Hugging Face's logo

majentik
/
gemma-4-31B-RotorQuant-MLX-8bit

Image-Text-to-Text
MLX
Safetensors
gemma4
rotorquant
kv-cache-quantization
gemma
multimodal
quantized
8bit
8-bit precision
Model card Files
xet
 Community
majentik commited on
Commit
a4daafe
·
verified ·
1 Parent(s): 9f27870

fix: family model links majentik/gemma4-* -> majentik/gemma-4-* (reported by @shramee in gemma-4-31B-TurboQuant-MLX-4bit#1)

Browse files
Files changed (1) hide show
  1. README.md +18 -18
README.md CHANGED
@@ -115,25 +115,25 @@ This model requires approximately 31 GB of unified memory. Recommended hardware:
115
 
116
  ## Variants in this family
117
 
118
- (Showing 18 sibling variants under `majentik/gemma4-31b-*`. The current variant — `RotorQuant-MLX-8bit` — is **bolded**.)
119
 
120
  | Variant | Runtime | Approx size | Use case |
121
  |---|---|---|---|
122
- | [RotorQuant](https://huggingface.co/majentik/gemma4-31b-rotorquant) | runtime modifier | n/a | KV-cache root (weight-agnostic) |
123
- | [RotorQuant-AWQ-4bit](https://huggingface.co/majentik/gemma4-31b-rotorquant-awq-4bit) | transformers | ~19 GB | GPU 4-bit (AutoAWQ) |
124
- | [RotorQuant-AWQ-8bit](https://huggingface.co/majentik/gemma4-31b-rotorquant-awq-8bit) | transformers | ~34 GB | GPU 8-bit (AutoAWQ) |
125
- | [RotorQuant-GGUF-IQ4_XS](https://huggingface.co/majentik/gemma4-31b-rotorquant-gguf-IQ4_XS) | llama.cpp | ~27 GB | Lossy 4-bit, low-RAM CPU/edge |
126
- | [RotorQuant-GGUF-Q2_K](https://huggingface.co/majentik/gemma4-31b-rotorquant-gguf-Q2_K) | llama.cpp | ~19 GB | Lossy, low-RAM CPU/edge |
127
- | [RotorQuant-GGUF-Q3_K_M](https://huggingface.co/majentik/gemma4-31b-rotorquant-gguf-Q3_K_M) | llama.cpp | ~24 GB | Smaller 3-bit, CPU-friendly |
128
- | [RotorQuant-GGUF-Q4_K_M](https://huggingface.co/majentik/gemma4-31b-rotorquant-gguf-Q4_K_M) | llama.cpp | ~34 GB | Balanced default |
129
- | [RotorQuant-GGUF-Q5_K_M](https://huggingface.co/majentik/gemma4-31b-rotorquant-gguf-Q5_K_M) | llama.cpp | ~41 GB | Higher fidelity, more RAM |
130
- | [RotorQuant-GGUF-Q8_0](https://huggingface.co/majentik/gemma4-31b-rotorquant-gguf-Q8_0) | llama.cpp | ~65 GB | Near-lossless reference |
131
- | [RotorQuant-MLX-2bit](https://huggingface.co/majentik/gemma4-31b-rotorquant-mlx-2bit) | mlx-lm | ~9.9 GB | Apple Silicon, smallest |
132
- | [RotorQuant-MLX-4bit](https://huggingface.co/majentik/gemma4-31b-rotorquant-mlx-4bit) | mlx-lm | ~19 GB | Apple Silicon balanced |
133
  | **RotorQuant-MLX-8bit** | mlx-lm | ~37 GB | Apple Silicon reference |
134
- | [TurboQuant](https://huggingface.co/majentik/gemma4-31b-turboquant) | runtime modifier | n/a | KV-cache root (weight-agnostic) |
135
- | [TurboQuant-AWQ-4bit](https://huggingface.co/majentik/gemma4-31b-turboquant-awq-4bit) | transformers | ~19 GB | GPU 4-bit (AutoAWQ) |
136
- | [TurboQuant-AWQ-8bit](https://huggingface.co/majentik/gemma4-31b-turboquant-awq-8bit) | transformers | ~34 GB | GPU 8-bit (AutoAWQ) |
137
- | [TurboQuant-MLX-2bit](https://huggingface.co/majentik/gemma4-31b-turboquant-mlx-2bit) | mlx-lm | ~9.9 GB | Apple Silicon, smallest |
138
- | [TurboQuant-MLX-4bit](https://huggingface.co/majentik/gemma4-31b-turboquant-mlx-4bit) | mlx-lm | ~19 GB | Apple Silicon balanced |
139
- | [TurboQuant-MLX-8bit](https://huggingface.co/majentik/gemma4-31b-turboquant-mlx-8bit) | mlx-lm | ~37 GB | Apple Silicon reference |
 
