Instructions to use amgadtewfik/atf with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- MLX
How to use amgadtewfik/atf with MLX:
# Download the model from the Hub pip install huggingface_hub[hf_xet] hf download amgadtewfik/atf --local-dir atf
- Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- LM Studio
- Atomic Chat
Update README.md
Browse files
README.md
CHANGED
|
@@ -36,11 +36,12 @@ That's within reach of the M4's ~84 GB/s memory-bandwidth ceiling — these
|
|
| 36 |
formats are now bandwidth-bound, not kernel-bound, on Apple hardware.
|
| 37 |
|
| 38 |
## Models
|
| 39 |
-
| Model |
|
| 40 |
-
|---|---|---
|
| 41 |
-
| Qwen3.5-9B
|
| 42 |
-
| Qwen3.8-27B
|
| 43 |
-
| Qwen-AgentWorld-35B-A3B |
|
|
|
|
| 44 |
|
| 45 |
*Benchmarks: Apple M4, 16 GB unified memory, greedy decode, 4k context, warm
|
| 46 |
pass. All three run comfortably on a 16 GB Mac.*
|
|
|
|
| 36 |
formats are now bandwidth-bound, not kernel-bound, on Apple hardware.
|
| 37 |
|
| 38 |
## Models
|
| 39 |
+
| Model | Size | Prefill | Decode |
|
| 40 |
+
|---|---|---|---|
|
| 41 |
+
| Qwen3.5-9B-IQ4_NL.atf | 5.52 GB | 65.00 tok/s (274 tok, 4.20s) | 14.04 tok/s (104 tok, 7.40s) |
|
| 42 |
+
| Qwen3.8-27B-UD-Q2_K_XL.atf | 9.52 GB | 18.90 tok/s (274 tok, 14.50s) | 4.97 tok/s (96 tok, 19.30s) |
|
| 43 |
+
| Qwen-AgentWorld-35B-A3B-UD-IQ2_M.atf | 11.64 GB | 24.30 tok/s (274 tok, 11.30s) | 6.93 tok/s (105 tok, 15.20s) |
|
| 44 |
+
|
| 45 |
|
| 46 |
*Benchmarks: Apple M4, 16 GB unified memory, greedy decode, 4k context, warm
|
| 47 |
pass. All three run comfortably on a 16 GB Mac.*
|