/𝖑𝖒𝖌/ 𝕷𝖔𝖈𝖆𝖑 𝕸𝖔𝖉𝖊𝖑 𝕲𝖊𝖓𝖊𝖗𝖆𝖑 𝕽𝖊𝖘𝖔𝖚𝖗𝖈𝖊 𝕲𝖚𝖎𝖉𝖊

V 6.56 | Helped Making Waifus Real Since 2023

Spingfield Artist: 豚さん | シューティングバー春田さん

This Rentry Guide is will serve as a quick guide for anons looking into working with Large Language Models like LlaMA or Pygmalion. While there are other detailed Rentry pages and guides around, however this Rentry will be geared towards users that are less technically inclined or just starting off.

𝕿𝖆𝖇𝖑𝖊 𝖔𝖋 𝕮𝖔𝖓𝖙𝖊𝖓𝖙𝖘

  1. Quick Rundown on Large Language Models:
    1. I. Work in Progress
  2. General Resources / Guides for Large Language Models (LLM) & Datasets
    1. LlaMA Models, Guides & Resources:
      1. Alpaca (LlaMA) Models, Guides & Resources:
      2. GPT4-ALL (LlaMA) Models, Guides & Resources:
      3. Vicuna (LlaMA) Models, Guides & Resources:
      4. Opan Assistant (Oasst-LlaMA) Models, Guides & Resources:
      5. WizardLM (LlaMA) Models, Guides & Resources:
      6. SuperCOT (LlaMA) Models, Guides & Resources:
      7. BluemoonRP (LlaMA) Models, Guides & Resources:
      8. Plugins & Other Projects for LlaMA models:
    2. ChatGLM "Chinese ChatGPT" Models, Guides & Resources:
    3. Pygmalion Models, Guides & Resources:
    4. Mosaic Pretrained Transformers (MPT) Models, Guides & Resources:
    5. RedPajama Models, Guides & Resources:
    6. Other Datasets & Related Resources:
  3. Related Projects for Large Language Model (LLM) & Guides
    1. General LLM Resources:
    2. Text Generation/Webui, Interfaces & Related Extensions:
    3. Machine Learning (ML) Framework Projects:
    4. Training Guides & Resources:
    5. Pruning Guides & Resources:
    6. Finetuning Guides & Resources:
    7. Quantization Guides & Resources:
    8. Benchmarking Guides & Resources:
  4. Guides/Resources for Story Generation, RP & ERP
    1. ERP, RP, & Story Resources:
  5. Character Guides & Resources
    1. Character Generators & General Information:
    2. Other Information for Characters & Waifus:
  6. All Things Miku
  7. All Other things
  8. Other Known Places for LLM (ML/AI) Discussions
  9. Archived Update Logs

Quick Rundown on Large Language Models:

This first section is still a working draft. Will work this section slowly and "git better" with another page as a draft until it's finished. Having a big incomplete section isn't good to look at imo.

I. Work in Progress

General Resources / Guides for Large Language Models (LLM) & Datasets

Tag Information || Last Updated: May 20th, 2023

I. ⇲ = Forked from Project
II. ↳ = Related Additional and/or Useful Resource
III. ◕ = Active Project || ◔ = Semi-active Project || ◎ = Unconfirmed Abandoned Project || ○ = Confirmed Abandoned Project
IV. Ŏ = Open Source || ⌀ = Closed Source || Ĩ= Institutional ⊻ University Projects || $ = Private ⊻ Corporate Projects
V. Strike out Entry = Project or model entry 404'd or was deleted
VI. ✅= Unfiltered/Uncucked Verified || 🚮 = Filtered/Cucked Verified (Not all links have been thoroughly verified)
VII. 🔔= User/Org. updated the quantization's for GGML model(s) after the recent change with llama.cpp
VIII. GH = GitHub || HF = Hugging Face

This section represents the current LLMs, datasets, guides and whatnot that's currently being developed (as well as an archive). You can find more detailed information (including download & torrent links) at this Rentry: https://rentry.org/lmg_models which is maintained by another great anon. That Rentry would more useful for anons who simply want access to the latest and greatest models and datasets. Charts below here should hopefully help serve as a reference (geared towards newer anons) about each project's activity status, quick information about the different models/datasets, guides, and a place to find other Rentry pages.

