/π–‘π–’π–Œ/ π•·π–”π–ˆπ–†π–‘ π•Έπ–”π–‰π–Šπ–‘ π•²π–Šπ–“π–Šπ–—π–†π–‘ π•Ώπ–Šπ–’π–•π–‘π–†π–™π–Š

V 1.20 | Helped Making Waifus Real Since 2023


Artist: keenH | Source: https://www.pixiv.net/en/artworks/76986607 Artist: keenH | HK416-C


News Section Updated (07/07/2023)


Main Body for OP Template:

β–ΊPrevious Threads >>9xxxxxxx & >>9xxxxxxx

β–ΊNews
>(07/01) OpenChat and OpenCoderPlus models released
>(06/29) LMSYS releases longchat 7B/13B with 16k context
>(06/29) OpenOrca dataset releases
>(06/29) koboldcpp v1.33 adds CUDA and extended context support
>(06/29) OpenFlamingo v2, Airoboros 65B GPT-4 v1.4 releases
>(06/28) Salesforce releases XGen family of models (7B)
>(06/26) Ooba's webui adds support for extended context with exllama
>(06/24) WizardLM-33B-V1.0-Uncensored releases
>(06/23) SuperHOT 30B 8k prototype + extending context write up releases

β–ΊFAQ & Wiki
>Main FAQ
https://rentry.org/er2qd

β–ΊGeneral LLM Guides & Resources
>Newb Guide
https://rentry.org/local_LLM_guide
>LlaMA Guide
https://rentry.org/TESFT-LLaMa
>Machine Learning Roadmap
https://rentry.org/machine-learning-roadmap
>Novice's LLM Training Guide
https://rentry.org/llm-training
>Local Models Papers
https://rentry.org/LocalModelsPapers
>Quantization Guide
https://rentry.org/easyquantguide
>lmg General Resources
https://rentry.org/lmg-resources
>ROCm AMD Guide
https://rentry.org/eq3hg

β–ΊModel DL Links, & Guides
>Model Links & DL
https://rentry.org/lmg_models
>lmg Related Links
https://rentry.org/LocalModelsLinks

β–ΊText Gen. UI
>Text Gen. WebUI
https://github.com/oobabooga/text-generation-webui
>KoboldCPP
https://github.com/LostRuins/koboldcpp
>KoboldAI
https://github.com/0cc4m/KoboldAI
>SimpleLlama
https://github.com/NO-ob/simpleLlama

β–ΊERP/RP/Story Gen.
>RolePlayBot
https://rentry.org/RPBT
>ERP/RP Data Collection
https://rentry.org/qib8f
>LLaMA RP Proxy
https://rentry.org/better-llama-roleplay

β–ΊOther Resources
>Drama Rentry
https://rentry.org/lmg-drama
>Miku
https://rentry.org/lmg-resources#all-things-miku
>Baking Template
https://rentry.org/lmg_template
>Benchmark Prompts
https://pastebin.com/LmRhwUCA
>Simple Proxy for WebUI (+output quality)
https://github.com/anon998/simple-proxy-for-tavern
>Additional Links
https://rentry.org/lmg_template#additional-resource-links

Follow-up Re-cap Post (Optional, but great to have)

β–ΊRecent Highlights from the Previous Thread: >>92XXXXXX (Previous OP Thread Reference)

--Anon goes over the repo process of the LLaMA-Adapter >>92XXXXXX
--Anon releases reverse Miku >>92XXXXXX
--Anons brings up the idea of a mascot for /lmg/ >>92XXXXXX >>>>92XXXXXX
--Oobabooga merged commit to fix one-click-install issue >>92XXXXXX
--Anons proposes a good use for wawario2's long-term memory >>92XXXXXX >>92XXXXXX
--etc...

β–ΊRecent Highlight Post from the Previous Thread: >>92XXXXXX (Previous Recap Post Reference)

Description: Links:
llama.cpp https://github.com/ggerganov/llama.cpp
Open LlaMA https://github.com/openlm-research/open_llama
Other Rentry Pages: -_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_
Pygmalion. Windows Guide https://rentry.org/Pyggymancy
Pygmalion Linux Guide https://rentry.org/pygmalion-local
Moe Archives https://rentry.org/ashh2
Ayumi ERP Rating https://rentry.org/ayumi_erp_rating
Another LLM Roleplay Rankings https://rentry.co/ALLMRR
Popular HF Model Contributors: -_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_
HF Account: Eachadea https://huggingface.co/eachadea
HF Account: Camelids https://huggingface.co/camelids
HF Account: Waifu Workshop https://huggingface.co/waifu-workshop
HF Account: Ehartford https://huggingface.co/ehartford
HF Account: TheBloke https://huggingface.co/TheBloke

