File size: 12,950 Bytes
3f7cfab |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 |
### Code to consider including:
[flan-alpaca](https://github.com/declare-lab/flan-alpaca)<br />
[text-generation-webui](https://github.com/oobabooga/text-generation-webui)<br />
[minimal-llama](https://github.com/zphang/minimal-llama/)<br />
[finetune GPT-NeoX](https://nn.labml.ai/neox/samples/finetune.html)<br />
[GPTQ-for_LLaMa](https://github.com/qwopqwop200/GPTQ-for-LLaMa/compare/cuda...Digitous:GPTQ-for-GPT-NeoX:main)<br />
[OpenChatKit on multi-GPU](https://github.com/togethercomputer/OpenChatKit/issues/20)<br />
[Non-Causal LLM](https://huggingface.co/docs/transformers/main/en/model_doc/gptj#transformers.GPTJForSequenceClassification)<br />
[OpenChatKit_Offload](https://github.com/togethercomputer/OpenChatKit/commit/148b5745a57a6059231178c41859ecb09164c157)<br />
[Flan-alpaca](https://github.com/declare-lab/flan-alpaca/blob/main/training.py)<br />
### Some open source models:
[GPT-NeoXT-Chat-Base-20B](https://huggingface.co/togethercomputer/GPT-NeoXT-Chat-Base-20B/tree/main)<br />
[GPT-NeoX](https://huggingface.co/docs/transformers/model_doc/gpt_neox)<br />
[GPT-NeoX-20B](https://huggingface.co/EleutherAI/gpt-neox-20b)<br />
[Pythia-6.9B](https://huggingface.co/EleutherAI/pythia-6.9b)<br />
[Pythia-12B](https://huggingface.co/EleutherAI/neox-ckpt-pythia-12b)<br />
[Flan-T5-XXL](https://huggingface.co/google/flan-t5-xxl)<br />
[GPT-J-Moderation-6B](https://huggingface.co/togethercomputer/GPT-JT-Moderation-6B)<br />
[OIG safety models](https://laion.ai/blog/oig-dataset/#safety-models)<br />
[BigScience-mT0](https://huggingface.co/mT0)<br />
[BigScience-XP3](https://huggingface.co/datasets/bigscience/xP3)<br />
[BigScience-Bloomz](https://huggingface.co/bigscience/bloomz)<br />
### Some create commons models that would be interesting to use:
[Galactica-120B](https://huggingface.co/facebook/galactica-120b)<br />
[LLaMa-small-pt](https://huggingface.co/decapoda-research/llama-smallint-pt)<br />
[LLaMa-64b-4bit](https://huggingface.co/maderix/llama-65b-4bit/tree/main)<br />
### Papers/Repos
[Self-improve](https://arxiv.org/abs/2210.11610)<br />
[Coding](https://arxiv.org/abs/2303.17491)<br />
[self-reflection](https://arxiv.org/abs/2303.11366)<br />
[RLHF](https://arxiv.org/abs/2204.05862)<br />
[DERA](https://arxiv.org/abs/2303.17071)<br />
[HAI Index Report 2023](https://aiindex.stanford.edu/report/)<br />
[LLaMa](https://arxiv.org/abs/2302.13971)<br />
[GLM-130B](https://github.com/THUDM/GLM-130B)<br />
[RWKV RNN](https://github.com/BlinkDL/RWKV-LM)<br />
[Toolformer](https://arxiv.org/abs/2302.04761)<br />
[GPTQ](https://github.com/qwopqwop200/GPTQ-for-LLaMa)<br />
[Retro](https://www.deepmind.com/publications/improving-language-models-by-retrieving-from-trillions-of-tokens)<br />
[Clinical_outperforms](https://arxiv.org/abs/2302.08091)<br />
[Chain-Of-Thought](https://github.com/amazon-science/mm-cot)<br />
[scaling law1](https://arxiv.org/abs/2203.15556)<br />
[Big-bench](https://github.com/google/BIG-bench)<br />
[Natural-Instructions](https://github.com/allenai/natural-instructions)<br />
### Other projects:
[StackLLaMa](https://huggingface.co/blog/stackllama)<br />
[Alpaca-CoT](https://github.com/PhoebusSi/alpaca-CoT)<br />