115
 
116
  ## Variants in this family
117
 
118
+ (Showing 18 sibling variants under `majentik/gemma-4-31b-*`. The current variant — `RotorQuant-MLX-8bit` — is **bolded**.)
119
 
120
  | Variant | Runtime | Approx size | Use case |
121
  |---|---|---|---|
122
+ | [RotorQuant](https://huggingface.co/majentik/gemma-4-31b-rotorquant) | runtime modifier | n/a | KV-cache root (weight-agnostic) |
123
+ | [RotorQuant-AWQ-4bit](https://huggingface.co/majentik/gemma-4-31b-rotorquant-awq-4bit) | transformers | ~19 GB | GPU 4-bit (AutoAWQ) |
124
+ | [RotorQuant-AWQ-8bit](https://huggingface.co/majentik/gemma-4-31b-rotorquant-awq-8bit) | transformers | ~34 GB | GPU 8-bit (AutoAWQ) |
125
+ | [RotorQuant-GGUF-IQ4_XS](https://huggingface.co/majentik/gemma-4-31b-rotorquant-gguf-IQ4_XS) | llama.cpp | ~27 GB | Lossy 4-bit, low-RAM CPU/edge |
126
+ | [RotorQuant-GGUF-Q2_K](https://huggingface.co/majentik/gemma-4-31b-rotorquant-gguf-Q2_K) | llama.cpp | ~19 GB | Lossy, low-RAM CPU/edge |
127
+ | [RotorQuant-GGUF-Q3_K_M](https://huggingface.co/majentik/gemma-4-31b-rotorquant-gguf-Q3_K_M) | llama.cpp | ~24 GB | Smaller 3-bit, CPU-friendly |
128
+ | [RotorQuant-GGUF-Q4_K_M](https://huggingface.co/majentik/gemma-4-31b-rotorquant-gguf-Q4_K_M) | llama.cpp | ~34 GB | Balanced default |
129
+ | [RotorQuant-GGUF-Q5_K_M](https://huggingface.co/majentik/gemma-4-31b-rotorquant-gguf-Q5_K_M) | llama.cpp | ~41 GB | Higher fidelity, more RAM |
130
+ | [RotorQuant-GGUF-Q8_0](https://huggingface.co/majentik/gemma-4-31b-rotorquant-gguf-Q8_0) | llama.cpp | ~65 GB | Near-lossless reference |
131
+ | [RotorQuant-MLX-2bit](https://huggingface.co/majentik/gemma-4-31b-rotorquant-mlx-2bit) | mlx-lm | ~9.9 GB | Apple Silicon, smallest |
132
+ | [RotorQuant-MLX-4bit](https://huggingface.co/majentik/gemma-4-31b-rotorquant-mlx-4bit) | mlx-lm | ~19 GB | Apple Silicon balanced |
133
  | **RotorQuant-MLX-8bit** | mlx-lm | ~37 GB | Apple Silicon reference |
134
+ | [TurboQuant](https://huggingface.co/majentik/gemma-4-31b-turboquant) | runtime modifier | n/a | KV-cache root (weight-agnostic) |
135
+ | [TurboQuant-AWQ-4bit](https://huggingface.co/majentik/gemma-4-31b-turboquant-awq-4bit) | transformers | ~19 GB | GPU 4-bit (AutoAWQ) |
136
+ | [TurboQuant-AWQ-8bit](https://huggingface.co/majentik/gemma-4-31b-turboquant-awq-8bit) | transformers | ~34 GB | GPU 8-bit (AutoAWQ) |
137
+ | [TurboQuant-MLX-2bit](https://huggingface.co/majentik/gemma-4-31b-turboquant-mlx-2bit) | mlx-lm | ~9.9 GB | Apple Silicon, smallest |
138
+ | [TurboQuant-MLX-4bit](https://huggingface.co/majentik/gemma-4-31b-turboquant-mlx-4bit) | mlx-lm | ~19 GB | Apple Silicon balanced |
139
+ | [TurboQuant-MLX-8bit](https://huggingface.co/majentik/gemma-4-31b-turboquant-mlx-8bit) | mlx-lm | ~37 GB | Apple Silicon reference |