Add ollama #687
Add ollama #687
Conversation
boxabirds
requested review from
osanseviero,
SBrandeis,
gary149,
Wauplin and
julien-c
as code owners
|
Thanks for the PR! Unfortunately, I don't think this would work as |
|
Right so the deeplink doesn’t work but the other fields should be ok.
…On Mon, 20 May 2024 at 15:10, Omar Sanseviero ***@***.***> wrote:
Thanks for the PR! Unfortunately, I don't think this would work as ollama
re-hosts the models themselves. E.g. an example url generated by your PR is
https://ollama.com/library/meta-llama/Meta-Llama-3-8B which is a 404 on
ollama
—
Reply to this email directly, view it on GitHub
<#687 (comment)>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/AABD62IWWTG5TSHKJPWNAFTZDH735AVCNFSM6AAAAABH7VXB2OVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDCMRQGU2DCNJWGA>
.
You are receiving this because you authored the thread.Message ID:
***@***.***>
|
|
The only goal of this button is to work with deeplinks. The button serves no other purpose |
There was a problem hiding this comment.
Hi @boxabirds - Just a heads-up we are thinking about this a bit more in our current sprint. We might have some exciting updates soon! 🤗
|
i thinik we could revive this PR but switch to a copy-pasteable command (like llama.cpp) rather than a deeplink |
|
@boxabirds ok for you if @mishig25 from our team takes over your PR and completes it? |
Very popular local LLM inference server.
|
Ready to be reviewed Order of operations
|
| @@ -389,6 +399,13 @@ export const LOCAL_APPS = { | |||
| displayOnModelPage: (model) => model.library_name === "diffusers" && model.pipeline_tag === "text-to-image", | |||
| deeplink: (model) => new URL(`https://models.invoke.ai/huggingface/${model.id}`), | |||
| }, | |||
| ollama: { | |||
| prettyLabel: "Ollama", | |||
| docsUrl: "https://ollama.com", | |||
There was a problem hiding this comment.
Probably have dedicated docs on the hub for this, but we can make a follow-up PR later on
Moving this function from moon to hf.js/gguf as I need it for #687 ```ts const quantLabel = parseGGUFQuantLabel("abc-Q4.gguf") console.log(quantLabel) // Q4 ``` ### Order of operations - [ ] merge #967 (this PR) & deploy `@hf.js/gguf` - [ ] merge #687 & deploy `@hf.js/tasks` - [ ] merge moon --------- Co-authored-by: Xuan Son Nguyen <[email protected]> Co-authored-by: Julien Chaumond <[email protected]>
|
@coyotte508 could you help me with correctly importing hfjs/gguf inside hfjs/tasks? However, the tests are failing & |
you can add (but not thrilled to add a dep to Also #887 woudl have fixed it maybe (no need for "prepare" anymore) |
im guessing we could merge tasks and gguf... |
As suggested in #687 (comment)
Very popular local LLM inference server.