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 />