dmatora

Results 101 comments of dmatora

Yeah, it's also an issue with QwQ model, particularly when Void tries to update the file, code is mixed with thinking output, which breaks the app

Until firecrawl can function self hosted this doesn't make much sense. Right now firecrawl requires expensive scraping APIs to be able to make Google searches, which costs more than firecrawl...

@eric09230 Yeah, we spoke with @thadius83 about it here https://github.com/mendableai/firecrawl/issues/1140#issuecomment-2647882020

are you using '--tensor-parallel-size 8'? 17840 seems to be small (at least for 70b) I basically have same question. I was getting `ValueError: The model's max seq len (131072) is...

Well ollama got KV cache quantisation support and Qwen released 2.5 models, which 32B outsmarting Llama 3.1 70B so now I don't need multiple A100 to get GPT4 brain level,...

What about `TEMPLATE` section for DeepSeek Modelfile? Doesn't it need one?

Ok, found a tip at issue #8571 ``` FROM merged_file.gguf TEMPLATE """{{- if .System }}{{ .System }}{{ end }} {{- range $i, $_ := .Messages }} {{- $last := eq...

@nickscamara message.toolInvocations[0].result.data.analysis is NOT supposed to be visible for the user right? I am currently forcing it to be shown, just so that I can see some sort of research,...