Remove links to retired duplicate brand variants
Browse files
README.md
CHANGED
|
@@ -96,7 +96,6 @@ upstream options described above (`-ctk/-ctv q8_0`, `OLLAMA_KV_CACHE_TYPE`).
|
|
| 96 |
| FP16 (original) | ~4 GB | -- |
|
| 97 |
| **8-bit quantized** | **~2 GB** | **This model** |
|
| 98 |
| 4-bit quantized | ~1.2 GB | [RotorQuant-MLX-4bit](https://huggingface.co/majentik/gemma-4-E2B-RotorQuant-MLX-4bit) |
|
| 99 |
-
| 2-bit quantized | ~0.6 GB | [RotorQuant-MLX-2bit](https://huggingface.co/majentik/gemma-4-E2B-RotorQuant-MLX-2bit) |
|
| 100 |
|
| 101 |
## Hardware Requirements
|
| 102 |
|
|
@@ -111,8 +110,6 @@ This model requires approximately 2 GB of unified memory. Recommended hardware:
|
|
| 111 |
|
| 112 |
- [google/gemma-4-E2B](https://huggingface.co/google/gemma-4-E2B) -- Base model
|
| 113 |
- [majentik/gemma-4-E2B-RotorQuant-MLX-4bit](https://huggingface.co/majentik/gemma-4-E2B-RotorQuant-MLX-4bit) -- MLX 4-bit variant
|
| 114 |
-
- [majentik/gemma-4-E2B-RotorQuant-MLX-2bit](https://huggingface.co/majentik/gemma-4-E2B-RotorQuant-MLX-2bit) -- MLX 2-bit variant
|
| 115 |
-
- [majentik/gemma-4-E2B-TurboQuant-MLX-8bit](https://huggingface.co/majentik/gemma-4-E2B-TurboQuant-MLX-8bit) -- TurboQuant MLX 8-bit variant
|
| 116 |
- [RotorQuant GitHub](https://github.com/scrya-com/rotorquant)
|
| 117 |
- [MLX Framework](https://github.com/ml-explore/mlx)
|
| 118 |
|
|
@@ -141,9 +138,6 @@ This model requires approximately 2 GB of unified memory. Recommended hardware:
|
|
| 141 |
| [RotorQuant-GGUF-Q4_K_M](https://huggingface.co/majentik/gemma-4-e2b-rotorquant-gguf-Q4_K_M) | llama.cpp | ~2.2 GB | Balanced default |
|
| 142 |
| [RotorQuant-GGUF-Q5_K_M](https://huggingface.co/majentik/gemma-4-e2b-rotorquant-gguf-Q5_K_M) | llama.cpp | ~2.6 GB | Higher fidelity, more RAM |
|
| 143 |
| [RotorQuant-GGUF-Q8_0](https://huggingface.co/majentik/gemma-4-e2b-rotorquant-gguf-Q8_0) | llama.cpp | ~4.2 GB | Near-lossless reference |
|
| 144 |
-
| [RotorQuant-MLX-2bit](https://huggingface.co/majentik/gemma-4-e2b-rotorquant-mlx-2bit) | mlx-lm | ~655 MB | Apple Silicon, smallest |
|
| 145 |
| [RotorQuant-MLX-4bit](https://huggingface.co/majentik/gemma-4-e2b-rotorquant-mlx-4bit) | mlx-lm | ~1.2 GB | Apple Silicon balanced |
|
| 146 |
| **RotorQuant-MLX-8bit** | mlx-lm | ~2.4 GB | Apple Silicon reference |
|
| 147 |
| [TurboQuant-MLX-2bit](https://huggingface.co/majentik/gemma-4-e2b-turboquant-mlx-2bit) | mlx-lm | ~655 MB | Apple Silicon, smallest |
|
| 148 |
-
| [TurboQuant-MLX-4bit](https://huggingface.co/majentik/gemma-4-e2b-turboquant-mlx-4bit) | mlx-lm | ~1.2 GB | Apple Silicon balanced |
|
| 149 |
-
| [TurboQuant-MLX-8bit](https://huggingface.co/majentik/gemma-4-e2b-turboquant-mlx-8bit) | mlx-lm | ~2.4 GB | Apple Silicon reference |
|
|
|
|
| 96 |
| FP16 (original) | ~4 GB | -- |
|
| 97 |
| **8-bit quantized** | **~2 GB** | **This model** |
|
| 98 |
| 4-bit quantized | ~1.2 GB | [RotorQuant-MLX-4bit](https://huggingface.co/majentik/gemma-4-E2B-RotorQuant-MLX-4bit) |
|
|
|
|
| 99 |
|
| 100 |
## Hardware Requirements
|
| 101 |
|
|
|
|
| 110 |
|
| 111 |
- [google/gemma-4-E2B](https://huggingface.co/google/gemma-4-E2B) -- Base model
|
| 112 |
- [majentik/gemma-4-E2B-RotorQuant-MLX-4bit](https://huggingface.co/majentik/gemma-4-E2B-RotorQuant-MLX-4bit) -- MLX 4-bit variant
|
|
|
|
|
|
|
| 113 |
- [RotorQuant GitHub](https://github.com/scrya-com/rotorquant)
|
| 114 |
- [MLX Framework](https://github.com/ml-explore/mlx)
|
| 115 |
|
|
|
|
| 138 |
| [RotorQuant-GGUF-Q4_K_M](https://huggingface.co/majentik/gemma-4-e2b-rotorquant-gguf-Q4_K_M) | llama.cpp | ~2.2 GB | Balanced default |
|
| 139 |
| [RotorQuant-GGUF-Q5_K_M](https://huggingface.co/majentik/gemma-4-e2b-rotorquant-gguf-Q5_K_M) | llama.cpp | ~2.6 GB | Higher fidelity, more RAM |
|
| 140 |
| [RotorQuant-GGUF-Q8_0](https://huggingface.co/majentik/gemma-4-e2b-rotorquant-gguf-Q8_0) | llama.cpp | ~4.2 GB | Near-lossless reference |
|
|
|
|
| 141 |
| [RotorQuant-MLX-4bit](https://huggingface.co/majentik/gemma-4-e2b-rotorquant-mlx-4bit) | mlx-lm | ~1.2 GB | Apple Silicon balanced |
|
| 142 |
| **RotorQuant-MLX-8bit** | mlx-lm | ~2.4 GB | Apple Silicon reference |
|
| 143 |
| [TurboQuant-MLX-2bit](https://huggingface.co/majentik/gemma-4-e2b-turboquant-mlx-2bit) | mlx-lm | ~655 MB | Apple Silicon, smallest |
|
|
|
|
|
|