LlaMA Models, Guides & Resources:

Project/Dataset Name: (Hyperlinked) User(s) / Org. Project Description: Additional Info: (Framework, Inference, Quant. etc…) Entry Tags:
llama.cpp ⇲LlaMA ggerganov GH Repository CPU // LLaMA Port in C++ // 4-bit quant. support ◕ Ŏ
AgustusCletus/llama.cpp ⇲llama.cpp AgustusCletus GH Repository CPU ◎ Ŏ
llama-7b-ggml Camelids HF Quant. Model GGML Quant. q5_1 ◕ Ŏ 🔔
llama-13b-ggml Camelids HF Quant. Model GGML Quant. q5_1 ◕ Ŏ 🔔
llama-33b-ggml Camelids HF Quant. Model GGML Quant. q5_1 ◕ Ŏ 🔔
LLaMa-65B Quantized CRD716 HG Quant. Models GGML Quant. q4_0 q4_1 q4_2 q5_0 q5_1 ◕ Ŏ 🔔

Alpaca (LlaMA) Models, Guides & Resources:

Project/Dataset Name: (Hyperlinked) User(s) / Org. Project Description: Additional Info: (Framework, Inference, Quant. etc…) Entry Tags:
Stanford Alpaca ⇲LlaMA 7B Stanford University: Tatsu Lab GH Repository Code & Documentation ◕ Ŏ Ĩ
Alpaca LoRA ⇲Stanford Alpaca Tloen GH LoRA Instruct-tune LoRA for GPUs ◕ Ŏ
Alpaca-13b (Native) Chavinlo HF Trained Model GPU // ↳PyTorch ◕ Ŏ
GPT4-X-Alpaca ⇲Alpaca Native Chavinlo HF Finetuned Model GPU // ↳PyTorch // Dataset: ↳GPTeacher ◕ Ŏ
GPT4-X-Alpaca 13b Native 4bit ⇲Alpaca Native anon8231489123 HF Finetuned & Quant. Models Inference: Multi. // Dataset: ↳GPTeacher // GPTQ Quant. GGML Quant. 4_1 ◕ Ŏ
Alpacino13b Digitous HF "Triple Model Merge" GPU Inf: ↳Triton // ↳Alpaca-LoRA, ↳CoT-LoRA, ↳Storytelling-LoRA ◕ Ŏ
Alpacino-13b-ggml-q5_1 ⇲Alpacino13b Camelids HF Quant. Model CPU Inf. // GGML Quant. q5_1 ◕ Ŏ 🔔
Alpaca-LoRA-30B-elina LLMs HF LoRA Checkpoint "...recommended to be used with transformers >= 4.29" ◕ Ŏ
GPT4-x-AlpacaDente2-30b ⇲Alpaca-LoRA-30B-elina Aeala HF Finetuned Model GPU // ↳Open-Ass 2nd Finetune ◕ Ŏ

GPT4-ALL (LlaMA) Models, Guides & Resources:

Project/Dataset Name: (Hyperlinked) User(s) / Org. Project Description: Additional Info: (Framework, Inference, Quant. etc…) Entry Tags:
GPT4-ALL Nomic-AI GH Repository Code & Documentation // w/↳LoRA ◕ Ŏ
Using GPT4-ALL ↳llama.cpp ggerganov GH Repository Related Code & Documentation ◕ Ŏ

Vicuna (LlaMA) Models, Guides & Resources:

Project/Dataset Name: (Hyperlinked) User(s) / Org. Project Description: Additional Info: (Framework, Inference, Quant. etc…) Entry Tags:
Vicuna ⇲LlaMA Multiple Universities Uni. Project Finetuned LLaMA model with ShareGPT dataset. ◕ Ŏ/⌀ Ĩ
LlaVA Multiple Uni's, Microsoft GH Repository Visual Instruction Tune. ◕ Ŏ Ĩ $
legacy ggml vicuna 13b 4bit Eachadea HF Trained Model GGML(OLD) Quant. ○ Ŏ
legacy ggml vicuna 7b 4bit Eachadea HF Trained Model GGML(OLD) Quant. ○ Ŏ
ShareGPT Vicuna unfiltered anon8231489123 HF Refined Dataset Dataset: ↳ShareGPT ◕ Ŏ
Vicuna 13b free Reeducator HF Finetuned & Quant. Models Dataset: ↳ShareGPT_Vicuna_unfiltered // GGML Quant. f16, q5_0, GPTQ 4-bit CUDA. ◕ Ŏ 🔔
ggml vicuna 7b 1.1 Eachadea HF Finetuned & Quant. Models Dataset: ↳ShareGPT_Vicuna_unfiltered // GGML Quant. 4_0, 5_0, 5_1 ◕ Ŏ🔔
ggml vicuna 13b 1.1 Eachadea HF Finetuned & Quant. Models Dataset: ↳ShareGPT_Vicuna_unfiltered // GGML Quant. 4_0, 5_0, 5_1, 8_0 ◕ Ŏ🔔
Vicuna 13b cocktail Reeducator HF Finetuned & Quant. Models Datasets: ↳ShareGPT_Vicuna_unfiltered, ↳bluemoon_roleplay_300k_vicuna, ↳GPTeacher-Vicuna // GGML Quant. 5_0, 8_0, GPTQ 4-bit CUDA ◕ Ŏ🔔
VicUnlocked alpaca 30b 4bit Aeala HF Finetuned & Quant. Models Dataset: cleaned ↳ShareGPT // Alpaca format // GPTQ 4-bit CUDA ◕ Ŏ
VicUnlocked alpaca half 30b LoRA Aeala HF LoRA Instruct-tune Dataset: ↳ShareGPT_Vicuna_unfiltered // Benchmarks: ↳Wikitext2, ↳c4-new, ↳ptb-new // ⇲VicUnlocked alpaca half 30b LoRA ◕ Ŏ
VicUnlocked alpaca 30b Aeala HF Finetuned Model GPU // ↳PyTorch // Dataset: cleaned ↳ShareGPT // Merge: ⇲VicUnlocked alpaca half 30b LoRA ◕ Ŏ
VicUnLocked 30b LoRA Neko Institute of Science HF LoRA Instruct-tune Convert tool: ↳vicuna_to_alpaca ◕ Ŏ

