Spaces:
Running
Running
Jeff Myers II
commited on
Commit
·
cb3e313
1
Parent(s):
297702e
Removed 8-bit quantization and changed model_id to google/gemma-3n-E4B-it-litert-preview
Browse files
Gemma.py
CHANGED
@@ -1,8 +1,6 @@
|
|
1 |
-
from transformers import pipeline
|
2 |
-
from transformers.utils import quantization_config
|
3 |
from huggingface_hub import login
|
4 |
import spaces
|
5 |
-
import torch
|
6 |
import json
|
7 |
import os
|
8 |
|
@@ -12,17 +10,18 @@ class GemmaLLM:
|
|
12 |
def __init__(self):
|
13 |
login(token=os.environ.get("GEMMA_TOKEN"))
|
14 |
|
15 |
-
quant_config = quantization_config.BitsAndBytesConfig(
|
16 |
-
|
17 |
-
|
18 |
-
|
19 |
-
)
|
20 |
|
21 |
-
model_id = "google/gemma-3-4b-it"
|
22 |
-
|
23 |
-
|
|
|
24 |
|
25 |
-
self.model = pipeline("text-generation",
|
26 |
|
27 |
@spaces.GPU
|
28 |
def generate(self, message) -> str:
|
|
|
1 |
+
from transformers import pipeline
|
|
|
2 |
from huggingface_hub import login
|
3 |
import spaces
|
|
|
4 |
import json
|
5 |
import os
|
6 |
|
|
|
10 |
def __init__(self):
|
11 |
login(token=os.environ.get("GEMMA_TOKEN"))
|
12 |
|
13 |
+
# quant_config = quantization_config.BitsAndBytesConfig(
|
14 |
+
# load_in_8bit=True,
|
15 |
+
# llm_int8_threshold=6.0,
|
16 |
+
# llm_int8_has_fp16_weight=False,
|
17 |
+
# )
|
18 |
|
19 |
+
# model_id = "google/gemma-3-4b-it"
|
20 |
+
model_id = "google/gemma-3n-E4B-it-litert-preview"
|
21 |
+
# model = AutoModelForCausalLM.from_pretrained(model_id, quantization_config=quant_config)
|
22 |
+
# tokenizer = AutoTokenizer.from_pretrained(model_id)
|
23 |
|
24 |
+
self.model = pipeline("text-generation", model_id)
|
25 |
|
26 |
@spaces.GPU
|
27 |
def generate(self, message) -> str:
|