#dgx_spark — Public Fediverse posts

Ngl I have a nvidia dgx spark at work and it’s so much fun to have a little self contained box that is reasonably capable and fits into a backpack.

I’ve been running so many machine learning experiments (from “traditional learning” to “spectrum of vision models” to llms and transformers) just because I can do it without setting up any api keys and cloud compute nonsense or having my laptop work itself to the heat-death of the universe on its puny i7.

Tempted to go into a cabin in the woods with zero internet and just constrain myself at programming, llm augmented sure, but completely autonomously on 70w of compute. Wild to think that this is where we are now, 4-5 years after copilot alpha.

#llms #llm #nvidia #dgx_spark

Ngl I have a nvidia dgx spark at work and it’s so much fun to have a little self contained box that is reasonably capable and fits into a backpack.

I’ve been running so many machine learning experiments (from “traditional learning” to “spectrum of vision models” to llms and transformers) just because I can do it without setting up any api keys and cloud compute nonsense or having my laptop work itself to the heat-death of the universe on its puny i7.

Tempted to go into a cabin in the woods with zero internet and just constrain myself at programming, llm augmented sure, but completely autonomously on 70w of compute. Wild to think that this is where we are now, 4-5 years after copilot alpha.

#llms #llm #nvidia #dgx_spark

Ngl I have a nvidia dgx spark at work and it’s so much fun to have a little self contained box that is reasonably capable and fits into a backpack.

I’ve been running so many machine learning experiments (from “traditional learning” to “spectrum of vision models” to llms and transformers) just because I can do it without setting up any api keys and cloud compute nonsense or having my laptop work itself to the heat-death of the universe on its puny i7.

Tempted to go into a cabin in the woods with zero internet and just constrain myself at programming, llm augmented sure, but completely autonomously on 70w of compute. Wild to think that this is where we are now, 4-5 years after copilot alpha.

#llms #llm #nvidia #dgx_spark

Ngl I have a nvidia dgx spark at work and it’s so much fun to have a little self contained box that is reasonably capable and fits into a backpack.

I’ve been running so many machine learning experiments (from “traditional learning” to “spectrum of vision models” to llms and transformers) just because I can do it without setting up any api keys and cloud compute nonsense or having my laptop work itself to the heat-death of the universe on its puny i7.

Tempted to go into a cabin in the woods with zero internet and just constrain myself at programming, llm augmented sure, but completely autonomously on 70w of compute. Wild to think that this is where we are now, 4-5 years after copilot alpha.

#llms #llm #nvidia #dgx_spark

Ngl I have a nvidia dgx spark at work and it’s so much fun to have a little self contained box that is reasonably capable and fits into a backpack.

I’ve been running so many machine learning experiments (from “traditional learning” to “spectrum of vision models” to llms and transformers) just because I can do it without setting up any api keys and cloud compute nonsense or having my laptop work itself to the heat-death of the universe on its puny i7.

Tempted to go into a cabin in the woods with zero internet and just constrain myself at programming, llm augmented sure, but completely autonomously on 70w of compute. Wild to think that this is where we are now, 4-5 years after copilot alpha.

#llms #llm #nvidia #dgx_spark