home.social

#mixtral8x7b — Public Fediverse posts

Live and recent posts from across the Fediverse tagged #mixtral8x7b, aggregated by home.social.

fetched live
  1. Für das Tuxedo Notebook wollte ich gestern optimistisch RAM kaufen. Aktuell sind 32GB drin. Für Spielereien mit Ollama und lokal installiertem mixtral:8x7b ist das etwas mager. 128er Riegel sind gar keine mehr verfügbar und 2x64GB kosten 940.- 😲
    Hätte ich doch nur auf meinen Instinkt gehört und beim Kauf des Notebooks direkt 128GB bestellt - damals, als die Preise noch normal waren.

    #ram #ai #ollama #localllm #mixtral8x7b

  2. Für das Tuxedo Notebook wollte ich gestern optimistisch RAM kaufen. Aktuell sind 32GB drin. Für Spielereien mit Ollama und lokal installiertem mixtral:8x7b ist das etwas mager. 128er Riegel sind gar keine mehr verfügbar und 2x64GB kosten 940.- 😲
    Hätte ich doch nur auf meinen Instinkt gehört und beim Kauf des Notebooks direkt 128GB bestellt - damals, als die Preise noch normal waren.

    #ram #ai #ollama #localllm #mixtral8x7b

  3. Au taff, on vient de recycler un vieux coucou avec 6 cartes GTX 1070 Ti, des trucs de 2017 dits "désuets" par #Nvidia.
    #llamacpp démarre un #Mixtral8x7b réparti sur les cartes... Et ça va plus vite que ChatGPT.
    Une interface codée en #gradio de 30 lignes.
    Si ça c'est pas rentabiliser !

    PS: on va virer CUDA et utiliser #Vulkan d'ici peu

  4. Just discovered that DDG search offers an AI chat feature! You can choose between GPT 3.5, Mixtral 8x7B, and more. It's incredibly useful and awesome, definitely worth a try! .5

  5. @catguy Sure. For image generation #sdxl based models, and my favourite #llm is #mixtral8x7b. These two are great choices. I would really like to host mixtral on my server using #ollama #docker.

  6. @catguy Sure. For image generation #sdxl based models, and my favourite #llm is #mixtral8x7b. These two are great choices. I would really like to host mixtral on my server using #ollama #docker.

  7. Due to #dhl, my #hp #elitedesk arrives way later than expected. I'm already so excited to upgrade it. Bought it nearly #barebones and i'm planning to equip it with a 2TB nvme for main storage, 2TB 2.5" hdd for backups, an intel core i5 (not quite sure which one rn), and 32GB of ram (64GB if #mixtral8x7b runs smooth that way).

  8. Due to #dhl, my #hp #elitedesk arrives way later than expected. I'm already so excited to upgrade it. Bought it nearly #barebones and i'm planning to equip it with a 2TB nvme for main storage, 2TB 2.5" hdd for backups, an intel core i5 (not quite sure which one rn), and 32GB of ram (64GB if #mixtral8x7b runs smooth that way).

  9. artificialanalysis.ai/models/m

    Quality vs. Throughput.

    mixtral-8x7b-instruct on groq

    Wow! What a difference with the LPU.

    This makes it really cheaper.

    #LLMs #Mistral #mixtral8x7b #groq

  10. artificialanalysis.ai/models/m

    Quality vs. Throughput.

    mixtral-8x7b-instruct on groq

    Wow! What a difference with the LPU.

    This makes it really cheaper.

    #LLMs #Mistral #mixtral8x7b #groq

  11. Tested #Mixtral8x7b #LLM in #AI "lab". Amazing precise answers about private documents. Something like #PrivateGPT / #LocalGPT but way better. Sadly, can't host publicly.
    Already full of ideas for #business cases.
    Now, I need €€€€... €, to get hardware for homelab 😎🤣💶💶💶

  12. Also, #mixtral8x7b is such a great #llm! I'm using it daily and i'm happy to see it getting integrated into my main search engine.

  13. Also, #mixtral8x7b is such a great #llm! I'm using it daily and i'm happy to see it getting integrated into my main search engine.

  14. Everybody’s talking about Mistral, an upstart French challenger to OpenAI - Enlarge / An illustration of a robot holding a French flag, figurativel... - arstechnica.com/?p=1989986 #largelanguagemodels #machinelearning #arthurmensch #mixtral8x7b #anthropic #mistralai #chatgpt #chatgtp #gpt-3.5 #mistral #mixtral #biz#france #llama2 #openai #gpt-3 #gpt-4 #llama #paris #meta #ai

  15. Everybody’s talking about Mistral, an upstart French challenger to OpenAI - Enlarge / An illustration of a robot holding a French flag, figurativel... - arstechnica.com/?p=1989986 #largelanguagemodels #machinelearning #arthurmensch #mixtral8x7b #anthropic #mistralai #chatgpt #chatgtp #gpt-3.5 #mistral #mixtral #biz#france #llama2 #openai #gpt-3 #gpt-4 #llama #paris #meta #ai

  16. 💡Mistral AI annuncia un nuovo LLM, Mixtral 8x7B
    Mistral AI lancia Mixtral 8x7B, è il modello LLM open-weight più potente con una licenza permissiva e il miglior modello in assoluto per quanto riguarda il compromesso costi/prestazioni.

    gomoot.com/mistral-ai-il-rival
    #GPT #Llama #Mistral7B #MistralAI #Mixtral #Mixtral8x7B #ai #ia #intelligenzaartificiale @MistralAI @dchaplot @PierreStock @GuillaumeLample

  17. 💡Mistral AI annuncia un nuovo LLM, Mixtral 8x7B
    Mistral AI lancia Mixtral 8x7B, è il modello LLM open-weight più potente con una licenza permissiva e il miglior modello in assoluto per quanto riguarda il compromesso costi/prestazioni.

    gomoot.com/mistral-ai-il-rival
    #GPT #Llama #Mistral7B #MistralAI #Mixtral #Mixtral8x7B #ai #ia #intelligenzaartificiale @MistralAI @dchaplot @PierreStock @GuillaumeLample