[ColossalAIChat](https://github.com/hpcaitech/ColossalAI/tree/main/applications/Chat)<br />
[EasyLM](https://github.com/young-geng/EasyLM.git)<br />
[Koala](https://bair.berkeley.edu/blog/2023/04/03/koala/)<br />
[Vicuna](https://vicuna.lmsys.org/)<br />
[Flan-Alpaca](https://github.com/declare-lab/flan-alpaca)<br />
[FastChat](https://chat.lmsys.org/)<br />
[alpaca-lora](https://github.com/h2oai/alpaca-lora)<br />
[alpaca.http](https://github.com/Nuked88/alpaca.http)<br />
[chatgpt-retrieval-pllugin](https://github.com/openai/chatgpt-retrieval-plugin)<br />
[subtl.ai docs search on private docs](https://www.subtl.ai/)<br />
[gertel](https://gretel.ai/)<br />
[alpaca_lora_4bit](https://github.com/johnsmith0031/alpaca_lora_4bit)<br />
[alpaca_lora_4bit_readme](https://github.com/s4rduk4r/alpaca_lora_4bit_readme)<br />
[code alpaca](https://github.com/sahil280114/codealpaca)<br />
[serge](https://github.com/nsarrazin/serge)<br />
[BlinkDL](https://huggingface.co/spaces/BlinkDL/ChatRWKV-gradio)<br />
[RWKV-LM](https://github.com/BlinkDL/RWKV-LM)<br />
[MosaicCM](https://github.com/mosaicml/examples#large-language-models-llms)<br />
[OpenAI Plugins](https://openai.com/blog/chatgpt-plugins)<br />
[GPT3.5-Turbo-PGVector](https://github.com/gannonh/gpt3.5-turbo-pgvector)<br />
[LLaMa-Adapter](https://github.com/ZrrSkywalker/LLaMA-Adapter)<br />
[llama-index](https://github.com/jerryjliu/llama_index)<br />
[minimal-llama](https://github.com/zphang/minimal-llama/)<br />
[llama.cpp](https://github.com/ggerganov/llama.cpp)<br />
[ggml](https://github.com/ggerganov/ggml)<br />
[mmap](https://justine.lol/mmap/)<br />
[lamma.cpp more](https://til.simonwillison.net/llms/llama-7b-m2)<br />
[TargetedSummarization](https://github.com/helliun/targetedSummarization)<br />
[OpenFlamingo](https://laion.ai/blog/open-flamingo/)<br />
[Auto-GPT](https://github.com/Torantulino/Auto-GPT)<br />
### Apache2/etc. Data
[OIG 43M instructions](https://laion.ai/blog/oig-dataset/) [direct HF link](https://huggingface.co/datasets/laion/OIG)<br />
[More on OIG](https://laion.ai/blog/oig-dataset/)<br />
[DataSet Viewer](https://huggingface.co/datasets/viewer/?dataset=squad)<br />
[Anthropic RLHF](https://huggingface.co/datasets/Anthropic/hh-rlhf)<br />
[WebGPT_Comparisons](https://huggingface.co/datasets/openai/webgpt_comparisons)<br />
[Self_instruct](https://github.com/yizhongw/self-instruct)<br />
[20BChatModelData](https://github.com/togethercomputer/OpenDataHub)<br />
### Apache2/MIT/BSD-3 Summarization Data
[xsum for Summarization](https://huggingface.co/datasets/xsum)<br />
[Apache2 Summarization](https://huggingface.co/datasets?task_categories=task_categories:summarization&license=license:apache-2.0&sort=downloads)<br />
[MIT summarization](https://huggingface.co/datasets?task_categories=task_categories:summarization&license=license:mit&sort=downloads)<br />
[BSD-3 summarization](https://huggingface.co/datasets?task_categories=task_categories:summarization&license=license:bsd-3-clause&sort=downloads)<br />
[OpenRail](https://huggingface.co/datasets?task_categories=task_categories:summarization&license=license:openrail&sort=downloads)<br />
[Summarize_from_feedback](https://huggingface.co/datasets/openai/summarize_from_feedback)<br />
