Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

VideoGPT+ Inference code, Simple Demo on Google Colab #18

Open
wants to merge 1 commit into
base: main
Choose a base branch
from
Open

VideoGPT+ Inference code, Simple Demo on Google Colab #18

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
279 changes: 279 additions & 0 deletions VideoGPT+_Demo.ipynb
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,279 @@
{
"nbformat": 4,
"nbformat_minor": 0,
"metadata": {
"colab": {
"provenance": [],
"machine_shape": "hm",
"gpuType": "A100",
"include_colab_link": true
},
"kernelspec": {
"name": "python3",
"display_name": "Python 3"
},
"language_info": {
"name": "python"
},
"accelerator": "GPU"
},
"cells": [
{
"cell_type": "markdown",
"metadata": {
"id": "view-in-github",
"colab_type": "text"
},
"source": [
"<a href=\"https://colab.research.google.com/github/Yogesh914/VideoGPT-plus/blob/main/VideoGPT%2B_Demo.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
]
},
{
"cell_type": "markdown",
"source": [
"# VideoGPT+ Demo 🎥"
],
"metadata": {
"id": "CP7--7iTuciZ"
}
},
{
"cell_type": "markdown",
"source": [
"## Setup"
],
"metadata": {
"id": "Wcomhj6fukGQ"
}
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"id": "D_h0ivDItx21"
},
"outputs": [],
"source": [
"!git clone https://github.com/mbzuai-oryx/VideoGPT-plus"
]
},
{
"cell_type": "code",
"source": [
"%cd VideoGPT-plus"
],
"metadata": {
"id": "Ds8NftGDyBBW"
},
"execution_count": null,
"outputs": []
},
{
"cell_type": "code",
"source": [
"!pip install huggingface_hub\n",
"\n",
"from huggingface_hub import notebook_login\n",
"notebook_login()"
],
"metadata": {
"id": "J02UHN8WvU4g"
},
"execution_count": null,
"outputs": []
},
{
"cell_type": "code",
"source": [
"!git clone https://huggingface.co/OpenGVLab/InternVideo2-Stage2_1B-224p-f4"
],
"metadata": {
"id": "I7qQvagisjQe"
},
"execution_count": null,
"outputs": []
},
{
"cell_type": "code",
"source": [
"!mkdir OpenGVLab\n",
"!mv InternVideo2-Stage2_1B-224p-f4 OpenGVLab/"
],
"metadata": {
"id": "TL5-SUvt1Ih7"
},
"execution_count": null,
"outputs": []
},
{
"cell_type": "code",
"source": [
"!git clone https://huggingface.co/MBZUAI/VideoGPT-plus_Phi3-mini-4k"
],
"metadata": {
"id": "n1IEfAC8uMN3"
},
"execution_count": null,
"outputs": []
},
{
"cell_type": "code",
"source": [
"!pip install shortuuid timm einops flash-attn decord mmengine ninja peft"
],
"metadata": {
"id": "S_UQhfdEuw4E"
},
"execution_count": null,
"outputs": []
},
{
"cell_type": "markdown",
"source": [
"## Inference"
],
"metadata": {
"id": "kXx6QFlUumnQ"
}
},
{
"cell_type": "code",
"source": [
"from tqdm import tqdm\n",
"from videogpt_plus.conversation import conv_templates\n",
"from videogpt_plus.model.builder import load_pretrained_model\n",
"from videogpt_plus.mm_utils import tokenizer_image_token, get_model_name_from_path\n",
"from eval.vcgbench.inference.ddp import *\n",
"from eval.video_encoding import _get_rawvideo_dec\n",
"import traceback\n",
"import pandas as pd\n",
"import torch\n",
"import os\n",
"\n",
"# Disable parameter resetting for specific layers\n",
"setattr(torch.nn.Linear, \"reset_parameters\", lambda self: None)\n",
"setattr(torch.nn.LayerNorm, \"reset_parameters\", lambda self: None)\n",
"\n",
"# Configuration\n",
"model_path = \"/content/VideoGPT-plus/VideoGPT-plus_Phi3-mini-4k/vcgbench\"\n",
"model_base = \"microsoft/Phi-3-mini-4k-instruct\"\n",
"video_path = \"path/to/video\"\n",
"qs = \"PROMPT\"\n",
"\n",
"conv_mode = \"phi3_instruct\"\n",
