Spaces:
Sleeping
Sleeping
MaduRox commited on
Commit ·
bdef46f
1
Parent(s): 1307aaa
docs: remove studio link and add paradigm info to API tab
Browse files
app.py
CHANGED
|
@@ -151,8 +151,7 @@ def run_benchmark_endpoint(context_lengths_str: str = "128,256,512,1024,2048") -
|
|
| 151 |
with gr.Blocks(title="Kalpanā API — ZeroGPU NVIDIA A100", theme=gr.themes.Soft()) as demo:
|
| 152 |
gr.Markdown(
|
| 153 |
"# ⚡ Kalpanā RIF O(1) Memory API & Inference Server\n"
|
| 154 |
-
"Production constant-memory neural inference powered by **NVIDIA GPU** (dedicated).
|
| 155 |
-
"👉 **Visual Studio Frontend:** [Kalpana RIF Studio](https://huggingface.co/spaces/MaduRox/Kalpana-RIF-Studio)"
|
| 156 |
)
|
| 157 |
with gr.Tabs():
|
| 158 |
with gr.TabItem("▶ Interactive Test Console"):
|
|
@@ -306,6 +305,8 @@ with gr.Blocks(title="Kalpanā API — ZeroGPU NVIDIA A100", theme=gr.themes.Sof
|
|
| 306 |
"""
|
| 307 |
### 🌐 Public REST API Specification
|
| 308 |
|
|
|
|
|
|
|
| 309 |
You can query this Kalpanā RIF Neural Engine programmatically from **cURL, PowerShell, Python, or JavaScript**.
|
| 310 |
|
| 311 |
---
|
|
|
|
| 151 |
with gr.Blocks(title="Kalpanā API — ZeroGPU NVIDIA A100", theme=gr.themes.Soft()) as demo:
|
| 152 |
gr.Markdown(
|
| 153 |
"# ⚡ Kalpanā RIF O(1) Memory API & Inference Server\n"
|
| 154 |
+
"Production constant-memory neural inference powered by **NVIDIA GPU** (dedicated)."
|
|
|
|
| 155 |
)
|
| 156 |
with gr.Tabs():
|
| 157 |
with gr.TabItem("▶ Interactive Test Console"):
|
|
|
|
| 305 |
"""
|
| 306 |
### 🌐 Public REST API Specification
|
| 307 |
|
| 308 |
+
> **Paradigm B (External RIF Cache Injection)**: The REST API exposes the exact same generation loop as the Interactive Test Console. It natively loads a standard LLM and intercepts its KV generation via our external RIF memory engine to provide O(1) memory during inference streaming.
|
| 309 |
+
|
| 310 |
You can query this Kalpanā RIF Neural Engine programmatically from **cURL, PowerShell, Python, or JavaScript**.
|
| 311 |
|
| 312 |
---
|