nodes/asaddi-ComfyUI-YALLM-node
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
59 Commits 2 Branches 8 Tags
master
Commit Graph
29 Commits
Author SHA1 Message Date
Allan Saddi 85423e6600 fix: Also explicitly request no streaming. 
From experience elsewhere, there seem to be some providers that default
to streaming.

Also update dependencies in pyproject.toml to match requirements.txt.
 2026-10-14 17:49:47 -07:00
Allan Saddi 033d1ac2a3 refactor: Drop openai package dependency. 
Just use `requests` directly. May not be as robust, but I'll fix it as
problems come up.

Works for me.
 2026-10-14 15:59:34 -07:00
Allan Saddi 6b8d59999e Sometimes remote APIs return no choices... 2025-08-09 15:45:10 -07:00
Allan Saddi e3d782ad80 Text Latch: Use my common utils, ensure text is saved. 2024-11-29 22:08:12 -08:00
Allan Saddi 8eaee8f69b Add Prepend/Append node. 2024-11-21 02:36:10 -08:00
Allan Saddi 6617531142 Text Latch: Make input lazy. For real. 2024-11-16 02:24:47 -08:00
Allan Saddi a61cd0ac51 Revert "Text Latch: Make input lazy." 
This reverts commit 40db9f5a39.

Apparently, saying a lazy input is needed makes it required even if
it's defined as optional.
 2024-11-16 01:22:34 -08:00
Allan Saddi 40db9f5a39 Text Latch: Make input lazy. 2024-11-16 00:00:20 -08:00
Allan Saddi b286b92eab Values of the "ui" dictionary need to be iterable. 
Discovered while working on other nodes, explains weird necessity for
"join."
 2024-11-06 23:43:31 -08:00
Allan Saddi f1d61967cc Initial ruff pass 2024-11-05 02:19:01 -08:00
Allan Saddi d2c3263d8f Text Latch: Add a "replace" switch. 
Also update all examples and images using the new Text Latch.
 2024-11-03 15:21:48 -08:00
Allan Saddi 17751018d5 Forgot to mark "Text Latch" as an output node. 2024-11-03 00:43:04 -07:00
Allan Saddi 7c7f6f5e1d Add "Text Latch" node. 
Basically, "Show Text" that remembers its last input.
 2024-11-02 23:54:42 -07:00
Allan Saddi b08648b09b Add "LLM Provider" node which dynamically fetches models. 2024-10-23 18:46:29 -07:00
Allan Saddi abeba47ece Bugfix: Docs for "os.environ/" did not match reality. 
Also ensure LLM Model node is re-executed when associated config changes.
 2024-10-23 04:11:37 -07:00
Allan Saddi 561d2d58f5 Minor refactor of ModelDefinition/Models methods. 2024-10-17 04:59:37 -07:00
Allan Saddi 922c10a011 Add support for prompting with an image. 2024-10-06 00:04:13 -07:00
Allan Saddi ff479d9d1f Conditionally refresh models.yaml based on mtime. 2024-10-02 12:01:50 -07:00
Allan Saddi 92f24e2172 Warn if models.yaml doesn't exist and use default. 2024-10-02 11:36:29 -07:00
Allan Saddi 3c2a3aa36b More clean up, especially around the seed. 2024-09-29 02:58:16 -07:00
Allan Saddi 49d37f79b8 Fix output bug due to model node refactor. 2024-09-28 23:41:11 -07:00
Allan Saddi 283c0a0a77 Add README, LICENSE, sample workflow, etc. 2024-09-27 16:55:55 -07:00
Allan Saddi ac17a8cbf2 Once again, split model into its own node. 
Also some minor cleanup/consistency.
 2024-09-27 15:42:02 -07:00
Allan Saddi 246e9e45e0 Also support specifying sampler order, since it's so trivial. 2024-09-17 16:55:52 -07:00
Allan Saddi e19649e093 Add nodes for basic samplers (including temperature). 2024-09-17 16:43:15 -07:00
Allan Saddi 8d0932733c Read LLM model information from a config file. 
Now we're down to only 1 custom node...
 2024-09-15 23:07:52 -07:00
Allan Saddi 4808631eb3 A little saner model handling. 
Get rid of LLMPrompt as all we really needed was to make the inputs multiline.
No longer force user_prompt/system_prompt to be inputs, let user decide.
 2024-09-12 17:08:16 -07:00
Allan Saddi 83d7edea05 Factor out LLM config so we can fetch models only when there's a change. 2024-09-12 16:28:10 -07:00
Allan Saddi ccad2ee3f9 Move node titles to the classes themselves. 
Rename main file to match what I see in ComfyUI & elsewhere.
 2024-09-09 22:26:03 -07:00