"stop_str = \"<|end|>\"\n",
"temperature = 0.0\n",
"\n",
"# Load pretrained model and tokenizer\n",
"model_path = os.path.expanduser(model_path)\n",
"model_name = get_model_name_from_path(model_path)\n",
"tokenizer, model, image_processor, context_len = load_pretrained_model(\n",
" model_path, model_base, model_name\n",
")\n",
"\n",
"# Configure model for multimodal use\n",
"mm_use_im_start_end = getattr(model.config, \"mm_use_im_start_end\", False)\n",
"mm_use_im_patch_token = getattr(model.config, \"mm_use_im_patch_token\", True)\n",
"if mm_use_im_patch_token:\n",
" tokenizer.add_tokens([DEFAULT_IMAGE_PATCH_TOKEN], special_tokens=True)\n",
"if mm_use_im_start_end:\n",
" tokenizer.add_tokens(\n",
" [DEFAULT_IM_START_TOKEN, DEFAULT_IM_END_TOKEN], special_tokens=True\n",
" )\n",
"model.resize_token_embeddings(len(tokenizer))\n",
"\n",
"# Load vision towers\n",
"vision_tower = model.get_vision_tower()\n",
"vision_tower.load_model(model.config.mm_vision_tower)\n",
"video_processor = vision_tower.image_processor\n",
"\n",
"image_vision_tower = model.get_image_vision_tower()\n",
"image_vision_tower.load_model()\n",
"image_processor = image_vision_tower.image_processor\n",
"\n",
"# Move model to GPU\n",
"model = model.to(\"cuda\")\n",
"\n",
"# Process video if it exists\n",
"if os.path.exists(video_path):\n",
" video_frames, context_frames, slice_len = _get_rawvideo_dec(\n",
" video_path,\n",
" image_processor,\n",
" video_processor,\n",
" max_frames=NUM_FRAMES,\n",
" image_resolution=224,\n",
" num_video_frames=NUM_FRAMES,\n",
" num_context_images=NUM_CONTEXT_IMAGES,\n",
" )\n",
"\n",
"# Prepare query string with image tokens\n",
"if model.config.mm_use_im_start_end:\n",
" qs = (\n",
" DEFAULT_IM_START_TOKEN\n",
" + DEFAULT_IMAGE_TOKEN * slice_len\n",
" + DEFAULT_IM_END_TOKEN\n",
" + \"\\n\"\n",
" + qs\n",
" )\n",
"else:\n",
" qs = DEFAULT_IMAGE_TOKEN * slice_len + \"\\n\" + qs\n",
"\n",
"# Create conversation prompt\n",
"conv = conv_templates[conv_mode].copy()\n",
"conv.append_message(conv.roles[0], qs)\n",
"conv.append_message(conv.roles[1], None)\n",
"prompt = conv.get_prompt()\n",
"\n",
"input_ids = (\n",
" tokenizer_image_token(\n",
" prompt, tokenizer, IMAGE_TOKEN_INDEX, return_tensors=\"pt\"\n",
" )\n",
" .unsqueeze(0)\n",
" .cuda()\n",
")\n",
"\n",
"# Generate output\n",
"with torch.inference_mode():\n",
" output_ids = model.generate(\n",
" input_ids,\n",
" images=torch.stack(video_frames).half().cuda(),\n",
" context_images=torch.stack(context_frames).half().cuda(),\n",
" do_sample=True if temperature > 0 else False,\n",
" temperature=temperature,\n",
" top_p=None,\n",
" num_beams=1,\n",
" max_new_tokens=1024,\n",
" use_cache=True,\n",
" )\n",
"\n",
"# Validate output\n",
"input_token_len = input_ids.shape[1]\n",
"n_diff_input_output = (\n",
" (input_ids != output_ids[:, :input_token_len]).sum().item()\n",
")\n",
"if n_diff_input_output > 0:\n",
" print(\n",
" f\"[Warning] {n_diff_input_output} output_ids are not the same as the input_ids\"\n",
" )\n",
"\n",
"# Decode and clean up output\n",
"outputs = tokenizer.batch_decode(\n",
" output_ids[:, input_token_len:], skip_special_tokens=True\n",
")[0]\n",
"outputs = outputs.strip()\n",
"if outputs.endswith(stop_str):\n",
" outputs = outputs[: -len(stop_str)]\n",
"outputs = outputs.strip()\n",
"outputs = outputs.replace(\"<|end|>\", \"\")\n",
"outputs = outputs.strip()\n",
"\n",
"print(outputs)"
],
"metadata": {
"id": "OUIabt9fuB3_"
},
"execution_count": null,
"outputs": []
}
]
}