Opan Assistant (Oasst-LlaMA) Models, Guides & Resources:

Project/Dataset Name: (Hyperlinked) User(s) / Org. Project Description: Additional Info: (Framework, Inference, Quant. etc…) Entry Tags:
Oasst LlaMA 13B Black-Engineer HF Finetuned and Quant. Model GGML(OLD) Quant. 4_0 ○ Ŏ
Oasst-sft-6-llama-30b OpenAssistant HF Weights xor ◕ Ŏ

WizardLM (LlaMA) Models, Guides & Resources:

Project/Dataset Name: (Hyperlinked) User(s) / Org. Project Description: Additional Info: (Framework, Inference, Quant. etc…) Entry Tags:
WizardLM nlpxucan GH Repository Trained model w/ full evolved instructions // ↳LlaMA 7B ◕ Ŏ
WizardVicunaLM Melodysdreamj GH Repository "Wizard's dataset + ChatGPT's conversation extension + Vicuna's tuning method" ◕ Ŏ
WizardLM-7B TheBloke HF Finetuned & Quant. Models Merging of delta weights w/ original Llama7B model // GGML Quant. 4_0, 5_0, 5_1, 8_0 ◕ Ŏ 🔔
WizardLM alpaca evol instruct 70k unfiltered Ehartford HF Refined Dataset "...WizardLM dataset victor123/evol_instruct_70k, removing instances of blatant alignment..." ◕ Ŏ
WizardLM alpaca claude evol inst. 70k Norquinal HF Refined Dataset Dataset: ↳Claude // Instruction: ↳WizardLM ◕ Ŏ
WizardLM-7B Uncensored Ehartford HF Finetuned Model GPU // ↳PyTorch // Dataset: ↳WizardLM alpaca evol instruct 70k unfiltered ◕ Ŏ ✅
WizardLM-7B Uncensored GPTQ ⇲WizardLM-7B Uncensored TheBloke HF Finetuned & Quant. Models Dataset: ↳WizardLM alpaca evol instruct 70k unfiltered // GPTQ 4-bit CUDA ◕ Ŏ ✅
WizardLM-7B Uncensored GGML TheBloke HF Finetuned & Quant. Models Dataset: ↳WizardLM alpaca evol instruct 70k unfiltered // GGML Quant. 4_0, 5_0, 5_1, 8_0 ◕ Ŏ ✅🔔
wizard-vicuna-13b Junelee HF Model Implementation GPU // ↳PyTorch ◕ Ŏ
Wizard-Vicuna-7B-Uncensored ⇲wizard-vicuna-13b Ehartford HF Trained Model GPU // ↳PyTorch // Dataset: ↳WizardLM alpaca evol instruct 70k unfiltered ◕ Ŏ

SuperCOT (LlaMA) Models, Guides & Resources:

Project/Dataset Name: (Hyperlinked) User(s) / Org. Project Description: Additional Info: (Framework, Inference, Quant. etc…) Entry Tags:
llama-13b-supercot-ggml-q5_1 Camelids HF Finetuned & Quant. Models GGML Quant. q5_1 ◕ Ŏ🔔
llama-33b-supercot-ggml-q5_1 Camelids HF Finetuned & Quant. Models GGML Quant. q5_1 ◕ Ŏ🔔

BluemoonRP (LlaMA) Models, Guides & Resources:

Project/Dataset Name: (Hyperlinked) User(s) / Org. Project Description: Additional Info: (Framework, Inference, Quant. etc…) Entry Tags:
Bluemoonrp-13b Reeducator HF Finetuned & Quant. Models Dataset: ↳bluemoon_roleplay_300k_vicuna // GGML Quant. q5_0 GPTQ Quant. Safetensors ◕ Ŏ 🔔
Bluemoon_RP_300k Gozfarb HF Pruned RP Dataset ↳Vicuna ◕ Ŏ

