Is there a way to set `max_tokens` on `runWithTools`?
Is there a way to set
max_tokensrunWithTools?
NWant to run resnet50 but for image embeddings, not classifications. Is this possible?
PDoes Cloudflare run some kind of moderation check on input to models?
I'm looking into using Workers AI to identify content which may be CSAM related, as unfortunately OpenAI's free moderation API was not sufficient for my use case and mostly gave me false negatives.
Hopefully the user-generated input I'll give to models won't trigger some kind of flag/check, so just want to make sure that my use case is alright.
I'm looking into using Workers AI to identify content which may be CSAM related, as unfortunately OpenAI's free moderation API was not sufficient for my use case and mostly gave me false negatives.
Hopefully the user-generated input I'll give to models won't trigger some kind of flag/check, so just want to make sure that my use case is alright.
AHello, I build a LLMs/AI chat platform build and design for Cloudflare worker AI. with many features like function calling, image to text, text to image, text to speech. all build with Cloudflare worker. https://github.com/akazwz/antonai
GitHub
a LLMs/AI chat platform build and design for Cloudflare worker AI. with many features like function calling, image to text, text to image, text to speech - akazwz/antonai
CIf you are looking for an Image-To-Text app, here is one built with Llava + SvelteKit https://floor-is-llava.pages.dev/ (code in the footer)
CWe do offer the LlamaGuard model: https://ai.meta.com/research/publications/llama-guard-llm-based-input-output-safeguard-for-human-ai-conversations/ Does this look like it might assist with what you are after?
We introduce Llama Guard, an LLM-based input-output safeguard model geared towards Human-AI conversation use cases. Our model incorporates a safety risk...
U?
RNo. You would have to modify runWithTools to be able to do that, e.g. https://github.com/cloudflare/ai-utils/compare/main...Raylight-PWL:ai-utils:Raylight-patch-maxTokens-option.
U:frown:
UI think a good idea would be to have either max_tokens be the max tokens for each call, or have the user able to configure the max tokens for each call
Uthe second option being something like initial_max_tokens, tool_max_tokens, and possibly trim_max_tokens (very scuffed index names, but you get the point)
UIts been really good
UUon normal CF ai i had to guess so much 
Uai-utils is definitely awesome in comparison
UxD
Ui once tried to read the actual AI library on npm to figure out unknowns
UNever trying that again!
Ucompiled code :salute:
Unp!! :D
UI have a quick question
Uconsidering you do ai-utils shenanigans
Uare there any plans on tool usage being supported on the qwen models?
Ufrom what I read, the actual model itself supports it, but I'm obviously not entirely sure on it
UIt's not documented as so, and when I tried I could not get it to work haha
Uthen again I could also have been doing it wrong.
UNo, it just never actually ran the tool code
Ubut I also couldn't get it to work on the documented model so I was probably doing it wrong
Uone secaroo
Uif i can find the project within the mess of my projects folder...
UUIt's supposed to run the tool when the user sends "Follow me" or "Come here" etc etc
Unot sure what I did wrong :P
U(my code would parse the string after-the-fact)
UAIModel is defined as
@cf/qwen/qwen1.5-14b-chat-awq, ai is env.aiUI can dm you my worker ID if needed
