Spaces:
Sleeping
Sleeping
Commit History
-
Max tokens
0ffe6c3 -
Dynamic max tokens
7e29a25 -
Refactor Icons
a979074 -
Advanced generation options (#30)
Unverified signature18f6e61
-
label: 'Max Tokens'
14e362d -
fix step
d91b043 -
Improve config: add top_p, top_k, repetition_penalty and their correct ranges (#29)
Unverified signatured5e14b5 -
No need for double-try
5f94ff7 -
put on key down in correct place
3b5cc6c -
Improve inference functions (#28)
Unverified signature573aa88 -
Settings dynamically change (#27)
Unverified signature86574c0 -
fix undefined
e9a4671 -
Check if model support system prompt (#26)
Unverified signaturedd66861 -
format
35f4e45 -
Rm jsonMode
413bc00 -
Fetch compatible models & their tokenizer files server side (#25)
Unverified signature7956c78 -
Stop deleteing messages in pairs automatically (#24)
Unverified signatureIf a user breaks chat template by deleting messages erroneously, then be ite779e06 -
Restructure to have `InferencePlayground.svelte` component (#23)
Unverified signatureRestructure to have InferencePlayground.svelte component so that it will be easier to move this into moon-landing1189124
-
Replace `process.write` with `console.log` (#22)
Unverified signaturece34835
-
fix placeholder msg in code snippets
78b198f -
Support bash snippets (#21)
Unverified signaturef1b84d3 -
Support python snippets (#20)
Unverified signature635c529 -
Render current messages in snippets (#19)
Unverified signaturee08d702 -
Improve JS snippet (#18)
Unverified signature05476c1 -
Feature: snippets tab for python & curl working (#17)
Unverified signaturec02e352 -
Improve CodeSnipetts.svelte (#16)
Unverified signature79d3242 -
Use highlight.js (#15)
Unverified signature877b369 -
Support config options when running models in comparison (#14)
Unverified signature3d94876
-
format
eb79b3d -
Feature: chat with two models simultaneously (#11)
Unverified signaturef2e5687 -
format
5213b80 -
Created the needed Conversation.svelte component (#10)
Unverified signature502cb81 -
Rm unused ModelSelectorV2.svelte (#9)
Unverified signaturebb96856 -
Remove Playground.svelte container component & just use +page.svelte (#8)
Unverified signature4323d20 -
Rename componentns to better names (#7)
Unverified signature5705707 -
Rm commented out section about PRO status & model deployment (#6)
Unverified signature8baea3c
-
Improve messages UX (#5)
Unverified signature# Improve messages UX When sending messages to an assistant, the messages has to alternate in user/assistant pairs. In the inference-playground, we allow editing or deleting messages, which can break this "alternate in user/assistant pairs" format. This PR fixes the issue65cc8c1 -
better word
299cbc7 -
pulsating animation for non-streaming messages to give feedback to the user
38c4650 -
during generation, messages should not be edited
8cb08ea -
red color for abort btn
62d3ef2 -
ability to abort non-streaming messages too
5143227 -
Same shortcut as other chatbots
e531b1e -
Sort models lexiographically
d914174 -
Use type `ChatCompletionInputMessage` from hfjs/tasks
38434c2 -
List vurrently running conversational models from hub
3138e12
-
sticky
a5e3e9f -
tweaks
42614e7 -
dark
5037fc5 -
testing stuff
781a165