### Ambiguous License Data
[GPT-4-LLM](https://github.com/Instruction-Tuning-with-GPT-4/GPT-4-LLM)<br />
[GPT4All](https://huggingface.co/datasets/nomic-ai/gpt4all_prompt_generations)<br />
[LinkGPT4](https://github.com/lm-sys/FastChat/issues/90#issuecomment-1493250773)<br />
[ShareGPT52K](https://huggingface.co/datasets/RyokoAI/ShareGPT52K)<br />
[ShareGPT_Vicuna](https://huggingface.co/datasets/anon8231489123/ShareGPT_Vicuna_unfiltered)<br />
[ChatLogs](https://chatlogs.net/)<br />
[Alpaca-CoT](https://github.com/PhoebusSi/alpaca-CoT)<br />
[LaMini-LM](https://github.com/mbzuai-nlp/LaMini-LM)<br />
### Non-commercial Data
[GPT-3 based Alpaca Cleaned](https://github.com/gururise/AlpacaDataCleaned)<br />
### Prompt ENGR
[Prompt/P-tuning](https://github.com/huggingface/peft)<br />
[Prompt/P-tuing Nemo/NVIDIA](https://docs.nvidia.com/deeplearning/nemo/user-guide/docs/en/main/nlp/nemo_megatron/prompt_learning.html)<br />
[Info](https://lilianweng.github.io/posts/2023-03-15-prompt-engineering/)<br />
[Info2](https://github.com/dair-ai/Prompt-Engineering-Guide)<br />
[Prompt-Tuning](https://arxiv.org/abs/2104.08691)<br />
[P-tuning v2](https://arxiv.org/abs/2110.07602)<br />
[babyagi](https://github.com/yoheinakajima/babyagi/blob/main/babyagi.py#L97-L134)<br />
[APE](https://www.promptingguide.ai/techniques/ape)<br />
### Validation
[Bleu/Rouge/Meteor/Bert-Score](https://arize.com/blog-course/generative-ai-metrics-bleu-score/)<br />
### Generate Hyperparameters
[hot-to-generate](https://huggingface.co/blog/how-to-generate)<br />
[Notes_on_Transformers Chpt5](https://christianjmills.com/posts/transformers-book-notes/chapter-5/index.html)<br />
[Notes_on_Transformers_Chpt10](https://christianjmills.com/posts/transformers-book-notes/chapter-10/index.html)<br />
### Embeddings
[OpenAI Expensive?](https://medium.com/@nils_reimers/openai-gpt-3-text-embeddings-really-a-new-state-of-the-art-in-dense-text-embeddings-6571fe3ec9d9)<br />
[Leaderboard](https://huggingface.co/spaces/mteb/leaderboard)<br />
### Commercial products
[OpenAI](https://platform.openai.com/docs/guides/fine-tuning/advanced-usage)<br />
[OpenAI Tokenizer](https://platform.openai.com/tokenizer)<br />
[OpenAI Playground](https://platform.openai.com/playground)<br />
[OpenAI Chat](https://chat.openai.com/chat?)<br />
[OpenAI GPT-4 Chat](https://chat.openai.com/chat?model=gpt-4)<br />
[cohere](https://cohere.io/)<br />
[coherefinetune](https://docs.cohere.ai/reference/finetune)<br />
[DocsBotAI](https://docsbot.ai/)<br />
[Perplexity](https://www.perplexity.ai/)<br />
[VoiceFlow](https://www.voiceflow.com/)<br />
[NLPCloud](https://nlpcloud.com/effectively-using-gpt-j-gpt-neo-gpt-3-alternatives-few-shot-learning.html)<br />
### Multinode inference
[FasterTransformer](https://github.com/triton-inference-server/fastertransformer_backend#multi-node-inference)<br />
[Kubernetes Triton](https://developer.nvidia.com/blog/deploying-nvidia-triton-at-scale-with-mig-and-kubernetes/)<br />
### Faster inference
[text-generation-inference](https://github.com/huggingface/text-generation-inference)<br />
[Optimum](https://github.com/huggingface/optimum)<br />
### Semi-Open source Semi-Commercial products
[OpenAssistant](https://open-assistant.io/)<br />
