Spaces:
Runtime error
Runtime error
Update app.py
Browse files
app.py
CHANGED
@@ -7,6 +7,18 @@ import transformers
|
|
7 |
from transformers import Pix2StructForConditionalGeneration, Pix2StructProcessor
|
8 |
from peft import PeftModel
|
9 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
10 |
## CoT prompts
|
11 |
|
12 |
def _add_markup(table):
|
@@ -24,7 +36,6 @@ def _add_markup(table):
|
|
24 |
# just use the raw table if parsing fails
|
25 |
return table
|
26 |
|
27 |
-
|
28 |
_TABLE = """Year | Democrats | Republicans | Independents
|
29 |
2004 | 68.1% | 45.0% | 53.0%
|
30 |
2006 | 58.0% | 42.0% | 53.0%
|
@@ -36,7 +47,6 @@ _TABLE = """Year | Democrats | Republicans | Independents
|
|
36 |
|
37 |
_INSTRUCTION = 'Read the table below to answer the following questions.'
|
38 |
|
39 |
-
|
40 |
_TEMPLATE = f"""First read an example then the complete question for the second table.
|
41 |
------------
|
42 |
{_INSTRUCTION}
|
@@ -56,7 +66,6 @@ A: Let's find the row of year 2007, that's Row 3. Let's extract the numbers on R
|
|
56 |
|
57 |
## alpaca-lora
|
58 |
|
59 |
-
# debugging...
|
60 |
assert (
|
61 |
"LlamaTokenizer" in transformers._import_structure["models.llama"]
|
62 |
), "LLaMA is now in HuggingFace's main branch.\nPlease reinstall it: pip uninstall transformers && pip install git+https://github.com/huggingface/transformers.git"
|
@@ -67,17 +76,6 @@ tokenizer = LlamaTokenizer.from_pretrained("decapoda-research/llama-7b-hf")
|
|
67 |
BASE_MODEL = "decapoda-research/llama-7b-hf"
|
68 |
LORA_WEIGHTS = "tloen/alpaca-lora-7b"
|
69 |
|
70 |
-
if torch.cuda.is_available():
|
71 |
-
device = "cuda"
|
72 |
-
else:
|
73 |
-
device = "cpu"
|
74 |
-
|
75 |
-
try:
|
76 |
-
if torch.backends.mps.is_available():
|
77 |
-
device = "mps"
|
78 |
-
except:
|
79 |
-
pass
|
80 |
-
|
81 |
if device == "cuda":
|
82 |
model = LlamaForCausalLM.from_pretrained(
|
83 |
BASE_MODEL,
|
@@ -147,7 +145,10 @@ def get_response_from_openai(prompt, model="gpt-3.5-turbo", max_output_tokens=25
|
|
147 |
return ret
|
148 |
|
149 |
## deplot models
|
150 |
-
|
|
|
|
|
|
|
151 |
processor_deplot = Pix2StructProcessor.from_pretrained("google/deplot")
|
152 |
|
153 |
def evaluate(
|
@@ -199,7 +200,9 @@ def evaluate(
|
|
199 |
|
200 |
def process_document(image, question, llm):
|
201 |
# image = Image.open(image)
|
202 |
-
inputs = processor_deplot(images=image, text="Generate the underlying data table for the figure below:", return_tensors="pt")
|
|
|
|
|
203 |
predictions = model_deplot.generate(**inputs, max_new_tokens=512)
|
204 |
table = processor_deplot.decode(predictions[0], skip_special_tokens=True).replace("<0x0A>", "\n")
|
205 |
|
|
|
7 |
from transformers import Pix2StructForConditionalGeneration, Pix2StructProcessor
|
8 |
from peft import PeftModel
|
9 |
|
10 |
+
|
11 |
+
if torch.cuda.is_available():
|
12 |
+
device = "cuda"
|
13 |
+
else:
|
14 |
+
device = "cpu"
|
15 |
+
|
16 |
+
try:
|
17 |
+
if torch.backends.mps.is_available():
|
18 |
+
device = "mps"
|
19 |
+
except:
|
20 |
+
pass
|
21 |
+
|
22 |
## CoT prompts
|
23 |
|
24 |
def _add_markup(table):
|
|
|
36 |
# just use the raw table if parsing fails
|
37 |
return table
|
38 |
|
|
|
39 |
_TABLE = """Year | Democrats | Republicans | Independents
|
40 |
2004 | 68.1% | 45.0% | 53.0%
|
41 |
2006 | 58.0% | 42.0% | 53.0%
|
|
|
47 |
|
48 |
_INSTRUCTION = 'Read the table below to answer the following questions.'
|
49 |
|
|
|
50 |
_TEMPLATE = f"""First read an example then the complete question for the second table.
|
51 |
------------
|
52 |
{_INSTRUCTION}
|
|
|
66 |
|
67 |
## alpaca-lora
|
68 |
|
|
|
69 |
assert (
|
70 |
"LlamaTokenizer" in transformers._import_structure["models.llama"]
|
71 |
), "LLaMA is now in HuggingFace's main branch.\nPlease reinstall it: pip uninstall transformers && pip install git+https://github.com/huggingface/transformers.git"
|
|
|
76 |
BASE_MODEL = "decapoda-research/llama-7b-hf"
|
77 |
LORA_WEIGHTS = "tloen/alpaca-lora-7b"
|
78 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
79 |
if device == "cuda":
|
80 |
model = LlamaForCausalLM.from_pretrained(
|
81 |
BASE_MODEL,
|
|
|
145 |
return ret
|
146 |
|
147 |
## deplot models
|
148 |
+
if device == "cuda":
|
149 |
+
model_deplot = Pix2StructForConditionalGeneration.from_pretrained("google/deplot", torch_dtype=torch.bfloat16).to(0)
|
150 |
+
else:
|
151 |
+
model_deplot = Pix2StructForConditionalGeneration.from_pretrained("google/deplot")
|
152 |
processor_deplot = Pix2StructProcessor.from_pretrained("google/deplot")
|
153 |
|
154 |
def evaluate(
|
|
|
200 |
|
201 |
def process_document(image, question, llm):
|
202 |
# image = Image.open(image)
|
203 |
+
inputs = processor_deplot(images=image, text="Generate the underlying data table for the figure below:", return_tensors="pt")
|
204 |
+
if device == "cuda":
|
205 |
+
inputs = inputs.to(0, torch.bfloat16)
|
206 |
predictions = model_deplot.generate(**inputs, max_new_tokens=512)
|
207 |
table = processor_deplot.decode(predictions[0], skip_special_tokens=True).replace("<0x0A>", "\n")
|
208 |
|