Plugins & Other Projects for LlaMA models:

Project/Dataset Name: (Hyperlinked) Info/Type: User(s) / Org. Tags: Resource Links:
Llama Hub Library/Plugin emptycrown ◕ Ŏ https://github.com/emptycrown/llama-hub
LLaMA retrieval plugin Plugin lastmile-ai ◔ Ŏ https://github.com/lastmile-ai/llama-retrieval-plugin
LlaMA Tokenizer Tokenizer Hugg. Internal Test. ◕ Ŏ https://huggingface.co/hf-internal-testing/llama-tokenizer/tree/main
LlaMA Index Data Augment. Jerryjliu ◕ Ŏ https://github.com/jerryjliu/llama_index

ChatGLM "Chinese ChatGPT" Models, Guides & Resources:

Project/Dataset Name: (Hyperlinked) User(s) / Org. Project Description: Additional Info: (Framework, Inference, Quant. etc…) Entry Tags:
ChatGLM-6B Tsinghua University GH Bilingual Dialogue LM ... ◕ Ŏ Ĩ
ChatGLM-130B Tsinghua University GH Bilingual Dialogue LM ... ◕ Ŏ Ĩ

Pygmalion Models, Guides & Resources:

Project/Dataset Name: (Hyperlinked) User(s) / Org. Project Description: Additional Info: (Framework, Inference, Quant. etc…) Entry Tags:
Pygmalion Pygmalion HF Account ... ◕
↳Pygmalion Windows Guide Anonymous Rentry Guide Pygmalion guide for Windows users ◕
↳Pygmalion Linux Guide Anonymous Rentry Guide Pygmalion guide for Linux users ◕
Pygmalion 350m ⇲BK OPT-350M Pygmalion HG Finetuned Model Datasets: "↳various sources" ◕ Ŏ
Pygmalion 1.3B ⇲pythia-1.3b-deduped Pygmalion HG Finetuned Model Datasets: "↳various sources" ◕ Ŏ
Pygmalion 2.7B ⇲gpt-neo-2.7B Pygmalion HG Finetuned Model Datasets: "↳various sources" ◕ Ŏ
Pygmalion 6B ⇲GPT-J-6B Pygmalion HG Finetuned Model Datasets: "↳various sources" ◕ Ŏ
Pygmalion 7B ↳LLaMA-7B Pygmalion HF Instruct Finetune xor // "...biased towards fiction writing and conversation." ◕ Ŏ
Metharme-13b ↳LLaMA-13B Pygmalion HF Instruct Finetune xor // "...biased towards fiction writing and conversation." ◕ Ŏ
Pygmalion-13b ↳LLaMA-13B Pygmalion HF Instruct Finetune xor // "...biased towards fiction writing and conversation." ◕ Ŏ

Mosaic Pretrained Transformers (MPT) Models, Guides & Resources:

Project/Dataset Name: (Hyperlinked) User(s) / Org. Project Description: Additional Info: (Framework, Inference, Quant. etc…) Entry Tags:
MPT-7B (Base) (Story Writer) (Instruct) (Chat) Mosaicml ↳ModusToolbox "MosaicML platform for LLM pretraining, finetuning, and inference" ◕ Ŏ Ĩ

RedPajama Models, Guides & Resources:

Project/Dataset Name: (Hyperlinked) User(s) / Org. Project Description: Additional Info: (Framework, Inference, Quant. etc…) Entry Tags:
RedPajama-Data RedPajama (GitHub) GH Data Repository ... ◕ Ŏ Ĩ
RedPajama INCITE 3B v1 Base-3B, Instruct-3B, Chat-3B Togethercomputer (RedPajama-HF) HF Native Models Multi. Inf. // ↳GPT-NeoX ◕ Ŏ Ĩ
RedPajama INCITE 7B v1 Base-7B, Instruct-7B, Chat-7B Togethercomputer (RedPajama-HF) HF Native Models Multi. Inf. // ↳GPT-NeoX ◕ Ŏ Ĩ
koishi-instruct-metharme Ewof HF Dataset Repository Datasets: A lot ◕ Ŏ
Koishi Instruct 3B ⇲INCITE Base 3B Ewof HF Native Finetune Model GPU // Dataset: ↳koishi-instruct-metharme ◕ Ŏ
Koishi Instruct 3B (GGML) f32 ⇲Koishi Instruct-3B Mongolian Basket Weaving HF Finetuned & Quant. Model GGML Quant. f32 ◕ Ŏ
Project/Dataset Name: (Hyperlinked) Info/Type: User(s) / Org. Tags: Resource Links:
ShareGPT ChatGPT Datalogs Domeccleston ◕ Ŏ https://github.com/domeccleston/sharegpt
GPTeacher GPT4 Modular Datasets Teknium1 ◕ Ŏ https://github.com/teknium1/GPTeacher
Encyclopedia Britannica Loaded Britannica Dataset Davanstrien ◕ Ŏ https://huggingface.co/datasets/davanstrien/encyclopedia_britannica_loaded
Embedding Wikipedia Guide & Resource Nils Reimers, Jay Alammar Cohere ◕ Ŏ $ https://txt.cohere.com/embedding-archives-wikipedia/
PKU SafeRLHF 10K PKU SafeRLHF Dataset Peking Uni. PKU-Align. ◕ Ŏ Ĩ 🚮 https://huggingface.co/datasets/PKU-Alignment/PKU-SafeRLHF-10K

