How to use multiple GPUs for inference? #12

waltonfuture · 2024-08-23T04:59:00Z

I want to use multiple GPUs for inference, and I use device_map='auto' to load the model. However, I always met that problem: Expected all tensors to be on the same device, but found at least two devices, cuda:1 and cuda:0!

Can you help me with that? Thanks a lot!

runninglsy · 2024-09-03T04:08:56Z

Below is an example of running AIDC-AI/Ovis1.5-Llama3-8B on two GPUs:

import torch
from PIL import Image
from transformers import AutoModelForCausalLM

device_map = {
    "visual_tokenizer": 0,
    "vte": 0,
    "llm.model.embed_tokens": 0,
    "llm.model.norm": 0,
    "llm.lm_head": 0,
    "llm.model.layers.0": 0,
    "llm.model.layers.1": 0,
    "llm.model.layers.2": 0,
    "llm.model.layers.3": 0,
    "llm.model.layers.4": 0,
    "llm.model.layers.5": 0,
    "llm.model.layers.6": 0,
    "llm.model.layers.7": 0,
    "llm.model.layers.8": 0,
    "llm.model.layers.9": 0,
    "llm.model.layers.10": 0,
    "llm.model.layers.11": 0,
    "llm.model.layers.12": 0,
    "llm.model.layers.13": 0,
    "llm.model.layers.14": 1,
    "llm.model.layers.15": 1,
    "llm.model.layers.16": 1,
    "llm.model.layers.17": 1,
    "llm.model.layers.18": 1,
    "llm.model.layers.19": 1,
    "llm.model.layers.20": 1,
    "llm.model.layers.21": 1,
    "llm.model.layers.22": 1,
    "llm.model.layers.23": 1,
    "llm.model.layers.24": 1,
    "llm.model.layers.25": 1,
    "llm.model.layers.26": 1,
    "llm.model.layers.27": 1,
    "llm.model.layers.28": 1,
    "llm.model.layers.29": 1,
    "llm.model.layers.30": 1,
    "llm.model.layers.31": 1
}

# load model
model = AutoModelForCausalLM.from_pretrained("AIDC-AI/Ovis1.5-Llama3-8B",
                                             torch_dtype=torch.bfloat16,
                                             multimodal_max_length=8192,
					     device_map=device_map,
                                             trust_remote_code=True)
text_tokenizer = model.get_text_tokenizer()
visual_tokenizer = model.get_visual_tokenizer()
conversation_formatter = model.get_conversation_formatter()

# enter image path and prompt
image_path = input("Enter image path: ")
image = Image.open(image_path)
text = input("Enter prompt: ")
query = f'<image>\n{text}'
prompt, input_ids = conversation_formatter.format_query(query)
input_ids = torch.unsqueeze(input_ids, dim=0).cuda()
attention_mask = torch.ne(input_ids, text_tokenizer.pad_token_id).cuda()
pixel_values = [visual_tokenizer.preprocess_image(image).to(
    dtype=visual_tokenizer.dtype, device=visual_tokenizer.device)]

# generate output
with torch.inference_mode():
    gen_kwargs = dict(
        max_new_tokens=1024,
        do_sample=False,
        top_p=None,
        top_k=None,
        temperature=None,
        repetition_penalty=None,
        eos_token_id=model.generation_config.eos_token_id,
        pad_token_id=text_tokenizer.pad_token_id,
        use_cache=True
    )
    output_ids = model.generate(input_ids, pixel_values=pixel_values, attention_mask=attention_mask, **gen_kwargs)[0]
    output = text_tokenizer.decode(output_ids, skip_special_tokens=True)
    print(f'Output: {output}')

cyj95 · 2024-09-11T02:34:15Z

Below is an example of running AIDC-AI/Ovis1.5-Llama3-8B on two GPUs:

import torch
from PIL import Image
from transformers import AutoModelForCausalLM

device_map = {
    "visual_tokenizer": 0,
    "vte": 0,
    "llm.model.embed_tokens": 0,
    "llm.model.norm": 0,
    "llm.lm_head": 0,
    "llm.model.layers.0": 0,
    "llm.model.layers.1": 0,
    "llm.model.layers.2": 0,
    "llm.model.layers.3": 0,
    "llm.model.layers.4": 0,
    "llm.model.layers.5": 0,
    "llm.model.layers.6": 0,
    "llm.model.layers.7": 0,
    "llm.model.layers.8": 0,
    "llm.model.layers.9": 0,
    "llm.model.layers.10": 0,
    "llm.model.layers.11": 0,
    "llm.model.layers.12": 0,
    "llm.model.layers.13": 0,
    "llm.model.layers.14": 1,
    "llm.model.layers.15": 1,
    "llm.model.layers.16": 1,
    "llm.model.layers.17": 1,
    "llm.model.layers.18": 1,
    "llm.model.layers.19": 1,
    "llm.model.layers.20": 1,
    "llm.model.layers.21": 1,
    "llm.model.layers.22": 1,
    "llm.model.layers.23": 1,
    "llm.model.layers.24": 1,
    "llm.model.layers.25": 1,
    "llm.model.layers.26": 1,
    "llm.model.layers.27": 1,
    "llm.model.layers.28": 1,
    "llm.model.layers.29": 1,
    "llm.model.layers.30": 1,
    "llm.model.layers.31": 1
}

# load model
model = AutoModelForCausalLM.from_pretrained("AIDC-AI/Ovis1.5-Llama3-8B",
                                             torch_dtype=torch.bfloat16,
                                             multimodal_max_length=8192,
					     device_map=device_map,
                                             trust_remote_code=True)
text_tokenizer = model.get_text_tokenizer()
visual_tokenizer = model.get_visual_tokenizer()
conversation_formatter = model.get_conversation_formatter()

# enter image path and prompt
image_path = input("Enter image path: ")
image = Image.open(image_path)
text = input("Enter prompt: ")
query = f'<image>\n{text}'
prompt, input_ids = conversation_formatter.format_query(query)
input_ids = torch.unsqueeze(input_ids, dim=0).cuda()
attention_mask = torch.ne(input_ids, text_tokenizer.pad_token_id).cuda()
pixel_values = [visual_tokenizer.preprocess_image(image).to(
    dtype=visual_tokenizer.dtype, device=visual_tokenizer.device)]

# generate output
with torch.inference_mode():
    gen_kwargs = dict(
        max_new_tokens=1024,
        do_sample=False,
        top_p=None,
        top_k=None,
        temperature=None,
        repetition_penalty=None,
        eos_token_id=model.generation_config.eos_token_id,
        pad_token_id=text_tokenizer.pad_token_id,
        use_cache=True
    )
    output_ids = model.generate(input_ids, pixel_values=pixel_values, attention_mask=attention_mask, **gen_kwargs)[0]
    output = text_tokenizer.decode(output_ids, skip_special_tokens=True)
    print(f'Output: {output}')

what's the script for Ovis1.5-Gemma2-9B

AntyRia · 2024-09-20T12:34:47Z

Did you solve it

runninglsy · 2024-09-23T02:51:21Z

Below is an example of running AIDC-AI/Ovis1.5-Gemma2-9B on two GPUs:

import torch
from PIL import Image
from transformers import AutoModelForCausalLM

device_map = {
    "visual_tokenizer": 0,
    "vte": 0,
    "llm.model.embed_tokens": 0,
    "llm.model.norm": 0,
    "llm.lm_head": 0,
    "llm.model.layers.0": 0,
    "llm.model.layers.1": 0,
    "llm.model.layers.2": 0,
    "llm.model.layers.3": 0,
    "llm.model.layers.4": 0,
    "llm.model.layers.5": 0,
    "llm.model.layers.6": 0,
    "llm.model.layers.7": 0,
    "llm.model.layers.8": 0,
    "llm.model.layers.9": 0,
    "llm.model.layers.10": 0,
    "llm.model.layers.11": 0,
    "llm.model.layers.12": 0,
    "llm.model.layers.13": 0,
    "llm.model.layers.14": 0,
    "llm.model.layers.15": 0,
    "llm.model.layers.16": 0,
    "llm.model.layers.17": 0,
    "llm.model.layers.18": 0,
    "llm.model.layers.19": 0,
    "llm.model.layers.20": 1,
    "llm.model.layers.21": 1,
    "llm.model.layers.22": 1,
    "llm.model.layers.23": 1,
    "llm.model.layers.24": 1,
    "llm.model.layers.25": 1,
    "llm.model.layers.26": 1,
    "llm.model.layers.27": 1,
    "llm.model.layers.28": 1,
    "llm.model.layers.29": 1,
    "llm.model.layers.30": 1,
    "llm.model.layers.31": 1,
    "llm.model.layers.32": 1,
    "llm.model.layers.33": 1,
    "llm.model.layers.34": 1,
    "llm.model.layers.35": 1,
    "llm.model.layers.36": 1,
    "llm.model.layers.37": 1,
    "llm.model.layers.38": 1,
    "llm.model.layers.39": 1,
    "llm.model.layers.40": 1,
    "llm.model.layers.41": 1
}

