Spaces:

codelion
/

dhara-chat

Running on Zero

App Files Files Community

dhara-chat

Commit History

Jul 10

03:42
Update README.md Running

codelion

Verified signature
838ad76
03:34
Move to ZeroGPU: bf16 on cuda + @spaces.GPU on streaming handlers⁠

codelion

Verified signature
81c437a

May 26

15:26
Multi-turn UX: new-chat button; example click resets conversation; 1k-token sliding-window history; 4th example (sleep)

codelion

Verified signature
a284b48
15:09
INT8 dynamic quant (~2x CPU, verified all 3 modes); add self-spec to chat mode selector

codelion

Verified signature
873a061
14:58
Compare/chat prompts: drop count-based "three tips" (looks truncated at 48 tok); use self-contained short-answer prompts (ocean / neural net / exercise)

codelion

Verified signature
0249a53
14:49
Fix chat crash: gradio 6 Chatbot content is a list of parts; flatten to text before apply_chat_template

codelion

Verified signature
1ed429f
14:42
Fix gradio 6.14: gr.Chatbot dropped the type= arg (messages is default)

codelion

Verified signature
f411959
14:37
Custom chat layout: compact mode/max-tokens/temp controls above input; single-column examples (readable); header link to codelion/dhara-250m

codelion

Verified signature
df2112c
14:28
Disable example caching (HF Spaces defaults cache_examples=True) so example clicks generate live, not pre-baked

codelion

Verified signature
4681623
14:23
Coherence: rep-penalty 1.3 in diffusion/self-spec (no more collapse); clip to last full sentence (no mid-word cutoff); exposed config panel (mode/max-tokens/temperature); greedy AR in compare; better default prompts

codelion

Verified signature
7c0285b
13:51
Fix gradio 6 build: examples as list-of-lists (additional_inputs); move theme/css to launch()

codelion

Verified signature
cf976b5
13:42
Chat mode toggle (AR/diffusion w/ live ▓▒░ denoise); streaming 3-mode compare reveal; max_new 128 + no_repeat_ngram

codelion

Verified signature
732ea5f
13:35
Fixed-height scrollable chat; pass attention_mask (silence pad==eos warning); _enc helper

codelion

Verified signature
a36cfa9
13:30
Denoiser-terminal UI (mask-glyph aesthetic, amber CRT) + streaming per-mode compare with progress

codelion

Verified signature
32de77f
13:19
Link model (models: codelion/dhara-250m) for Spaces-using-this-model widget

codelion

Verified signature
3ecfd57
13:02
Add torch to requirements (gradio SDK base has no torch)

codelion

Verified signature
7638ff6
12:59
remove Dockerfile (switched to gradio SDK)

codelion

Verified signature
619b3a2
12:59
gradio SDK 6.14.0 (proven to import with hub 1.16) + transformers 5.5 + messages-format app

codelion

Verified signature
51b846a
12:42
Pin gradio==4.44.1 + use 4.x tuples ChatInterface API (model already loads)

codelion

Verified signature
1a232f4
12:31
Switch to Docker SDK; pip-resolved transformers==5.5.0 + gradio (fixes tokenizer/HfFolder)

codelion

Verified signature
a54a16a
12:10
Pin coherent stack (py3.12, gradio4.44, transformers<5, hub<1) to fix HfFolder/audioop; title ChatDhara

codelion

Verified signature
a113c18
11:59
Fix py3.13 gradio import: add audioop-lts (audioop removed in 3.13)

codelion

Verified signature
e2c8b3f
11:52
dhara-chat tri-mode CPU demo

codelion

Verified signature
78e793b
11:51
initial commit

codelion

Verified signature
bca5486

Commit History

Jul 10

Update README.md Running

Move to ZeroGPU: bf16 on cuda + @spaces.GPU on streaming handlers⁠

May 26

Multi-turn UX: new-chat button; example click resets conversation; 1k-token sliding-window history; 4th example (sleep)

INT8 dynamic quant (~2x CPU, verified all 3 modes); add self-spec to chat mode selector

Compare/chat prompts: drop count-based "three tips" (looks truncated at 48 tok); use self-contained short-answer prompts (ocean / neural net / exercise)

Fix chat crash: gradio 6 Chatbot content is a list of parts; flatten to text before apply_chat_template

Fix gradio 6.14: gr.Chatbot dropped the type= arg (messages is default)

Custom chat layout: compact mode/max-tokens/temp controls above input; single-column examples (readable); header link to codelion/dhara-250m

Disable example caching (HF Spaces defaults cache_examples=True) so example clicks generate live, not pre-baked

Coherence: rep-penalty 1.3 in diffusion/self-spec (no more collapse); clip to last full sentence (no mid-word cutoff); exposed config panel (mode/max-tokens/temperature); greedy AR in compare; better default prompts

Fix gradio 6 build: examples as list-of-lists (additional_inputs); move theme/css to launch()

Chat mode toggle (AR/diffusion w/ live ▓▒░ denoise); streaming 3-mode compare reveal; max_new 128 + no_repeat_ngram

Fixed-height scrollable chat; pass attention_mask (silence pad==eos warning); _enc helper

Denoiser-terminal UI (mask-glyph aesthetic, amber CRT) + streaming per-mode compare with progress

Link model (models: codelion/dhara-250m) for Spaces-using-this-model widget

Add torch to requirements (gradio SDK base has no torch)

remove Dockerfile (switched to gradio SDK)

gradio SDK 6.14.0 (proven to import with hub 1.16) + transformers 5.5 + messages-format app

Pin gradio==4.44.1 + use 4.x tuples ChatInterface API (model already loads)

Switch to Docker SDK; pip-resolved transformers==5.5.0 + gradio (fixes tokenizer/HfFolder)

Pin coherent stack (py3.12, gradio4.44, transformers<5, hub<1) to fix HfFolder/audioop; title ChatDhara

Fix py3.13 gradio import: add audioop-lts (audioop removed in 3.13)

dhara-chat tri-mode CPU demo

initial commit