Instructions to use DeepNarra/Paham with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use DeepNarra/Paham with Transformers:

# Use a pipeline as a high-level helper
from transformers import pipeline

pipe = pipeline("text-generation", model="DeepNarra/Paham")

# pip install -U transformers accelerate
# Load model directly
from transformers import AutoModelForCausalLM
model = AutoModelForCausalLM.from_pretrained("DeepNarra/Paham", device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

llama.cpp

How to use DeepNarra/Paham with llama.cpp:

Install (macOS, Linux)

curl -LsSf https://llama.app/install.sh | sh
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf DeepNarra/Paham:TQ1_0
# Run inference directly in the terminal:
llama cli -hf DeepNarra/Paham:TQ1_0

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf DeepNarra/Paham:TQ1_0
# Run inference directly in the terminal:
llama cli -hf DeepNarra/Paham:TQ1_0

Use pre-built binary

# Download pre-built binary from:
# https://github.com/ggerganov/llama.cpp/releases
# Start a local OpenAI-compatible server with a web UI:
./llama-server -hf DeepNarra/Paham:TQ1_0
# Run inference directly in the terminal:
./llama-cli -hf DeepNarra/Paham:TQ1_0

Build from source code

git clone https://github.com/ggerganov/llama.cpp.git
cd llama.cpp
cmake -B build
cmake --build build -j --target llama-server llama-cli
# Start a local OpenAI-compatible server with a web UI:
./build/bin/llama-server -hf DeepNarra/Paham:TQ1_0
# Run inference directly in the terminal:
./build/bin/llama-cli -hf DeepNarra/Paham:TQ1_0

Use Docker

docker model run hf.co/DeepNarra/Paham:TQ1_0

Llama new
LM Studio
Jan

vLLM

How to use DeepNarra/Paham with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "DeepNarra/Paham"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "DeepNarra/Paham",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/DeepNarra/Paham:TQ1_0

SGLang

How to use DeepNarra/Paham with SGLang:

Install from pip and serve model

# Install SGLang from pip:
pip install sglang
# Start the SGLang server:
python3 -m sglang.launch_server \
    --model-path "DeepNarra/Paham" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "DeepNarra/Paham",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker images

docker run --gpus all \
    --shm-size 32g \
    -p 30000:30000 \
    -v ~/.cache/huggingface:/root/.cache/huggingface \
    --env "HF_TOKEN=<secret>" \
    --ipc=host \
    lmsysorg/sglang:latest \
    python3 -m sglang.launch_server \
        --model-path "DeepNarra/Paham" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "DeepNarra/Paham",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Ollama
How to use DeepNarra/Paham with Ollama:
```
ollama run hf.co/DeepNarra/Paham:TQ1_0
```
Unsloth Desktop
Docker Model Runner
How to use DeepNarra/Paham with Docker Model Runner:
```
docker model run hf.co/DeepNarra/Paham:TQ1_0
```

Lemonade

How to use DeepNarra/Paham with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull DeepNarra/Paham:TQ1_0