# load model
model = AutoModelForCausalLM.from_pretrained("AIDC-AI/Ovis1.5-Gemma2-9B",
                                             torch_dtype=torch.bfloat16,
                                             multimodal_max_length=8192,
					     device_map=device_map,
                                             trust_remote_code=True)
text_tokenizer = model.get_text_tokenizer()
visual_tokenizer = model.get_visual_tokenizer()
conversation_formatter = model.get_conversation_formatter()

# enter image path and prompt
image_path = input("Enter image path: ")
image = Image.open(image_path)
text = input("Enter prompt: ")
query = f'<image>\n{text}'
prompt, input_ids = conversation_formatter.format_query(query)
input_ids = torch.unsqueeze(input_ids, dim=0).cuda()
attention_mask = torch.ne(input_ids, text_tokenizer.pad_token_id).cuda()
pixel_values = [visual_tokenizer.preprocess_image(image).to(
    dtype=visual_tokenizer.dtype, device=visual_tokenizer.device)]

# generate output
with torch.inference_mode():
    gen_kwargs = dict(
        max_new_tokens=1024,
        do_sample=False,
        top_p=None,
        top_k=None,
        temperature=None,
        repetition_penalty=None,
        eos_token_id=model.generation_config.eos_token_id,
        pad_token_id=text_tokenizer.pad_token_id,
        use_cache=True
    )
    output_ids = model.generate(input_ids, pixel_values=pixel_values, attention_mask=attention_mask, **gen_kwargs)[0]
    output = text_tokenizer.decode(output_ids, skip_special_tokens=True)
    print(f'Output: {output}')

CVHub520 · 2024-09-23T09:09:03Z

@cyj95, @AntyRia, here’s the code, try running this:

from dataclasses import field, dataclass
from typing import Optional, Union, List

import torch
from PIL import Image

from ovis.model.modeling_ovis import Ovis
from ovis.util.constants import IMAGE_TOKEN


device_map = {
    "visual_tokenizer": 0,
    "vte": 0,
    "llm.model.embed_tokens": 0,
    "llm.model.norm": 0,
    "llm.lm_head": 0,
    "llm.model.layers.0": 0,
    "llm.model.layers.1": 0,
    "llm.model.layers.2": 0,
    "llm.model.layers.3": 0,
    "llm.model.layers.4": 0,
    "llm.model.layers.5": 0,
    "llm.model.layers.6": 0,
    "llm.model.layers.7": 0,
    "llm.model.layers.8": 0,
    "llm.model.layers.9": 0,
    "llm.model.layers.10": 0,
    "llm.model.layers.11": 0,
    "llm.model.layers.12": 0,
    "llm.model.layers.13": 0,
    "llm.model.layers.14": 0,
    "llm.model.layers.15": 0,
    "llm.model.layers.16": 0,
    "llm.model.layers.17": 0,
    "llm.model.layers.18": 0,
    "llm.model.layers.19": 0,
    "llm.model.layers.20": 1,
    "llm.model.layers.21": 1,
    "llm.model.layers.22": 1,
    "llm.model.layers.23": 1,
    "llm.model.layers.24": 1,
    "llm.model.layers.25": 1,
    "llm.model.layers.26": 1,
    "llm.model.layers.27": 1,
    "llm.model.layers.28": 1,
    "llm.model.layers.29": 1,
    "llm.model.layers.30": 1,
    "llm.model.layers.31": 1,
    "llm.model.layers.32": 1,
    "llm.model.layers.33": 1,
    "llm.model.layers.34": 1,
    "llm.model.layers.35": 1,
    "llm.model.layers.36": 1,
    "llm.model.layers.37": 1,
    "llm.model.layers.38": 1,
    "llm.model.layers.39": 1,
    "llm.model.layers.40": 1,
    "llm.model.layers.41": 1
}


