We take Security seriously at WebBrain #llm #ai #offlineAI #qwen #gemma #gemma4 #llamacpp #openllama #opensource #chrome #ollama #
June 27, 2026 at 1:01 PM
We take Security seriously at WebBrain #llm #ai #offlineAI #qwen #gemma #gemma4 #llamacpp #openllama #opensource #chrome #ollama #
I think openllama is pretty important.
December 10, 2024 at 9:42 PM
I think openllama is pretty important.
And every Kaggle competition, every research investment, every paper written or kernel produced, it has all led to advancing the state of the art that now makes it possible to have your own open source LLM (github.com/openlm-resea...) that is built from data from permissively licensed sources.
GitHub - openlm-research/open_llama: OpenLLaMA, a permissively licensed open source reproduction of Meta AI’s LLaMA 7B trained on the RedPajama dataset
OpenLLaMA, a permissively licensed open source reproduction of Meta AI’s LLaMA 7B trained on the RedPajama dataset - openlm-research/open_llama
github.com
November 27, 2024 at 10:35 AM
And every Kaggle competition, every research investment, every paper written or kernel produced, it has all led to advancing the state of the art that now makes it possible to have your own open source LLM (github.com/openlm-resea...) that is built from data from permissively licensed sources.
My inner dialog for people creating all these #startup companies depending on #OpenAI, and why baby #LLM as OpenLLama or Mistral can outpeform the others and even run on your machine.
Btw #MistralAI is probably going to be funded for billions soon
#AI is ruining us www.youtube.com/shorts/IpKbZ...
Btw #MistralAI is probably going to be funded for billions soon
#AI is ruining us www.youtube.com/shorts/IpKbZ...
Why baby #LLM as OpenLLama or Mistral can outpeform the others and even run on your machine.
#shorts My inner dialog for people creating all these #startup companies depending on OpenAI, and why baby #LLM as OpenLLama or Mistral can outpeform the oth...
www.youtube.com
November 26, 2023 at 5:50 PM
My inner dialog for people creating all these #startup companies depending on #OpenAI, and why baby #LLM as OpenLLama or Mistral can outpeform the others and even run on your machine.
Btw #MistralAI is probably going to be funded for billions soon
#AI is ruining us www.youtube.com/shorts/IpKbZ...
Btw #MistralAI is probably going to be funded for billions soon
#AI is ruining us www.youtube.com/shorts/IpKbZ...
if the only outcome is "smarter messaging/calendar management", then that's probably reasonably attainable (and will become more attainable) with OpenLLAMA and other open source tools as well
December 12, 2023 at 6:27 PM
if the only outcome is "smarter messaging/calendar management", then that's probably reasonably attainable (and will become more attainable) with OpenLLAMA and other open source tools as well
Unfortunately I'm not the one creating them, I'm using pinokio, huggingface.co, openllama and github repos to learn, and they all use requirements.txt ...
July 29, 2024 at 2:36 PM
Unfortunately I'm not the one creating them, I'm using pinokio, huggingface.co, openllama and github repos to learn, and they all use requirements.txt ...
I don't suppose any of y'all have tried OpenLLaMa yet?
May 24, 2023 at 11:44 PM
I don't suppose any of y'all have tried OpenLLaMa yet?
Making a base model is expensive, but customizing it is not. And on the free end of art, we have Stable Diffusion. On LLMs, we have LLaMA 2, which is free for noncommercial use, and OpenLLaMA, which is fully free. And so on - there's public AI models for near everything.
August 14, 2023 at 3:57 PM
Making a base model is expensive, but customizing it is not. And on the free end of art, we have Stable Diffusion. On LLMs, we have LLaMA 2, which is free for noncommercial use, and OpenLLaMA, which is fully free. And so on - there's public AI models for near everything.
May 3, 2023 at 10:23 PM
“Let’s use OpenLLaMA. We can use it offline. It will work when it is on the server in even if there is no internet connection,” Violet suggested.
