Personal space - OSCHINA - Chinese open source technology exchange community

HelloWorld # Import the Gradio library, which is a Python library used to create a Web interface Import Gradio as gr # Define a greeting function that accepts two parameters: # name: the name entered by the user # intensity: the number of exclamation marks (from the slider

Gradio interface assembly demonstration Blocks Event Listener State management

Daily records

03/31 09:55

seven hundred and eighty-seven

zero

LLM's development process The development process of large-scale language models can be roughly divided into four stages: pre training (base model), supervised fine-tuning (SFT), incentive modeling (RM model, the evaluation model for fine-tuning models), and reinforcement learning (RL model). Each stage is aimed at specific goals ..

LLM development process Pre training Supervised fine adjustment Reward modeling Reinforcement learning Dialogue model

Daily records

03/23 07:17

four hundred and forty-four

zero

We use Q (S, A) to represent the value action function. In the case of a limited number of states and actions, we are actually equivalent to maintaining a Q-table ..

Value function estimation Status and action Q-table Markov process Q-learning algorithm Sarsa algorithm

Daily records

03/14 06:27

1.7K

zero

Understanding the Three Perspectives of Marketing: Customer Value Creating Enterprise Business Philosophy Enterprise implementation of process marketing is a complete process for enterprises to understand customers, create customer value, disseminate customer value and deliver customer value. Cognitive customer value market research User interviews User observations found that Gu ..

Customer value creation Business philosophy Enterprise execution process market research User interviews User observation

Daily records

02/20 17:40

four hundred and seventy-eight

zero

Connect AI agent (V) Dify Dify is an open source agent development platform. There are two ways to use Dify, one is to use Dify's online platform. https://cloud.dify.ai . One is privatization deployment. Download address: https://github.com/lang...

AI agent Dify Platform Open Source Privatization deployment Workflow LLM model

Daily records

01/05 04:03

three hundred and fifty-one

zero

Data analysis Data usually refers to an article, an actor list, and a comment. Human sophistication extracts relationships from massive data. Relationships and entities form relationships. Fig. Neo4j is generally selected for database installation. Download address: https://we-yun....

neo4j jdk knowledge graph https community- helloworld unix java browser python Figure Database jdk8

Daily records

2024/12/26 04:26

three hundred and three

zero

Connected to AI's mathematical theoretical foundation, Matrix multiplication in Transformer's relevant self attention mechanism In Tensorflow's in-depth learning algorithm (II), we know that word2vec is trained, so we can get a preliminary word vector, which we call embeddi

k v cos token Feedforward neural network Deep learning openai Backpropagation seq2seq tensorflow word2vec overp

Daily records

2024/11/10 20:00

seven hundred and forty-one

zero

When AI agent (III) memory tool was used in the past, we directly interacted with the large model, and there was no way to systematically realize memory. Langchain provides a system level solution. As shown in the figure above, if the user asks a question, the system will first store the information in the memory (as shown in the figure above

memory langchain zero shot react chain Vector database invoke openai https prompt tools docs plan

Daily records

2024/11/08 17:29

1.2K

zero

Connect the multi-mode big model chapter Stable Diffusion code practice chapter to create a conda environment. Here is the source code download address of conda create - n diffusion -- clone pythoch conda activate diffusion: https://github.com/pesser/s...

Stable-Diffusion Self encoder LDM model CLIP model ControlNet IP-Adapter

Daily records

2024/10/17 17:21

1.3K

zero

After statistics (II) Analysis of variance (ANOVA) t-test vs analysis of variance t-test Comparison of mean values of two groups (H_0: μ _1=μ _2 ) (H_A: μ _1 ≠ μ _2 ) Analysis of variance Comparison of mean values of three groups or more (H_0: μ _1=μ _2=