@dataclass
class RunnerArguments:
    model_path: str
    max_new_tokens: int = field(default=512)
    do_sample: bool = field(default=False)
    top_p: Optional[float] = field(default=None)
    top_k: Optional[int] = field(default=None)
    temperature: Optional[float] = field(default=None)
    max_partition: int = field(default=9)


class OvisRunner:
    def __init__(self, args: RunnerArguments):
        self.model_path = args.model_path
        self.dtype = torch.bfloat16
        self.device = torch.cuda.current_device()
        self.dtype = torch.bfloat16
        self.model = Ovis.from_pretrained(self.model_path, torch_dtype=self.dtype, multimodal_max_length=8192, device_map=device_map)
        self.model = self.model.eval()
        self.eos_token_id = self.model.generation_config.eos_token_id
        self.text_tokenizer = self.model.get_text_tokenizer()
        self.pad_token_id = self.text_tokenizer.pad_token_id
        self.visual_tokenizer = self.model.get_visual_tokenizer()
        self.conversation_formatter = self.model.get_conversation_formatter()
        self.image_placeholder = IMAGE_TOKEN
        self.max_partition = args.max_partition
        self.gen_kwargs = dict(
            max_new_tokens=args.max_new_tokens,
            do_sample=args.do_sample,
            top_p=args.top_p,
            top_k=args.top_k,
            temperature=args.temperature,
            repetition_penalty=None,
            eos_token_id=self.eos_token_id,
            pad_token_id=self.pad_token_id,
            use_cache=True
        )

    def preprocess(self, inputs: List[Union[Image.Image, str]]):
        # for single image and single text inputs, ensure image ahead
        if len(inputs) == 2 and isinstance(inputs[0], str) and isinstance(inputs[1], Image.Image):
            inputs = reversed(inputs)

        # build query
        query = ''
        images = []
        for data in inputs:
            if isinstance(data, Image.Image):
                query += self.image_placeholder + '\n'
                images.append(data)
            elif isinstance(data, str):
                query += data.replace(self.image_placeholder, '')
            elif data is not None:
                raise RuntimeError(f'Invalid input type, expected `PIL.Image.Image` or `str`, but got {type(data)}')

        # format conversation
        prompt, input_ids, pixel_values = self.model.preprocess_inputs(
            query, images, max_partition=self.max_partition)
        attention_mask = torch.ne(input_ids, self.text_tokenizer.pad_token_id)
        input_ids = input_ids.unsqueeze(0).to(device=self.device)
        attention_mask = attention_mask.unsqueeze(0).to(device=self.device)
        if pixel_values is not None:
            pixel_values = [pixel_values.to(device=self.device, dtype=self.dtype)]
        else:
            pixel_values = [None]

        return prompt, input_ids, attention_mask, pixel_values

    def run(self, inputs: List[Union[Image.Image, str]]):
        prompt, input_ids, attention_mask, pixel_values = self.preprocess(inputs)
        output_ids = self.model.generate(
            input_ids,
            pixel_values=pixel_values,
            attention_mask=attention_mask,
            **self.gen_kwargs
        )
        output = self.text_tokenizer.decode(output_ids[0], skip_special_tokens=True)
        input_token_len = input_ids.shape[1]
        output_token_len = output_ids.shape[1]
        response = dict(
            prompt=prompt,
            output=output,
            prompt_tokens=input_token_len,
            total_tokens=input_token_len + output_token_len
        )
        return response


if __name__ == '__main__':
    text = <prompt>
    image_path = <image_path>
    model_path = <model_path>

    runner_args = RunnerArguments(model_path=model_path)
    runner = OvisRunner(runner_args)
    image = Image.open(image_path)
    response = runner.run([image, text])
    print(response['output'])

LuciusLan · 2024-09-30T10:55:38Z

Hi, for Ovis1.6-Gemma2-9B the device mapping issue persists for transformers version 4.45.1, seems the inconsistent device allocation is happening on HybridCache used (basically, when mannually set the device_map like previous posts, the cache is not set with same map so for layers 20 afterward, all cache tensors are on cuda0 but incoming key_states, value_states are on cuda1), and somehow passing use_cache=False in gen_kwargs still cannot disable usage of cache.
My current workaround is changing line 596-597 in modelling_ovis.py to:

layer_device_map = [0] * 20 + [1] * 22
kwargs['past_key_values'] = self._get_hybrid_cache_for_llm(
    getattr(kwargs, "num_beams", inputs_embeds.shape[0]), kwargs['max_new_tokens'] + inputs_embeds.shape[-2], 
    layer_device_map=layer_device_map)