November 17, 2024 at 2:55 PM
“Let’s use OpenLLaMA. We can use it offline. It will work when it is on the server in even if there is no internet connection,” Violet suggested.
might also consider
tiiuae/falcon-mamba-7b
StripedHyena-Hessian-7B
openllama
mpt-3b-8k
robowaifudev/megatron-gpt2-345m
PleIAs/Pleias-3b-Preview
Zyphra/Zamba2-2.7B
gpt-j
KnutJaegersberg/GPT-JX-3b
T5
tiiuae/falcon-mamba-7b
StripedHyena-Hessian-7B
openllama
mpt-3b-8k
robowaifudev/megatron-gpt2-345m
PleIAs/Pleias-3b-Preview
Zyphra/Zamba2-2.7B
gpt-j
KnutJaegersberg/GPT-JX-3b
T5
December 10, 2024 at 12:27 PM
might also consider
tiiuae/falcon-mamba-7b
StripedHyena-Hessian-7B
openllama
mpt-3b-8k
robowaifudev/megatron-gpt2-345m
PleIAs/Pleias-3b-Preview
Zyphra/Zamba2-2.7B
gpt-j
KnutJaegersberg/GPT-JX-3b
T5
tiiuae/falcon-mamba-7b
StripedHyena-Hessian-7B
openllama
mpt-3b-8k
robowaifudev/megatron-gpt2-345m
PleIAs/Pleias-3b-Preview
Zyphra/Zamba2-2.7B
gpt-j
KnutJaegersberg/GPT-JX-3b
T5
🚀 Skyrocketing！(200+ new stars today) 🚀
📦 openlm-research / open_llama
⭐ 4,959 (+236)
OpenLLaMA, a permissively licensed open source reproduction of Meta AI’s LLaMA 7B trained on the RedPajama dataset
📦 openlm-research / open_llama
⭐ 4,959 (+236)
OpenLLaMA, a permissively licensed open source reproduction of Meta AI’s LLaMA 7B trained on the RedPajama dataset
GitHub - openlm-research/open_llama: OpenLLaMA, a permissively licensed open source reproduction of Meta AI’s LLaMA 7B trained on the RedPajama dataset
OpenLLaMA, a permissively licensed open source reproduction of Meta AI’s LLaMA 7B trained on the RedPajama dataset - GitHub - openlm-research/open_llama: OpenLLaMA, a permissively licensed open sou...
github.com
June 20, 2023 at 6:58 PM
🚀 Skyrocketing！(200+ new stars today) 🚀
📦 openlm-research / open_llama
⭐ 4,959 (+236)
OpenLLaMA, a permissively licensed open source reproduction of Meta AI’s LLaMA 7B trained on the RedPajama dataset
📦 openlm-research / open_llama
⭐ 4,959 (+236)
OpenLLaMA, a permissively licensed open source reproduction of Meta AI’s LLaMA 7B trained on the RedPajama dataset
it's possible that he took something like OpenLLAMA and tuned it on some Xitter data, but that particular output still doesn't seem very likely under that scenario
December 9, 2023 at 5:13 PM
it's possible that he took something like OpenLLAMA and tuned it on some Xitter data, but that particular output still doesn't seem very likely under that scenario
Anderer Ansatz wäre ein generisches LLM (wie chatgpt oder openllama) per langchain mit dem eigenen Spezialwissen anzureichern, z.B. als PDF/Word Dokumente von euch, die mit Balkonsolar zu tun haben.
January 7, 2024 at 1:19 PM
Anderer Ansatz wäre ein generisches LLM (wie chatgpt oder openllama) per langchain mit dem eigenen Spezialwissen anzureichern, z.B. als PDF/Word Dokumente von euch, die mit Balkonsolar zu tun haben.
Sure. As you say, you can probably run OpenLlama on a budget VM. But I don’t think it’s going to ‘escape’ and mythos is the most ‘dangerous’ model yet! Etc etc. and I don’t think it can replace white color jobs and run unsupervised. Also Bluesky is terrible for conversations.
April 25, 2026 at 8:08 PM
Sure. As you say, you can probably run OpenLlama on a budget VM. But I don’t think it’s going to ‘escape’ and mythos is the most ‘dangerous’ model yet! Etc etc. and I don’t think it can replace white color jobs and run unsupervised. Also Bluesky is terrible for conversations.
📦 CStanKonrad / long_llama
⭐ 401 (+94)
🗒 Jupyter Notebook
LongLLaMA is a large language model capable of handling long contexts. It is based on OpenLLaMA and fine-tuned with the Focused Transformer (FoT) method.
⭐ 401 (+94)
🗒 Jupyter Notebook
LongLLaMA is a large language model capable of handling long contexts. It is based on OpenLLaMA and fine-tuned with the Focused Transformer (FoT) method.
GitHub - CStanKonrad/long_llama: LongLLaMA is a large language model capable of handling long contexts. It is based on OpenLLaMA and fine-tuned with the Focused Transformer (FoT) method.
