一文带你了解OpenAI Sora
Published:
This post explains the basics of OpenAI Sora, a new text-to-video technology developed by OpenAI. Check out the post for more details.
Posts by Tags
AIGC
Alignment
A Closer Look at ChatGPT
Published:
This post explains the details of the ChatGPT model and how it works. Check out the post for more details.
CLIP
DenseCLIP: Language-Guided Dense Prediction with Context-Aware Prompting
Published:
Paper reading and summary of DenseCLIP. Chech out the post for more details.
CoCoOp: Conditional Prompt Learning for Vision-Language Models
Published:
Paper reading and summary of CoCoOp. Chech out the post for more details.
CoOp: Learning to Prompt for Vision-Language Models
Published:
Paper reading and summary of CoOp. Chech out the post for more details.
CLIP: Learning Transferable Visual Models From Natural Language Supervision
Published:
Paper reading and summary of CLIP. Chech out the post for more details.
从MuZero到EfficientZero：Make MBRL Great Again
Published:
I recap the recent progress of MuZero and EfficientZero in the field of Model-Based Reinforcement Learning (MBRL). Chech out the post for more details.
ChatGPT
A Closer Look at ChatGPT
Published:
This post explains the details of the ChatGPT model and how it works. Check out the post for more details.
Experience Replay
Prioritized Experience Replay
Published:
Paper reading and summary of Prioritized Experience Replay (PER). Chech out the post for more details.
Foundation Model
Gato: A Generalist Agent (An RL Foundation Model)
Published:
Paper reading and summary of Gato. Chech out the post for more details.
Gato
Gato: A Generalist Agent (An RL Foundation Model)
Published:
Paper reading and summary of Gato. Chech out the post for more details.
Image Inpainting
Inpaint Anything: Segment Anything Meets Image Inpainting
Published:
We introduce Inpaint Anything (IA), a mask-free image inpainting system based on the Segment-Anything Model (SAM). IA offers a “clicking and filling” paradigm, combining different models to create a powerful, user-friendly pipeline for inpainting tasks. It supports three features: Remove Anything, Fill Anything, and Replace Anything, allowing users to remove objects, fill holes with generative content, or replace backgrounds with new scenes. Check out the post for more details!
Importance Sampling
【IS视角下的PPO】从Importance Sampling（IS）到PPO
Published:
I recap the Importance Sampling (IS) and Proximal Policy Optimization (PPO) algorithms in the context of RL. Chech out the post for more details.
Inpainting Anything
Inpaint Anything: Segment Anything Meets Image Inpainting
Published:
We introduce Inpaint Anything (IA), a mask-free image inpainting system based on the Segment-Anything Model (SAM). IA offers a “clicking and filling” paradigm, combining different models to create a powerful, user-friendly pipeline for inpainting tasks. It supports three features: Remove Anything, Fill Anything, and Replace Anything, allowing users to remove objects, fill holes with generative content, or replace backgrounds with new scenes. Check out the post for more details!
InstructGPT
A Closer Look at ChatGPT
Published:
This post explains the details of the ChatGPT model and how it works. Check out the post for more details.
LLM
A Closer Look at ChatGPT
Published:
This post explains the details of the ChatGPT model and how it works. Check out the post for more details.
Offline RL
【Offline RL】再读BCQ
Published:
Paper reading and summary of BCQ. Chech out the post for more details.
PPO
【IS视角下的PPO】从Importance Sampling（IS）到PPO
Published:
I recap the Importance Sampling (IS) and Proximal Policy Optimization (PPO) algorithms in the context of RL. Chech out the post for more details.
Parameter-Efficient Tuning
CoCoOp: Conditional Prompt Learning for Vision-Language Models
Published:
Paper reading and summary of CoCoOp. Chech out the post for more details.
CoOp: Learning to Prompt for Vision-Language Models
Published:
Paper reading and summary of CoOp. Chech out the post for more details.
Policy Gradient
光速推导REINFORCE
Published:
Paper reading and summary of REINFORCE. Chech out the post for more details.
Prompt Learning
DenseCLIP: Language-Guided Dense Prediction with Context-Aware Prompting
Published:
Paper reading and summary of DenseCLIP. Chech out the post for more details.
CoCoOp: Conditional Prompt Learning for Vision-Language Models
Published:
Paper reading and summary of CoCoOp. Chech out the post for more details.
CoOp: Learning to Prompt for Vision-Language Models
Published:
Paper reading and summary of CoOp. Chech out the post for more details.
RL
Gato: A Generalist Agent (An RL Foundation Model)
Published:
Paper reading and summary of Gato. Chech out the post for more details.
Prioritized Experience Replay
Published:
Paper reading and summary of Prioritized Experience Replay (PER). Chech out the post for more details.
【Offline RL】再读BCQ
Published:
Paper reading and summary of BCQ. Chech out the post for more details.
RL Transformer之Decision Transformer
Published:
Paper reading and summary of Decision Transformer. Chech out the post for more details.
【IS视角下的PPO】从Importance Sampling（IS）到PPO
Published:
I recap the Importance Sampling (IS) and Proximal Policy Optimization (PPO) algorithms in the context of RL. Chech out the post for more details.
光速推导REINFORCE
Published:
Paper reading and summary of REINFORCE. Chech out the post for more details.
RL Transformer之Trajectory Transformers
Published:
Paper reading and summary of Trajectory Transformers. Chech out the post for more details.
SAM
Inpaint Anything: Segment Anything Meets Image Inpainting
Published:
We introduce Inpaint Anything (IA), a mask-free image inpainting system based on the Segment-Anything Model (SAM). IA offers a “clicking and filling” paradigm, combining different models to create a powerful, user-friendly pipeline for inpainting tasks. It supports three features: Remove Anything, Fill Anything, and Replace Anything, allowing users to remove objects, fill holes with generative content, or replace backgrounds with new scenes. Check out the post for more details!
Segmentation
DenseCLIP: Language-Guided Dense Prediction with Context-Aware Prompting
Published:
Paper reading and summary of DenseCLIP. Chech out the post for more details.
Sora
一文带你了解OpenAI Sora
Published:
This post explains the basics of OpenAI Sora, a new text-to-video technology developed by OpenAI. Check out the post for more details.
Text-to-Video
一文带你了解OpenAI Sora
Published:
This post explains the basics of OpenAI Sora, a new text-to-video technology developed by OpenAI. Check out the post for more details.
Transformer
RL Transformer之Decision Transformer
Published:
Paper reading and summary of Decision Transformer. Chech out the post for more details.
RL Transformer之Trajectory Transformers
Published:
Paper reading and summary of Trajectory Transformers. Chech out the post for more details.
VLM
DenseCLIP: Language-Guided Dense Prediction with Context-Aware Prompting
Published:
Paper reading and summary of DenseCLIP. Chech out the post for more details.
CoCoOp: Conditional Prompt Learning for Vision-Language Models
Published:
Paper reading and summary of CoCoOp. Chech out the post for more details.
CoOp: Learning to Prompt for Vision-Language Models
Published:
Paper reading and summary of CoOp. Chech out the post for more details.
CLIP: Learning Transferable Visual Models From Natural Language Supervision
Published:
Paper reading and summary of CLIP. Chech out the post for more details.
从MuZero到EfficientZero：Make MBRL Great Again
Published:
I recap the recent progress of MuZero and EfficientZero in the field of Model-Based Reinforcement Learning (MBRL). Chech out the post for more details.