(Following the device map specified by previous posts, for other device map for more gpus, edit the layer_device_map accordingly)

momianhua · 2024-10-17T00:47:13Z

Below is an example of running AIDC-AI/Ovis1.5-Gemma2-9B on two GPUs:

import torch
from PIL import Image
from transformers import AutoModelForCausalLM

device_map = {
    "visual_tokenizer": 0,
    "vte": 0,
    "llm.model.embed_tokens": 0,
    "llm.model.norm": 0,
    "llm.lm_head": 0,
    "llm.model.layers.0": 0,
    "llm.model.layers.1": 0,
    "llm.model.layers.2": 0,
    "llm.model.layers.3": 0,
    "llm.model.layers.4": 0,
    "llm.model.layers.5": 0,
    "llm.model.layers.6": 0,
    "llm.model.layers.7": 0,
    "llm.model.layers.8": 0,
    "llm.model.layers.9": 0,
    "llm.model.layers.10": 0,
    "llm.model.layers.11": 0,
    "llm.model.layers.12": 0,
    "llm.model.layers.13": 0,
    "llm.model.layers.14": 0,
    "llm.model.layers.15": 0,
    "llm.model.layers.16": 0,
    "llm.model.layers.17": 0,
    "llm.model.layers.18": 0,
    "llm.model.layers.19": 0,
    "llm.model.layers.20": 1,
    "llm.model.layers.21": 1,
    "llm.model.layers.22": 1,
    "llm.model.layers.23": 1,
    "llm.model.layers.24": 1,
    "llm.model.layers.25": 1,
    "llm.model.layers.26": 1,
    "llm.model.layers.27": 1,
    "llm.model.layers.28": 1,
    "llm.model.layers.29": 1,
    "llm.model.layers.30": 1,
    "llm.model.layers.31": 1,
    "llm.model.layers.32": 1,
    "llm.model.layers.33": 1,
    "llm.model.layers.34": 1,
    "llm.model.layers.35": 1,
    "llm.model.layers.36": 1,
    "llm.model.layers.37": 1,
    "llm.model.layers.38": 1,
    "llm.model.layers.39": 1,
    "llm.model.layers.40": 1,
    "llm.model.layers.41": 1
}

# load model
model = AutoModelForCausalLM.from_pretrained("AIDC-AI/Ovis1.5-Gemma2-9B",
                                             torch_dtype=torch.bfloat16,
                                             multimodal_max_length=8192,
					     device_map=device_map,
                                             trust_remote_code=True)
text_tokenizer = model.get_text_tokenizer()
visual_tokenizer = model.get_visual_tokenizer()
conversation_formatter = model.get_conversation_formatter()

# enter image path and prompt
image_path = input("Enter image path: ")
image = Image.open(image_path)
text = input("Enter prompt: ")
query = f'<image>\n{text}'
prompt, input_ids = conversation_formatter.format_query(query)
input_ids = torch.unsqueeze(input_ids, dim=0).cuda()
attention_mask = torch.ne(input_ids, text_tokenizer.pad_token_id).cuda()
pixel_values = [visual_tokenizer.preprocess_image(image).to(
    dtype=visual_tokenizer.dtype, device=visual_tokenizer.device)]

# generate output
with torch.inference_mode():
    gen_kwargs = dict(
        max_new_tokens=1024,
        do_sample=False,
        top_p=None,
        top_k=None,
        temperature=None,
        repetition_penalty=None,
        eos_token_id=model.generation_config.eos_token_id,
        pad_token_id=text_tokenizer.pad_token_id,
        use_cache=True
    )
    output_ids = model.generate(input_ids, pixel_values=pixel_values, attention_mask=attention_mask, **gen_kwargs)[0]
    output = text_tokenizer.decode(output_ids, skip_special_tokens=True)
    print(f'Output: {output}')

what's the script for Ovis1.6-Gemma2-9B？How do I deploy on two cards?