LongLLaMA is a large language model capable of handling long contexts. It is based on OpenLLaMA and fine-tuned with the Focused Transformer (FoT) method. - GitHub - CStanKonrad/long_llama: LongLLaM...
github.com
July 12, 2023 at 1:50 AM
📦 CStanKonrad / long_llama
⭐ 401 (+94)
🗒 Jupyter Notebook
LongLLaMA is a large language model capable of handling long contexts. It is based on OpenLLaMA and fine-tuned with the Focused Transformer (FoT) method.
⭐ 401 (+94)
🗒 Jupyter Notebook
LongLLaMA is a large language model capable of handling long contexts. It is based on OpenLLaMA and fine-tuned with the Focused Transformer (FoT) method.
"Our results suggest that OpenLLaMA does remember having seen a specific piece of content during training" ?!?Big implicaitons for audits/IP? arxiv.org/abs/2310.15007
Did the Neurons Read your Book? Document-level Membership...
With large language models (LLMs) poised to become embedded in our daily lives, questions are starting to be raised about the dataset(s) they learned from. These questions range from potential...
arxiv.org
October 28, 2023 at 8:02 AM
"Our results suggest that OpenLLaMA does remember having seen a specific piece of content during training" ?!?Big implicaitons for audits/IP? arxiv.org/abs/2310.15007
A equipe do 22h.ai lançou um novo modelo de IA treinado em língua portuguesa. Tomei a liberdade de converter para um formato que pode rodar em qualquer PC.
huggingface.co/lucianosb/op...
#llm #openllama 🎲
huggingface.co/lucianosb/op...
#llm #openllama 🎲
lucianosb/open-cabrita3b-GGUF · Hugging Face
We’re on a journey to advance and democratize artificial intelligence through open source and open science.
huggingface.co
August 28, 2023 at 10:26 AM
A equipe do 22h.ai lançou um novo modelo de IA treinado em língua portuguesa. Tomei a liberdade de converter para um formato que pode rodar em qualquer PC.
huggingface.co/lucianosb/op...
#llm #openllama 🎲
huggingface.co/lucianosb/op...
#llm #openllama 🎲
I confess I have no real idea, but the last time I looked at Asahi, it seemed to have a lot of caveats and downsides
I mean, if it’s just running macOS and then OpenLlama… could you live with it? You don’t need to look at macOS much, I think.
I mean, if it’s just running macOS and then OpenLlama… could you live with it? You don’t need to look at macOS much, I think.
March 13, 2025 at 1:22 PM
I confess I have no real idea, but the last time I looked at Asahi, it seemed to have a lot of caveats and downsides
I mean, if it’s just running macOS and then OpenLlama… could you live with it? You don’t need to look at macOS much, I think.
I mean, if it’s just running macOS and then OpenLlama… could you live with it? You don’t need to look at macOS much, I think.
Docker Offload is an interesting pivot. My guess is they are using predefined contexts in Docker desktop to redirect to their cloud services. This seems like something which could be done locally. Now to find that Nvidia GTX1060 I have in my spare parts bin for running OpenLLaMA. 🙂
July 14, 2025 at 3:25 PM
Docker Offload is an interesting pivot. My guess is they are using predefined contexts in Docker desktop to redirect to their cloud services. This seems like something which could be done locally. Now to find that Nvidia GTX1060 I have in my spare parts bin for running OpenLLaMA. 🙂
Fucking hell.. and what crap model would they use? OpenAI, well then can't be critical of them, CoPilot, guess we can't say fuck microsoft's shitty behavior, GoogleAI, bet you wouldn't be summarizing their scummy ad tactics. Only OpenLLAMA, if they're going to do it, should be used.
June 11, 2025 at 9:54 PM
Fucking hell.. and what crap model would they use? OpenAI, well then can't be critical of them, CoPilot, guess we can't say fuck microsoft's shitty behavior, GoogleAI, bet you wouldn't be summarizing their scummy ad tactics. Only OpenLLAMA, if they're going to do it, should be used.
approximation (SCA) and penalty dual decomposition (PDD) algorithm to solve the problem. Experiments with the OpenLLaMA v2 model with 3 billion parameters demonstrate up to $95\%$ faster convergence and over $90\%$ lower total energy consumption [6/7 of https://arxiv.org/abs/2506.00011v1]
June 3, 2025 at 6:01 AM
approximation (SCA) and penalty dual decomposition (PDD) algorithm to solve the problem. Experiments with the OpenLLaMA v2 model with 3 billion parameters demonstrate up to $95\%$ faster convergence and over $90\%$ lower total energy consumption [6/7 of https://arxiv.org/abs/2506.00011v1]