Tag Information || Last Updated: May 16th, 2023

I. ⇲ = Forked from Project
II. ↳ = Related Additional and/or Useful Resource
III. ◕ = Active Project || ◔ = Semi-active Project || ◎ = Unconfirmed Abandoned Project || ○ = Confirmed Abandoned Project
IV. Ŏ = Open Source || ⌀ = Closed Source || Ĩ= Institutional ⊻ University Projects || $ = Private ⊻ Corporate Projects
V. Strike out Entry = Project or model entry 404'd or was deleted
VI. ✅= Unfiltered/Uncucked Verified || 🚮 = Filtered/Cucked Verified (Not all links have been thoroughly verified)

This section will focus on projects and repositories that deal with the creation/finetuning/pruning of large language models, frameworks, quantization, and benchmarking.

General LLM Resources:

Project Name: Type: Info. User(s) / Org. Tags: Resource Links:
Whisper.cpp Speech Recognition OpenAI Port ggerganov ◕ Ŏ https://github.com/ggerganov/whisper.cpp
AMD GPU ROCm GPU Platform Radeon ◕ Ŏ $ https://github.com/RadeonOpenCompute/ROCm
Model Torrents Rentry DL. Collection Anonymous ○ https://rentry.org/nur779
Model Torrents Backup Rentry DL. Collection Anonymous ◔ https://rentry.org/backupmdlist
LMG Models Rentry DL. Collection Anonymous ◕ https://rentry.org/lmg_models
Simple Proxy for Tavern Java Script ↳KoboldAI, ↳KoboldCPP, ↳SillyTavern anon998 ◕ Ŏ https://github.com/anon998/simple-proxy-for-tavern
Open-llama TensorFlow ↳RedPajama Dataset OpenLM Research ◕ Ŏ https://github.com/openlm-research/open_llama

Text Generation/Webui, Interfaces & Related Extensions:

Project Name: Type: Info. User(s) / Org. Tags: Resource Links:
Oobabooga's WebUI GPU Inf. Web User Interface Oobabooga ◕ Ŏ https://github.com/oobabooga/text-generation-webui
Wawa's TGW Ext. ↳Ooba WebUI Extension Long Term Memory Wawawario2 ◕ Ŏ https://github.com/wawawario2/text-generation-webui
KoboldCPP ⇲llama.cpp CPU Inf. w/↳CLBlast Web User Interface ↳GGML LostRuins ◕ Ŏ https://github.com/LostRuins/koboldcpp
KoboldAI ⇲llama.cpp CPU&GPU WebUI, ↳Pgyy ↳GPT-J KoboldAI ◕ Ŏ https://github.com/KoboldAI/KoboldAI-Client
Alpaca.cpp ⇲llama.cpp CPU Inf. CMD/Terminal Interface Antimatter15 ◔ Ŏ https://github.com/antimatter15/alpaca.cpp
Serge ↳llama.cpp CPU Inf. Web User Interface Nsarrazin ◕ Ŏ https://github.com/nsarrazin/serge
ChatLLaMA ↳Ooba WebUI Extension Backend for Discord Bots xNul ◕ Ŏ https://github.com/xNul/chat-llama-discord-bot
Oobabot ↳Ooba WebUI Extension Backend for Discord Bots Chrisrude ◕ Ŏ https://github.com/chrisrude/oobabot
SuperBIG ↳Ooba WebUI Extension Prompt / Context MGMT Syst. Kaiokendev ◕ Ŏ https://github.com/kaiokendev/superbig
SimpleLlama Webserver w/ ↳GPTQ Backend for Discord Bots NO-ob ◕ Ŏ https://github.com/NO-ob/simpleLlama/
Issho For LLaMA Models Web User Interface Stephen Tong ◕ Ŏ https://github.com/stong/issho https://issho.ai (Online Ver.)
TavernAI ↳KoboldAI ↳NovelAI Web User Interface TavernAI ◕ Ŏ https://github.com/TavernAI/TavernAI
SillyTavern ⇲TavernAI Multi Inf. Web User Interface Cohee1207 ◕ Ŏ SillyTavern, SillyTavern - dev
Agnaistic ⇲galatea-ui Web User Interface Hosted Ver. @ Agnai.chat LuminAI ◕ Ŏ https://github.com/luminai-companion/agn-ai

