do workers stay online after a request is made to handle further requests or do they get shutdown in
do workers stay online after a request is made to handle further requests or do they get shutdown instantly?
FUsually workers exit when they have done their work
I^ they can stay online for a little bit but different users will likely hit different servers (meaning they won't hit the same Worker) and Workers are not guaranteed to stay online at all.
ACan we please get... Qwen2.5-Coder-32B-Instruct on cloudflare AI?
SCan anyone make LoRA adapters work? I'm constantly getting
{"errors":[{"message":"AiError: AiError: Unknown internal error","code":3028}],"success":false,"result":{},"messages":[]}Cloudflare Docs
Cloudflare offers a few public LoRA adapters that are immediately ready for use.
�the ony one supports function: @hf/nousresearch/hermes-2-pro-mistral-7b so terrible?
it even retrieves kind of tool_calls which is not in pre-defined list
it even retrieves kind of tool_calls which is not in pre-defined list
�
CNews for multilingual embedding model ?
STried custom lora on gemma-2b too, same error
FCloudflare Workers AI + WordPress, I'm sharing this WordPress plugin I've created to use the CF REST API and enable webchat. Feedback, suggestions, and improvements are welcome! :MeowHeartCloudflare:
https://github.com/fernandodilland/workers-ai-chat
https://github.com/fernandodilland/workers-ai-chat
GitHub
Uses Cloudflare Workers AI in a web chat embedded via shortcode. - fernandodilland/workers-ai-chat
Aany news on this?
SSeems like there's no solution found online, but I'm sure there're working hard on that. Let's wait patiently.
Acan't wait on the updates 
seems like you guys are working on a lot of updates to AI 
seems like you guys are working on a lot of updates to AI I
this is quite awesome, but also interesting timing given Whisper V3 Turbo just got released with quality matching large-v2baseIf you're targeting translation though that makes sense, I think Turbo is only for direct transcription
Chmm guessing it's still cooking? It throws
"message": "5006: must be string, 'audio' must be array, must match exactly one schema in oneOf"Oh awesome 
Coo cool, working now!
Kare beta models still free to use? even image models? I think they used to be that way
KI'm also running into the issue where flex-1-schnell has an overactive nsfw filter.
"cyberpunk cat"Just a heads up, but @cf/google/gemma-7b-it-lora is returning mostly blank responses for me regardless of prompt.
Kwould love for
flex-1-schnellthe model on Replicate, has a megapixel param
LIs there any plan to support prompts for Whisper (/ are they supported currently just not documented)?
AI love how you guys can leak stuff unofficially haha! Cloudflare has a great culture man!
Awith a company the size of cloudflare... no one likes everything. You are doing good work by leaking haha 
THello, appears that @cf/meta/m2m100-1.2b is down or consistently returning 429s anyway, failing a bunch of workers I have
�this GA model is a night mare, unstable and truncate input unexpectedly
TYeah I would love a couple of other translating and embedding models in CF but everything seems to be about text and image generation at the moment
SHi
i am stuck with this error can anyone help me
Tensor error: failed to build tensor image: Tensor error: Unknown internal error: failed to decode u8
code:
app.post("/ai-process", async (c) => {
// Parse the form data from the request
const formData = await c.req.parseBody();
console.log("1", formData);
const text = formData.text || "Default text";
const image = formData.image;
console.log("2", image);
if (!image) {
return c.json({ error: "No image uploaded or invalid format" }, 400);
}
const imageArrayBuffer = await image.arrayBuffer();
const u8 = new Uint8Array(imageArrayBuffer);
// Prepare the AI request
const aiRequest = {
messages: [
{
role: "user",
content: text, // e.g., "What is the purchase price?"
},
],
image: Buffer.from(u8).toString("base64"),
};
console.log("3", aiRequest);
try {
// Call the AI API
const results = await c.env.AI.run(
"@cf/meta/llama-3.2-11b-vision-instruct",
aiRequest
);
console.log("4", results);
return c.json({
message: "Form processed successfully!",
response: results.response,
});
} catch (error) {
return c.json(
{
error:
},
500
);
}
});
i am stuck with this error can anyone help me
Tensor error: failed to build tensor image: Tensor error: Unknown internal error: failed to decode u8
code:
app.post("/ai-process", async (c) => {
// Parse the form data from the request
const formData = await c.req.parseBody();
console.log("1", formData);
const text = formData.text || "Default text";
const image = formData.image;
console.log("2", image);
if (!image) {
return c.json({ error: "No image uploaded or invalid format" }, 400);
}
const imageArrayBuffer = await image.arrayBuffer();
const u8 = new Uint8Array(imageArrayBuffer);
// Prepare the AI request
const aiRequest = {
messages: [
{
role: "user",
content: text, // e.g., "What is the purchase price?"
},
],
image: Buffer.from(u8).toString("base64"),
};
console.log("3", aiRequest);
try {
// Call the AI API
const results = await c.env.AI.run(
"@cf/meta/llama-3.2-11b-vision-instruct",
aiRequest
);
console.log("4", results);
return c.json({
message: "Form processed successfully!",
response: results.response,
});
} catch (error) {
return c.json(
{
error:
AI processing failed: ${error.message},},
500
);
}
});
Sam I crazy or is
@cf/qwen/qwen1.5-14b-chat-awq returning blank responsesSworks with other models
SSeems to work 90% of the time, sometimes it still returns blank responses
SNot too sure why or how to reproduce it consistently
