = self.pipeline(
|
34 |
+
prompt,
|
35 |
+
max_length=1000, # Added this line to set max_length
|
36 |
+
temperature=0.7, # Added this line to set temperature
|
37 |
+
top_p=0.9, # Added this line to set top_p
|
38 |
+
num_return_sequences=1, # Added this line to set num_return_sequences
|
39 |
+
pad_token_id=self.generation_config.pad_token_id,
|
40 |
+
eos_token_id=self.generation_config.eos_token_id
|
41 |
+
)
|
42 |
+
return {"generated_text": result}
|