Machine Learning (ML) Framework Projects:

Project Name: Type: User(s) / Org. Tags: Resource Links:
PyTorch Python Pkg. Facebook Inc. ◕ Ŏ $ https://github.com/pytorch/pytorch
Intel Extension ↳PyTorch Extension Intel Corp. ◕ Ŏ $ https://github.com/intel/intel-extension-for-pytorch
Transformers ML Library Huggingface ◕ Ŏ https://github.com/huggingface/transformers
GGML Tensor library ggerganov ◕ Ŏ https://github.com/ggerganov/ggml
Cuda Python Binding NVIDIA ◕ Ŏ $ https://github.com/NVIDIA/cuda-python
FlexGen Gen. Engine Multiple Entities ◕ Ŏ Ĩ $ https://github.com/FMInference/FlexGen
DeepSpeed Opt. Library Microsoft ◕ Ŏ $ https://github.com/microsoft/DeepSpeed
ColossalAI DL System Multiple Entities ◕ Ŏ Ĩ https://github.com/hpcaitech/ColossalAI
ModusToolbox Inference Engine Infineon ◕ Ŏ Ĩ https://github.com/Infineon/ml-inference
TensorFlow ML Framework TensorFlow, Google Brain ◕ Ŏ Ĩ https://github.com/tensorflow/tensorflow
ImageBind Vision & Pattern Rec. Facebook Inc. ◕ Ŏ $ https://github.com/facebookresearch/ImageBind
Megatron-LM Transformer NVIDIA ◕ Ŏ $ https://github.com/NVIDIA/Megatron-LM
GPT-NeoX ↳Megatron Transformer ↳DeepSpeed EleutherAI ◕ Ŏ Ĩ https://github.com/EleutherAI/gpt-neox
PKU Beaver RLHF Framework Peking Uni. PKU-Align. ◕ Ŏ Ĩ 🚮 https://github.com/PKU-Alignment/safe-rlhf
NVIDIA Triton Inference Inference Engine Triton Inference Server ◕ Ŏ $ https://github.com/triton-inference-server/server

Training Guides & Resources:

Project Name: Type: Info: User(s) / Org. Tags: Resource Links:
FastChat Training Platform w/ Web UI Multiple Universities ◕ Ŏ/⌀ Ĩ https://github.com/lm-sys/FastChat#readme
Ooba's Training Guide Train. Guide Ooba's WebUI Oobabooga ◕ Ŏ https://github.com/oobabooga/text-generation-webui/blob/main/docs/Training-LoRAs.md
DASLab GPTQ Post-Train. GPTQ Impl. IST Austria: Alistarh Group ◕ Ŏ Ĩ https://github.com/IST-DASLab/gptq
Pythia Training Huggingface EleutherA ◕ Ŏ Ĩ https://github.com/EleutherAI/pythia
Dataset Preparation Tool for LLM Training Training Tool ↳YAML Config. w/ text transformers Justin42 ◕ Ŏ https://github.com/Justin42/dsbuild

Pruning Guides & Resources:

Project Name: Type: Info: User(s) / Org. Tags: Resource Links:
SparseGPT ↳GPTQ IST-DASLab raw-WikiText2, PTB, C4-subset IST Austria: Alistarh Group ◕ Ŏ Ĩ https://github.com/IST-DASLab/sparsegpt
⇲SparseGPT for LLaMA ↳LLaMA Implementation AlpinDale ◔ Ŏ https://github.com/AlpinDale/sparsegpt-for-LLaMA

Finetuning Guides & Resources:

Project Name: Type: Info: User(s) / Org. Tags: Resource Links:
Low-Rank Adaptation of Large Language Models (LoRA) loralib Pre-Train. Microsoft ◕ Ŏ $ https://github.com/microsoft/LoRA
Parameter Efficient Fine Tuning Train. Method PEFT Huggingface ◕ Ŏ https://github.com/huggingface/peft
LlaMA Adapter Adapter Implementation ZrrSkywalker ◕ Ŏ https://github.com/ZrrSkywalker/LLaMA-Adapter
xTuring Fine-tune Interface INT4 w/ ↳LoRA Stochastic: Cambridge Uni. ◕ Ŏ Ĩ https://github.com/stochasticai/xturing
Axolotl Train. Method ↳LlaMA ↳Cerebras ↳Pythia Winglian OpenAccess AI Collective ◕ Ŏ https://github.com/OpenAccess-AI-Collective/axolotl
Dromedary Finetuning Min. Supervision IBM ◕ Ŏ https://github.com/IBM/Dromedary

