CMSManhattan / text-generation updated 1 day ago

JiRackUltra_14b

A fast and efficient 14B model optimized for CPU inference. The model was refactored with BitNet features and an updated tokenizer that includes new Routing, Media, Vision, Sound, Tool call, and Robotics tags. Built on a DeepSeek R1-14B architecture with native ternary (BitNet-style) support and ready-to-run GGUF quantizations. - JiRack is a cloud-ready model that helps save money on cloud infrastructure. It can be u...

Params
14.8 B
Context
131,072
Downloads 30d
1.0 M
Likes
2
Commercial use: allowed mit Not gated GGUF · SAFETENSORS 13 languages View on Hugging Face ↗

Download history

daily snapshots · 8 days
1.0 M229 K
Sep 14Sep 16Sep 19Sep 21

Can you run it?

Estimated VRAM at 8K context unless noted. Pick your hardware to see the verdict per quantization.

FileQuantSizeEst. VRAMVerdict on RTX 4090 · 24 GB
JiRackUltra_14b_Q2_K.gguf Q2_K 5.8 GB 9.1 GB ✅ Runs comfortably
JiRackUltra_14b_Q3_K_M.gguf Q3_K_M 7.3 GB 10.8 GB ✅ Runs comfortably
JiRackUltra_14b_Q4_K_M.gguf Q4_K_M 9.0 GB 12.6 GB ✅ Runs comfortably
JiRackUltra_14b.Q5_K_M.gguf Q5_K_M 10.5 GB 14.3 GB ✅ Runs comfortably
JiRackUltra_14b.Q6_K.gguf Q6_K 12.1 GB 16.1 GB ✅ Runs comfortably
JiRackUltra_14b.Q8_0.gguf Q8_0 15.7 GB 20.0 GB ⚠️ Tight — reduce context
JiRackUltra_14b.gguf GGUF 29.5 GB 35.2 GB ❌ Won’t fit
model.safetensors bf16 29.5 GB 35.2 GB ❌ Won’t fit
model.safetensors (bf16, full) bf16 + 128K ctx 29.5 GB 68.4 GB ❌ Won’t fit
Estimate: file size × 1.1 + KV cache at 8K + 0.5 GB overhead. Not a benchmark — how we calculate this.

Run it

copy-paste, exact tags checked against the Hub
~ · ollama · Q4_K_M
$ ollama run jirackultra:14b

# pin the quantization explicitly
$ ollama run jirackultra:14b-q4_k_m
est. VRAM 12.6 GBon RTX 4090 · 24 GBJSON API →

Specifications

Architecture
Qwen2ForCausalLM
Parameters
14.8 B
Tensor type
BF16
Context length
131,072
Vocabulary
152,064
Layers / heads
48 / 40
Licence
mit
First seen on the Hub
2026-08-03
Training datasets
undisclosed
Added to our catalog
2026-09-14
Compare with any text-generation model