runninglsy · 2024-10-22T04:40:21Z

Below is an example of running AIDC-AI/Ovis1.5-Gemma2-9B on two GPUs:

import torch
from PIL import Image
from transformers import AutoModelForCausalLM

device_map = {
    "visual_tokenizer": 0,
    "vte": 0,
    "llm.model.embed_tokens": 0,
    "llm.model.norm": 0,
    "llm.lm_head": 0,
    "llm.model.layers.0": 0,
    "llm.model.layers.1": 0,
    "llm.model.layers.2": 0,
    "llm.model.layers.3": 0,
    "llm.model.layers.4": 0,
    "llm.model.layers.5": 0,
    "llm.model.layers.6": 0,
    "llm.model.layers.7": 0,
    "llm.model.layers.8": 0,
    "llm.model.layers.9": 0,
    "llm.model.layers.10": 0,
    "llm.model.layers.11": 0,
    "llm.model.layers.12": 0,
    "llm.model.layers.13": 0,
    "llm.model.layers.14": 0,
    "llm.model.layers.15": 0,
    "llm.model.layers.16": 0,
    "llm.model.layers.17": 0,
    "llm.model.layers.18": 0,
    "llm.model.layers.19": 0,
    "llm.model.layers.20": 1,
    "llm.model.layers.21": 1,
    "llm.model.layers.22": 1,
    "llm.model.layers.23": 1,
    "llm.model.layers.24": 1,
    "llm.model.layers.25": 1,
    "llm.model.layers.26": 1,
    "llm.model.layers.27": 1,
    "llm.model.layers.28": 1,
    "llm.model.layers.29": 1,
    "llm.model.layers.30": 1,
    "llm.model.layers.31": 1,
    "llm.model.layers.32": 1,
    "llm.model.layers.33": 1,
    "llm.model.layers.34": 1,
    "llm.model.layers.35": 1,
    "llm.model.layers.36": 1,
    "llm.model.layers.37": 1,
    "llm.model.layers.38": 1,
    "llm.model.layers.39": 1,
    "llm.model.layers.40": 1,
    "llm.model.layers.41": 1
}

# load model
model = AutoModelForCausalLM.from_pretrained("AIDC-AI/Ovis1.5-Gemma2-9B",
                                             torch_dtype=torch.bfloat16,
                                             multimodal_max_length=8192,
					     device_map=device_map,
                                             trust_remote_code=True)
text_tokenizer = model.get_text_tokenizer()
visual_tokenizer = model.get_visual_tokenizer()
conversation_formatter = model.get_conversation_formatter()

# enter image path and prompt
image_path = input("Enter image path: ")
image = Image.open(image_path)
text = input("Enter prompt: ")
query = f'<image>\n{text}'
prompt, input_ids = conversation_formatter.format_query(query)
input_ids = torch.unsqueeze(input_ids, dim=0).cuda()
attention_mask = torch.ne(input_ids, text_tokenizer.pad_token_id).cuda()
pixel_values = [visual_tokenizer.preprocess_image(image).to(
    dtype=visual_tokenizer.dtype, device=visual_tokenizer.device)]

# generate output
with torch.inference_mode():
    gen_kwargs = dict(
        max_new_tokens=1024,
        do_sample=False,
        top_p=None,
        top_k=None,
        temperature=None,
        repetition_penalty=None,
        eos_token_id=model.generation_config.eos_token_id,
        pad_token_id=text_tokenizer.pad_token_id,
        use_cache=True
    )
    output_ids = model.generate(input_ids, pixel_values=pixel_values, attention_mask=attention_mask, **gen_kwargs)[0]
    output = text_tokenizer.decode(output_ids, skip_special_tokens=True)
    print(f'Output: {output}')

what's the script for Ovis1.6-Gemma2-9B？How do I deploy on two cards?

The code is available at: https://huggingface.co/AIDC-AI/Ovis1.6-Gemma2-9B/discussions/6

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

How to use multiple GPUs for inference? #12

How to use multiple GPUs for inference? #12

waltonfuture commented Aug 23, 2024

runninglsy commented Sep 3, 2024

cyj95 commented Sep 11, 2024

AntyRia commented Sep 20, 2024

runninglsy commented Sep 23, 2024

CVHub520 commented Sep 23, 2024 •

edited

Loading

LuciusLan commented Sep 30, 2024

momianhua commented Oct 17, 2024

runninglsy commented Oct 22, 2024

How to use multiple GPUs for inference? #12

How to use multiple GPUs for inference? #12

Comments

waltonfuture commented Aug 23, 2024

runninglsy commented Sep 3, 2024

cyj95 commented Sep 11, 2024

AntyRia commented Sep 20, 2024

runninglsy commented Sep 23, 2024

CVHub520 commented Sep 23, 2024 • edited Loading

LuciusLan commented Sep 30, 2024

momianhua commented Oct 17, 2024

runninglsy commented Oct 22, 2024

CVHub520 commented Sep 23, 2024 •

edited

Loading