Quantization Guides & Resources:

Project Name: Type: Info: User(s) / Org. Tags: Resource Links:
GPTQ LlaMA 4-bit Quantization Method Qwopqwop200 ◕ Ŏ https://github.com/qwopqwop200/GPTQ-for-LLaMa
RPTQ LlaMA 3 and 4-bit Quantization Method AlpinDale ◕ Ŏ https://github.com/AlpinDale/RPTQ-for-LLaMA
Reduced-Kobold Impl. of ↳GPTQ1&↳SparseGPT KoboldAI & Ooba's UI Mstnegate ◔ Ŏ https://github.com/mstnegate/reduced-kobold
GPTQ 4 LLaMa ⇲ROCm ↳GPTQ-for-LlaMA WapaMario63 ◕ Ŏ https://github.com/WapaMario63/GPTQ-for-LLaMa-ROCm

Benchmarking Guides & Resources:

Project Name: Type: Info: User(s) / Org. Tags: Resource Links:
LLM Benchmark Prompts Pastebin Prompts Anonymous ◕ https://pastebin.com/LmRhwUCA

Guides/Resources for Story Generation, RP & ERP

Story/RP/ERP Generation Tag Information || Last Updated: May 13th, 2023

I. 🏢 = Safe for work (SFW) || 🔞 = Not Safe for Work (NSFW)
II. ◕ = Active Project || ◔ = Semi-active Project || ◎ = Unconfirmed Abandoned Project || ○ = Confirmed Abandoned Project
III. ↳ = Related Additional and/or Useful Resource
IV. Strike out Entry = Project or model entry 404'd or was deleted

This section will focus on resources for Story writing, Roll-Playing, Erotic-Roll-Playing using WebUIs for LLMs and other interfaces.

ERP, RP, & Story Resources:

Project Name: Type: Info: User(s) / Org. Tags: Resource Links:
RolePlayBot Guide Rentry Guide Anonymous ◕ https://rentry.org/RPBT
ERP/RP Data Collection Rentry Data Log Collection Anonymous ◕ https://rentry.org/qib8f
Anon's Guide to LLaMA Roleplay Rentry Data Log Collection Anonymous ◕ https://rentry.org/better-llama-roleplay

Character Guides & Resources

Tag Information || Last Updated: May 13th, 2023

I. 🏢 = Safe for work (SFW) || 🔞 = Not Safe for Work (NSFW)
II. ◕ = Active Project || ◔ = Semi-active Project || ◎ = Unconfirmed Abandoned Project || ○ = Confirmed Abandoned Project
III. ↳ = Related Additional and/or Useful Resource
IV. Strike out Entry = Project or model entry 404'd or was deleted

This section will focus on characters and resources for large language model interfaces, which will contain Rentry pages, generators, guides and other overall resources.

Character Generators & General Information:

Project Name: Type: Info: User(s) / Org. Tags: Resource Links:
Oobabooga's Character Guide GitHub Doc. Char. Guide Oobabooga ◕ Ŏ https://github.com/oobabooga/text-generation-webui/blob/main/docs/Custom-chat-characters.md
Oobabooga's Character Creator Char. Generator .json Oobabooga ◕ Ŏ https://oobabooga.github.io/character-creator.html
ZoltanAI Character Editor Char. Generator Tavern, Pyggy, CAI ZoltanAI ◕ Ŏ https://github.com/ZoltanAI/character-editor
Booru Collection for Pygmalion Char. Collection Booru Library .json Tavern ◕ https://booru.plus/+pygmalion
Meta Bot Lists Rentry Character Collection /aicg/ ◕ https://rentry.org/meta_bot_list
wAIfu - Tavern Cards Tavern Cards mega.nz file wAIfu ◕ https://mega.nz/folder/ySAHXaaK#jDHN_RpvFrVorZWYMtEAiA
CharacterHub Char. Library Tavern CAI CharacterHub ◕ https://www.characterhub.org

Other Information for Characters & Waifus:

Project & Model Names: Type: Info: User(s) / Org. Tags: Resource Links:
Girls Frontline Char. Files Rentry .json & Tavern Cards nymous ◕ https://rentry.org/GFLBots/ (AICG Bots - Not LLM) https://booru.plus/+pygmalion/@nymous (Booru Depository)
Wster's Characters Profile ↳Character Hub Wster ◕ https://www.characterhub.org/users/wster

All Things Miku

Project & Model Names: Type/Info: User(s) / Org. Tags: Resource Links:
OG Miku Pastebin Anonymous ◕ https://pastebin.com/5WVd28Um
Reverse Miku Pastebin Anonymous ○ https://pastebin.com/yG6xWMS
Miku.sh Pastebin Anonymous ◕ https://pastebin.com/vWKhETWS