Archived News:

Date: (MM/DD/YYYY) Description:
06/23/2023 Ooba's preset arena results and SuperHOT 16k prototype releases
06/22/2023 Vicuna 33B (preview), OpenLLaMA 7B scaled and MPT 30B released
06/20/2023 SuperHOT Prototype 2 w/ 8K context released >>94191797
06/18/2023 Minotaur 15B 8K, WizardLM 7B Uncensored v1.0 and Vicuna 1.3 released
06/17/2023 exllama support merged into ooba; API server rewrite merged into llama.cpp
06/16/2023 OpenLlama 13B released
06/16/2023 Airoboros GPT-4 v1.2 released
06/16/2023 Robin-33B-V2 released
06/16/2023 Dan's 30B Personality Engine LoRA released
06/14/2023 WizardCoder 15B Released
06/14/2023 CUDA full GPU acceleration merged in llama.cpp
06/10/2023 First Landmark Attention models released >>93993800 (Cross-thread)
06/08/2023 Openllama 3B and 7B released
06/07/2023 StarCoderPlus / StarChat-Ξ² released
06/07/2023 chronos-33b released
06/06/2023 RedPajama 7B released + Instruct&Chat
06/06/2023 WizardLM 30B v1.0 released
06/05/2023 k-quantization released for llama.cpp
06/03/2023 Nous-Hermes-13b released
06/03/2023 WizardLM-Uncensored-Falcon-40b released
05/27/2023 FalconLM release Falcon-7B & 40B, new foundational models
05/26/2023 BluemoonRP 30B 4K released
05/25/2023 QLoRA and 4bit bitsandbytes released
05/23/2023 exllama transformer rewrite offers around x2 t/s increases for GPU models
05/22/2023 SuperHOT 13B prototype & WizardLM Uncensored 30B released
05/22/2023 SuperHOT 13B prototype & WizardLM Uncensored 30B released
05/19/2023 RTX 30 series 15% performance gains, quantization breaking changes again >>93536523
05/19/2023 PygmalionAI release 13B Pyg & Meth
05/18/2023 VicunaUnlocked-30B released
05/14/2023 llama.cpp quantization change breaks current Q4 & Q5 models, must be quantized again
05/13/2023 llama.cpp GPU acceleration has been merged onto master >>93403996 >>93404319
05/10/2023 GPU-accelerated token generation >>93334002
05/06/2023 MPT 7B, 65k context model trained on 1T tokens: https://huggingface.co/mosaicml/mpt-7b-storywriter
05/05/2023 GPT4-x-AlpacaDente2-30b. https://huggingface.co/Aeala/GPT4-x-AlpacaDente2-30b
05/04/2023 Allegedly leaked document from Google, fretting over Open Source LLM's. https://www.semianalysis.com/p/google-we-have-no-moat-and-neither
05/04/2023 StarCoder, a 15.5B parameter models trained on 80+ programming languages: https://huggingface.co/bigcode/starcoderbase
04/30/2023 Uncucked Vicuna 13B released: https://huggingface.co/reeducator/vicuna-13b-free
04/30/2023 PygmalionAI release two 7B LLaMA-based models: https://huggingface.co/PygmalionAI
04/29/2023 GPT4 X Alpasta 30B Merge: https://huggingface.co/MetaIX/GPT4-X-Alpasta-30b-4bit
04/25/2023 Proxy script for Tavern via Kobold/webui, increases LLaMA: output quality https://github.com/anon998/simple-proxy-for-tavern
04/23/2023 OASS 30B released & quantized: https://huggingface.co/MetaIX/OpenAssistant-Llama-30b-4bit
04/22/2023 SuperCOT LoRA (by kaiokendev), merged by helpful anons: https://huggingface.co/tsumeone/llama-30b-supercot-4bit-128g-cuda https://huggingface.co/ausboss/llama-13b-supercot-4bit-128g
04/22/2023 OASS "releases" XORs again, deletes them soon after... again
04/21/2023 StableLM models performing terribly, are apparently broken: https://github.com/Stability-AI/StableLM/issues/30

Archived Updates

Week of (MM/WK/YYYY) Description:
05/WK-3/2023 Description and name updated for "Proxy for Tavern', shortened number of characters // Last template edit was utter dogshit. Reverted some changes, implemented anon user inputs and added an overflow chart down below. Should have about ~200 extra characters even with the news section as it is right now
05/WK-2/2023 Local LLM Guide added // google leaked doc news moved to the "Archived News" // https://rentry.org/LocalModelsLinks & https://rentry.org/LocalModelsPapers have returned, thank you Anon!
Edit
Pub: 06 May 2023 17:12 UTC
Edit: 07 Jul 2023 14:15 UTC
Views: 14159