Shawn Reed
@shekari.bsky.social
Self proclaimed good guy with some bad habits including Porsche, Ducati & sugar. Also very lucky to have @sparklyseaster.bsky.social in my life.
I think cooling towers are typically evaporative
November 3, 2025 at 8:19 PM
I think cooling towers are typically evaporative
Also contaminated water through cooling towers means contaminated air
November 3, 2025 at 7:56 PM
Also contaminated water through cooling towers means contaminated air
Cost. How do you get it where you need it.
November 3, 2025 at 7:55 PM
Cost. How do you get it where you need it.
Start with an rc2014 and build it first.
October 27, 2025 at 4:59 AM
Start with an rc2014 and build it first.
+1 on shell preference as well.
October 25, 2025 at 8:07 AM
+1 on shell preference as well.
I literally woke up and thought I can't wait to get back to testing multi-modal embeddings today. #Memgraph #MCP #llama.cpp
October 15, 2025 at 3:50 PM
I literally woke up and thought I can't wait to get back to testing multi-modal embeddings today. #Memgraph #MCP #llama.cpp
I can’t make this new tool work, it must be the tools fault.
October 10, 2025 at 12:29 AM
I can’t make this new tool work, it must be the tools fault.
Open source that can be audited and consensus between a few different models would be a start.
October 8, 2025 at 3:26 PM
Open source that can be audited and consensus between a few different models would be a start.
Dedicated button, no. Specific chip to support? Yes and proof. bsky.app/profile/shek...
After a day of tuning and testing models, another variable has entered the fray. #AMD #RYZENAIMAX #GMKtec
September 29, 2025 at 4:48 PM
Dedicated button, no. Specific chip to support? Yes and proof. bsky.app/profile/shek...
Some incredible results after tuning gpt-oss a bit. an FP16 GGUF Thanks Unsloth! with a 320k context all in VRAM and >100 tok/sec! #GPT-OSS #llama.cpp #AMD #ROCM
September 28, 2025 at 8:55 PM
Some incredible results after tuning gpt-oss a bit. an FP16 GGUF Thanks Unsloth! with a 320k context all in VRAM and >100 tok/sec! #GPT-OSS #llama.cpp #AMD #ROCM
I also found that ROCm 6.4 was way faster on prompt processing but Vulkan as a bit faster on the streaming speed so Gemini Flash helped me determine the break even point.
September 28, 2025 at 6:59 PM
I also found that ROCm 6.4 was way faster on prompt processing but Vulkan as a bit faster on the streaming speed so Gemini Flash helped me determine the break even point.
I think I will add in a gpt-oss and produce my own IQ4_NL quant of the Dolphin model as I did for the Gemma 3.
September 28, 2025 at 6:56 PM
I think I will add in a gpt-oss and produce my own IQ4_NL quant of the Dolphin model as I did for the Gemma 3.
Will try matching the kv cache quants to the model quant and some more common like q8_0 next.
September 28, 2025 at 4:27 PM
Will try matching the kv cache quants to the model quant and some more common like q8_0 next.