All Other things

Project & Model Names: Type: Info: User(s) / Org. Tags: Resource Links:
/lmg/ Baking Template Rentry Baking Template Anonymous ◕ https://rentry.org/lmg_template
/lmg/ Drama Rentry Rentry Charts with Rentry Drama pages Anonymous ◕ https://rentry.org/lmg-drama
/lmg/ Papers Backup Backup Rentry /lmg/ usefull papers backup Anonymous ◕ https://rentry.org/lmg_papers_backup
Local Models Related Links Backup Backup Rentry LM Related Links Anonymous ◕ https://rentry.org/lm_related_links_backup

Other Known Places for LLM (ML/AI) Discussions

PLEASE DO NOT SHILL ANY IRC, REDDIT, DISCORD SERVERS OR WHATNOT IN THE /LMG/ THREADS!! IT SHOULD BE KNOWN THAT THERE IS NO OFFICIAL /LMG/ DISCORD!! NONE OF THE PLACES BELOW REPRESENTS ALTERNATIVE COMMUNITIES OR PLACES FOR /LMG/. THEY ARE PLACES WHERE LLM'S IN GENERAL ARE DISCUSSED, DEVELOPED AND GENERALLY HAVE RESOURCES FOR LLM'S.

GitHub User(s) / Org. Tags: Discussion Links:
Ggerganov's llama.cpp ggerganov ◕ https://github.com/ggerganov/llama.cpp/discussions
Oobabooga's WebUI Oobabooga ◕ https://github.com/oobabooga/text-generation-webui/discussions
Koboldcpp LostRuins ◕ https://github.com/LostRuins/koboldcpp/discussions
KoboldAI KoboldAI ◕ https://github.com/KoboldAI/KoboldAI-Client/discussions
Wawawario2's long term memory Wawawario2 ◕ https://github.com/wawawario2/long_term_memory/discussions
Other Imageboards User(s) / Org. Tags: Discussion Links:
/robowaifu/ N.L ◕ https://alogs.space/robowaifu/catalog.html
IRC User(s) / Org. Tags: Discussion Links:
#Localmodels N.L ◕ irc.libera.chat #localmodels
Discord User(s) / Org. Tags: Discussion Links:
RedPajama N.L ◕ https://discord.com/invite/9Rk6sSeWEG
Stable Diffusion N.L ◕ discord.gg/stablediffusion
Pygmalion N.L ◕ https://discord.com/invite/ZHXEa3yywq
KoboldAI N.L ◕ https://koboldai.org/discord
ChatLLaMA N.L ◕ discord.gg/TcRGDV754Y
Reddit User(s) / Org. Tags: Discussion Links:
LocalLLaMA N.L ◕ https://www.reddit.com/r/LocalLLaMA/

Archived Update Logs

Week of (MM/WK/YYYY) Description:
05/wK-3/2023 Charts are now fixed, no more "...", Only the first FAQ section has missing information. Next comes the fine touches, adding the few missing entries, & work on the first FAQ section. // Gozfarb's HF got 404'd, F // Need to re-work the model & dataset chart section, it had more then a few mistakes with it's tags. Model and Dataset Charts should be completed this weekend, the checklist: (LlaMA Models, Alpaca (LlaMA), GPT4-ALL (LlaMA), Vicuna (LlaMA), Opan Assistant (Oasst-LlaMA), WizardLM (LlaMA), SuperCOT (LlaMA), BluemoonRP (LlaMA), "Plugins & Other Projects for LlaMA", ChatGLM, Pygmalion Models, MPT, RedPajama Models, "Other Datasets") // Recent change with llama.cpp made old GGML quantizations incompatible. Need to go through each model and add a 🔔 indicated the model was updated with the new change along with the chart changes.
05/WK-2/2023 Updates will be slower this week (Sunday through Friday)... I will be spending time with friends and family as well as going out for my birthday. After next week, I should be regularly updating this Rentry again. / Couldn't help myself and decided to work on this Rentry again while on break. / Cosmetic changes were made / Charts were mostly filled in / Character & RP/ERP charts were updated along with new tags info / More polishing, grammar & spelling corrections. / Since / make frequent changes on the fly, updates will be trickled in this box, then cycled weekly and archived at the bottom. / Hopefully Ver. 8 will be the first "finished version" of this Rentry with the general guide in the first section complete. Ver. 7 will mean the charts below should be completely organized and in order, only new entries should be made.
Edit
Raw PDF PNG
Pub: 29 Mar 2023 17:13 UTC
Edit: 29 May 2023 10:02 UTC
Views: 49436
new·what·how·langs·contacts