Pinned Repositories
awesome-algorithm
代码模板、脑图、Top100高频LeetCode题 - 极简实用的算法和数据结构（持续更新中..）
bert_crf_sequence_annotation
基于Pytorch+BERT+CRF的NLP序列标注模型，目前包括分词，词性标注，命名实体识别等
BERTweet
BERTweet: A pre-trained language model for English Tweets (EMNLP-2020)
KBQA-BERT
基于知识图谱的问答系统，BERT做命名实体识别和句子相似度，分为online和outline模式
learn-nlp-with-transformers
we want to create a repo to illustrate usage of transformers in chinese
leetcode
LeetCode Solutions: A Record of My Problem Solving Journey.( leetcode题解，记录自己的leetcode解题之路。)
nlp-competitions-list-review
复盘所有NLP比赛的TOP方案，只关注NLP比赛，持续更新中！
PaddleNLP
An NLP library with Awesome pre-trained Transformer models and easy-to-use interface, supporting wide-range of NLP tasks from research to industrial applications.
Paddlenlp-learning
Record for learning PaddleNLP
Research-on-turbulent-combustion-model
Research on turbulent combustion model based on back propagation neural network and convex optimization theory
Grant-Rk's Repositories
Grant-Rk/Paddlenlp-learning
Record for learning PaddleNLP
Grant-Rk/Research-on-turbulent-combustion-model
Research on turbulent combustion model based on back propagation neural network and convex optimization theory
Grant-Rk/awesome-algorithm
代码模板、脑图、Top100高频LeetCode题 - 极简实用的算法和数据结构（持续更新中..）
Grant-Rk/bert_crf_sequence_annotation
基于Pytorch+BERT+CRF的NLP序列标注模型，目前包括分词，词性标注，命名实体识别等
Grant-Rk/BERTweet
BERTweet: A pre-trained language model for English Tweets (EMNLP-2020)
Grant-Rk/KBQA-BERT
基于知识图谱的问答系统，BERT做命名实体识别和句子相似度，分为online和outline模式
Grant-Rk/learn-nlp-with-transformers
we want to create a repo to illustrate usage of transformers in chinese
Grant-Rk/leetcode
LeetCode Solutions: A Record of My Problem Solving Journey.( leetcode题解，记录自己的leetcode解题之路。)
Grant-Rk/nlp-competitions-list-review
复盘所有NLP比赛的TOP方案，只关注NLP比赛，持续更新中！
Grant-Rk/PaddleNLP
An NLP library with Awesome pre-trained Transformer models and easy-to-use interface, supporting wide-range of NLP tasks from research to industrial applications.
Grant-Rk/python-girlfriend-mood
:kissing_closed_eyes: Analyze her mood through her girlfriend's words ·女朋友聊天时的情绪波动图谱，没女朋友的可自用(雾）。
Grant-Rk/QASystemOnHepatopathyKG
基于neo4j肝病知识图谱的问答系统
Grant-Rk/Target-Offer
根据《剑指offer》中的题目，用python来进行编写
Grant-Rk/twitter_nlp
Twitter NLP Tools
Grant-Rk/WoBERT
以词为基本单位的中文BERT
Grant-Rk/bio-learning
bio-learning for single cell
Grant-Rk/Deep-Live-Cam
real time face swap and one-click video deepfake with only a single image (uncensored)
Grant-Rk/dgl
Python package built to ease deep learning on graph, on top of existing DL frameworks.
Grant-Rk/Firefly
Firefly(流萤): 中文对话式大语言模型(全量微调+QLoRA)，支持微调Aquila2、Baichuan2、CodeLlama、Llma2、Llama、Qwen、Baichuan、ChatGLM2、InternLM、Ziya、Bloom等大模型
Grant-Rk/Free-Chatgpt4
free chatgpt4
Grant-Rk/MedicalGPT
MedicalGPT: Training Your Own Medical GPT Model with ChatGPT Training Pipeline. 训练医疗大模型，实现包括二次预训练、有监督微调、奖励建模、强化学习训练。
Grant-Rk/Medusa
Medusa: Simple Framework for Accelerating LLM Generation with Multiple Decoding Heads
Grant-Rk/Megatron-LM
Ongoing research training transformer models at scale
Grant-Rk/mmf
A modular framework for vision & language multimodal research from Facebook AI Research (FAIR)
Grant-Rk/Open-sharing-of-CRF
Open sharing of CRF
Grant-Rk/Research
novel deep learning research works with PaddlePaddle
Grant-Rk/scFormer
Grant-Rk/sincere
Code for "Let's Stop Incorrect Comparisons in End-to-end Relation Extraction!", EMNLP 2020
Grant-Rk/transformers-benchmarks
Grant-Rk/UIE-ACL-310
有一个通用实体关系事件抽取的任务，需要使用到UIE模框架，而且需要将起部署到昇腾310服务器上，因为UIE模型底层使用的是ernie3.0，但是目前paddle官方还不支持ernie3.0模型在昇腾310上部署，所以才有了以下的操作，主要过程是，先试用paddle训练处模型，然后使用 paddle2onnx.command.c_paddle_to_onnx方法将paddle的模型转为onnx模型 ，因现在的onnx模型是动态的shape和散乱的算子形态，需要使用paddle自带的工具paddle2onnx.optimize将onnx模型先进行重塑，固定好shape的维度，将散乱的算子进行整合，命令如下： $ python -m paddle2onnx.optimize --input_model /home/user/lijiaqi/PaddleNLP/model_zoo/uie/export_new/model.onnx --output_model /home/user/lijiaqi/model_new_uie.onnx --input_shape_dict "{'att_mask':[1,512],'pos_ids':[1,512],'token_type_ids':[1,512],'input_ids':[1,512]}" 然后将onnx模型在使用ATC工具转为acl所需要的om模型，这一步后面会讲。 另外在使用acl部署的时候，paddle框架是不能使用的，acl使用到的模型和训练过程均需要自己实现，包括from_pretrain阶段的分词，建立词表，数据处理部分，这部分我已经实现完，纯python版本的实现