Skip to content

Commit

Permalink
Multiple images for eval.run_llava (haotian-liu#432)
Browse files Browse the repository at this point in the history
  • Loading branch information
HireTheHero authored Nov 4, 2023
1 parent f751d7b commit caf8993
Show file tree
Hide file tree
Showing 2 changed files with 73 additions and 19 deletions.
1 change: 1 addition & 0 deletions llava/constants.py
Original file line number Diff line number Diff line change
Expand Up @@ -10,3 +10,4 @@
DEFAULT_IMAGE_PATCH_TOKEN = "<im_patch>"
DEFAULT_IM_START_TOKEN = "<im_start>"
DEFAULT_IM_END_TOKEN = "<im_end>"
IMAGE_PLACEHOLDER = "<image-placeholder>"
91 changes: 72 additions & 19 deletions llava/eval/run_llava.py
Original file line number Diff line number Diff line change
@@ -1,42 +1,75 @@
import argparse
import torch

from llava.constants import IMAGE_TOKEN_INDEX, DEFAULT_IMAGE_TOKEN, DEFAULT_IM_START_TOKEN, DEFAULT_IM_END_TOKEN
from llava.constants import (
IMAGE_TOKEN_INDEX,
DEFAULT_IMAGE_TOKEN,
DEFAULT_IM_START_TOKEN,
DEFAULT_IM_END_TOKEN,
IMAGE_PLACEHOLDER,
)
from llava.conversation import conv_templates, SeparatorStyle
from llava.model.builder import load_pretrained_model
from llava.utils import disable_torch_init
from llava.mm_utils import tokenizer_image_token, get_model_name_from_path, KeywordsStoppingCriteria
from llava.mm_utils import (
tokenizer_image_token,
get_model_name_from_path,
KeywordsStoppingCriteria,
)

from PIL import Image

import requests
from PIL import Image
from io import BytesIO
import re


def image_parser(args):
out = args.image_file.split(args.sep)
return out


def load_image(image_file):
if image_file.startswith('http') or image_file.startswith('https'):
if image_file.startswith("http") or image_file.startswith("https"):
response = requests.get(image_file)
image = Image.open(BytesIO(response.content)).convert('RGB')
image = Image.open(BytesIO(response.content)).convert("RGB")
else:
image = Image.open(image_file).convert('RGB')
image = Image.open(image_file).convert("RGB")
return image


def load_images(image_files):
out = []
for image_file in image_files:
image = load_image(image_file)
out.append(image)
return out


def eval_model(args):
# Model
disable_torch_init()

model_name = get_model_name_from_path(args.model_path)
tokenizer, model, image_processor, context_len = load_pretrained_model(args.model_path, args.model_base, model_name)
tokenizer, model, image_processor, context_len = load_pretrained_model(
args.model_path, args.model_base, model_name
)

qs = args.query
if model.config.mm_use_im_start_end:
qs = DEFAULT_IM_START_TOKEN + DEFAULT_IMAGE_TOKEN + DEFAULT_IM_END_TOKEN + '\n' + qs
image_token_se = DEFAULT_IM_START_TOKEN + DEFAULT_IMAGE_TOKEN + DEFAULT_IM_END_TOKEN
if IMAGE_PLACEHOLDER in qs:
if model.config.mm_use_im_start_end:
qs = re.sub(IMAGE_PLACEHOLDER, image_token_se, qs)
else:
qs = re.sub(IMAGE_PLACEHOLDER, DEFAULT_IMAGE_TOKEN, qs)
else:
qs = DEFAULT_IMAGE_TOKEN + '\n' + qs
if model.config.mm_use_im_start_end:
qs = image_token_se + "\n" + qs
else:
qs = DEFAULT_IMAGE_TOKEN + "\n" + qs

if 'llama-2' in model_name.lower():
if "llama-2" in model_name.lower():
conv_mode = "llava_llama_2"
elif "v1" in model_name.lower():
conv_mode = "llava_v1"
Expand All @@ -46,7 +79,11 @@ def eval_model(args):
conv_mode = "llava_v0"

if args.conv_mode is not None and conv_mode != args.conv_mode:
print('[WARNING] the auto inferred conversation mode is {}, while `--conv-mode` is {}, using {}'.format(conv_mode, args.conv_mode, args.conv_mode))
print(
"[WARNING] the auto inferred conversation mode is {}, while `--conv-mode` is {}, using {}".format(
conv_mode, args.conv_mode, args.conv_mode
)
)
else:
args.conv_mode = conv_mode

Expand All @@ -55,10 +92,19 @@ def eval_model(args):
conv.append_message(conv.roles[1], None)
prompt = conv.get_prompt()

image = load_image(args.image_file)
image_tensor = image_processor.preprocess(image, return_tensors='pt')['pixel_values'].half().cuda()
image_files = image_parser(args)
images = load_images(image_files)
images_tensor = (
image_processor.preprocess(images, return_tensors="pt")["pixel_values"]
.half()
.cuda()
)

input_ids = tokenizer_image_token(prompt, tokenizer, IMAGE_TOKEN_INDEX, return_tensors='pt').unsqueeze(0).cuda()
input_ids = (
tokenizer_image_token(prompt, tokenizer, IMAGE_TOKEN_INDEX, return_tensors="pt")
.unsqueeze(0)
.cuda()
)

stop_str = conv.sep if conv.sep_style != SeparatorStyle.TWO else conv.sep2
keywords = [stop_str]
Expand All @@ -67,31 +113,38 @@ def eval_model(args):
with torch.inference_mode():
output_ids = model.generate(
input_ids,
images=image_tensor,
images=images_tensor,
do_sample=True,
temperature=0.2,
max_new_tokens=1024,
use_cache=True,
stopping_criteria=[stopping_criteria])
stopping_criteria=[stopping_criteria],
)

input_token_len = input_ids.shape[1]
n_diff_input_output = (input_ids != output_ids[:, :input_token_len]).sum().item()
if n_diff_input_output > 0:
print(f'[Warning] {n_diff_input_output} output_ids are not the same as the input_ids')
outputs = tokenizer.batch_decode(output_ids[:, input_token_len:], skip_special_tokens=True)[0]
print(
f"[Warning] {n_diff_input_output} output_ids are not the same as the input_ids"
)
outputs = tokenizer.batch_decode(
output_ids[:, input_token_len:], skip_special_tokens=True
)[0]
outputs = outputs.strip()
if outputs.endswith(stop_str):
outputs = outputs[:-len(stop_str)]
outputs = outputs[: -len(stop_str)]
outputs = outputs.strip()
print(outputs)


if __name__ == "__main__":
parser = argparse.ArgumentParser()
parser.add_argument("--model-path", type=str, default="facebook/opt-350m")
parser.add_argument("--model-base", type=str, default=None)
parser.add_argument("--image-file", type=str, required=True)
parser.add_argument("--query", type=str, required=True)
parser.add_argument("--conv-mode", type=str, default=None)
parser.add_argument("--sep", type=str, default=",")
args = parser.parse_args()

eval_model(args)

0 comments on commit caf8993

Please sign in to comment.