[OpenAssistant Repo](https://github.com/LAION-AI/Open-Assistant)<br />
[OpenChatKit](https://github.com/togethercomputer/OpenChatKit)<br />
[OpenChatKit2](https://github.com/togethercomputer/OpenDataHub)<br />
[OpenChatKit3](https://www.together.xyz/blog/openchatkit)<br />
[OpenChatKit4](https://github.com/togethercomputer/OpenChatKit/blob/main/training/README.md#arguments)<br />
[OpenChatKitPreview](https://api.together.xyz/open-chat?preview=1)<br />
[langchain](https://python.langchain.com/en/latest/)<br />
[langchain+pinecone](https://www.youtube.com/watch?v=nMniwlGyX-c)<br />
### Q/A docs
[HUMATA](https://www.humata.ai/)<br />
[OSSCHat](https://osschat.io/)<br />
[NeuralSearchCohere](https://txt.cohere.com/embedding-archives-wikipedia/)<br />
[ue5](https://github.com/bublint/ue5-llama-lora)<br />
### AutoGPT type projects
[AgentGPT](https://github.com/reworkd/AgentGPT)<br />
[Self-DEBUG](https://arxiv.org/abs/2304.05128)<br />
[BabyAGI](https://github.com/yoheinakajima/babyagi/)<br />
[AutoPR](https://github.com/irgolic/AutoPR)<br />
### Cloud fine-tune
[AWS](https://docs.aws.amazon.com/sagemaker/latest/dg/jumpstart-fine-tune.html)<br />
[AWS2](https://aws.amazon.com/blogs/machine-learning/training-large-language-models-on-amazon-sagemaker-best-practices/)<br />
### Chatbots:
[GPT4ALL Chat](https://github.com/nomic-ai/gpt4all-chat)<br />
[GLT4ALL](https://github.com/nomic-ai/gpt4all)<br />
[OASSST](https://open-assistant.io/chat)<br />
[FastChat](https://github.com/lm-sys/FastChat)<br />
[Dolly](https://huggingface.co/spaces/HuggingFaceH4/databricks-dolly)<br />
[HF Instructions](https://huggingface.co/spaces/HuggingFaceH4/instruction-model-outputs-filtered)<br />
[DeepSpeed Chat](https://github.com/microsoft/DeepSpeedExamples/tree/master/applications/DeepSpeed-Chat)<br />
[LoraChat](https://github.com/bupticybee/FastLoRAChat)<br />
[Tabby](https://github.com/TabbyML/tabby)<br />
[TalkToModel](https://github.com/dylan-slack/TalkToModel)<br />
[You.com](https://you.com/)<br />
### LangChain or Agent related
[Gradio Tools](https://github.com/freddyaboulton/gradio-tools)<br />
[LLM Agents](https://blog.langchain.dev/gradio-llm-agents/)<br />
[Meta Prompt](https://github.com/mbchang/meta-prompt)<br />
[HF Agents](https://huggingface.co/docs/transformers/transformers_agents)
[HF Agents Collab](https://colab.research.google.com/drive/1c7MHD-T1forUPGcC_jlwsIptOzpG3hSj)
[Einstein GPT](https://www.salesforce.com/products/einstein/overview/?d=cta-body-promo-8)
[SMOL-AI](https://github.com/smol-ai/developer)
[Pandas-AI](https://github.com/gventuri/pandas-ai/)
### Summaries
[LLMs](https://github.com/Mooler0410/LLMsPracticalGuide)<br />
### Deployment
[MLC-LLM](https://github.com/mlc-ai/mlc-llm)<br />
### Evaluations
[LMSYS (check for latest glob)](https://lmsys.org/blog/2023-05-25-leaderboard/)<br />
[LMSYS Chatbot Arena](https://chat.lmsys.org/?arena)<br />
[LMSYS Add model](https://github.com/lm-sys/FastChat/blob/main/docs/arena.md#how-to-add-a-new-model)<br />
[NLL](https://blog.gopenai.com/lmflow-benchmark-an-automatic-evaluation-framework-for-open-source-llms-ef5c6f142418)<br />
[HackAPrompt](https://www.aicrowd.com/challenges/hackaprompt-2023/leaderboards)<br />
|