Run and chat with the model

lemonade run user.Paham-TQ1_0

List all available models

lemonade list

Atomic Chat

Deeplines commited on 1 day ago

Commit

3e22aa0

verified ·

1 Parent(s): 8045715

Set Paham sovereign root hierarchy & clean formatting

Browse files

Files changed (1) hide show

README.md +8 -9

README.md CHANGED Viewed

@@ -17,7 +17,6 @@ tags:
 - tq1_0
 - gguf
 - safetensors
-base_model: DeepNarra/Hiraya-1-Reasoning
 datasets:
 - HuggingFaceFW/fineweb-edu
 - CohereForAI/aya_collection
@@ -61,22 +60,22 @@ library_name: transformers
 | **Exact Safetensors Parameter Count** | **`4,171,887` parameters** (`15.92 MB` in `model.safetensors`, `60` tensors) |
 | **Native Ternary GGUF Binary** | **`0.81 MB`** (`model-TQ1_0.gguf` — BitNet b1.58 Ternary `TQ1_0` at `1.58 bits/weight`) |
 | **Elastic Active Compute / Token** | **Gear 1 (Reflex):** `~1.2M` ➔ **Gear 2 (Standard):** `~2.5M` ➔ **Gear 3 (Sovereign Deep):** `4.17M` |
-| **Full-Scale H-MoE Companion** | [`DeepNarra/Hiraya-1-Reasoning`](https://huggingface.co/DeepNarra/Hiraya-1-Reasoning) (`908,645,372` params / `0.91B`) |
 | **Measured Inference Throughput** | **`1,680.7 tokens/sec`** (`9.52 ms` step latency on Tesla T4) |
 | **Hybrid Attention Mechanism** | **75% Lyapunov GLA Recurrence + 25% GLM-5 IndexShare Dynamic Sparse Attention (DSA)** |
-| **Continuous Latent Deliberation** | **768-D Coconut Symplectic Phase Space** (Measured energy drift $\Delta E = 2.37 	imes 10^{-16}$) |
-| **Speculative Leap Engine** | Latent-Guided Multi-Token Prediction ($k=4$ MTP) + Prompt Lookup Decoding ($k=50$ PLD) |
-| **Lifelong Awake Learning** | Zero-downtime hot-swappable **GGUF-QLoRA Sidecar Adapters** ($B=0$, $\Delta = 0.0	ext{e}+00$ init drift) |
 ---
 ## ⚡ Distinct Architectural Innovations (Unique to Paham)
 1. **Shannon Entropy 3-Gear Elastic Router:**
-   - Computes token-level Shannon entropy $H(x) = -\sum p_i \log p_i$ on the fly to route compute elastically:
-     - **Gear 1 (`GEAR_1_MAKO_REFLEX`, $H < 0.8$):** Bypasses heavy latent loops for instant reflexive token emission (`1,680.7 tok/s`).
-     - **Gear 2 (`GEAR_2_NARRA_STEADY`, $0.8 \le H < 2.0$):** Engages recurrent Lyapunov GLA memory across medium-complexity sequences.
-     - **Gear 3 (`GEAR_3_HIRAYA_SOVEREIGN`, $H \ge 2.0$):** Triggers full 768-D Coconut continuous latent deliberation and symplectic verification.
 2. **Lyapunov-IndexShare Hybrid Attention (`O(1)` PahamStateCache):**
    - Blends Narra's chunk-wise Gated Linear Attention (`6.0 B/rec` constant state pool) with GLM-5 Dynamic Sparse Attention (DSA), maintaining $\Delta = 0.0	ext{e}+00$ numerical parity on standard contexts.

 - tq1_0
 - gguf
 - safetensors
 datasets:
 - HuggingFaceFW/fineweb-edu
 - CohereForAI/aya_collection
 | **Exact Safetensors Parameter Count** | **`4,171,887` parameters** (`15.92 MB` in `model.safetensors`, `60` tensors) |
 | **Native Ternary GGUF Binary** | **`0.81 MB`** (`model-TQ1_0.gguf` — BitNet b1.58 Ternary `TQ1_0` at `1.58 bits/weight`) |
 | **Elastic Active Compute / Token** | **Gear 1 (Reflex):** `~1.2M` ➔ **Gear 2 (Standard):** `~2.5M` ➔ **Gear 3 (Sovereign Deep):** `4.17M` |
+| **Sovereign Root Status** | **Independent Top-Level Sovereign Architecture** (Omniversal 3-Gear Synthesis) |
 | **Measured Inference Throughput** | **`1,680.7 tokens/sec`** (`9.52 ms` step latency on Tesla T4) |
 | **Hybrid Attention Mechanism** | **75% Lyapunov GLA Recurrence + 25% GLM-5 IndexShare Dynamic Sparse Attention (DSA)** |
+| **Continuous Latent Deliberation** | **768-D Coconut Symplectic Phase Space** (Measured energy drift `ΔE = 2.37e-16`) |
+| **Speculative Leap Engine** | Latent-Guided Multi-Token Prediction (`k=4` MTP) + Prompt Lookup Decoding (`k=50` PLD) |
+| **Lifelong Awake Learning** | Zero-downtime hot-swappable **GGUF-QLoRA Sidecar Adapters** (`B = 0`, `Δ = 0.0e+00` init drift) |
 ---
 ## ⚡ Distinct Architectural Innovations (Unique to Paham)
 1. **Shannon Entropy 3-Gear Elastic Router:**
+   - Computes token-level Shannon entropy `H(x) = -∑ p_i log(p_i)` on the fly to route compute elastically:
+     - **Gear 1 (`GEAR_1_MAKO_REFLEX`, `H < 0.8`):** Bypasses heavy latent loops for instant reflexive token emission (`1,680.7 tok/s`).
+     - **Gear 2 (`GEAR_2_NARRA_STEADY`, `0.8 ≤ H < 2.0`):** Engages recurrent Lyapunov GLA memory across medium-complexity sequences.
+     - **Gear 3 (`GEAR_3_HIRAYA_SOVEREIGN`, `H ≥ 2.0`):** Triggers full 768-D Coconut continuous latent deliberation and symplectic verification.
 2. **Lyapunov-IndexShare Hybrid Attention (`O(1)` PahamStateCache):**
    - Blends Narra's chunk-wise Gated Linear Attention (`6.0 B/rec` constant state pool) with GLM-5 Dynamic Sparse Attention (DSA), maintaining $\Delta = 0.0	ext{e}+00$ numerical parity on standard contexts.