Qwen2.5-VL 7B: how much RAM do you need?
Qwen2.5-VL 7B: butuh RAM berapa?
About 4.6 GB for the weights at 4-bit, and about 6.1 GB in total with an 8K context (7.4 GB at 32K). Numbers are estimates.
Sekitar 4.6 GB untuk bobot pada 4-bit, dan sekitar 6.1 GB total dengan konteks 8K (7.4 GB pada 32K). Angkanya perkiraan.
Specs
Spesifikasi
| Parameters | Parameter | 8.3B |
|---|---|---|
| Layers / KV heads | Layer / kepala KV | 28 / 4 |
| Native context | Konteks asli | 125K tokens |
| License | Apache 2.0 |
Memory by context length
Memori menurut panjang konteks
| Context | Konteks | KV cache | Q4 | Q8 | FP16 |
|---|---|---|---|---|---|
| 4K | 0.2 GB | 5.9 GB | 9.4 GB | 17 GB | |
| 8K | 0.4 GB | 6.1 GB | 9.6 GB | 17 GB | |
| 16K | 0.9 GB | 6.5 GB | 10 GB | 17 GB | |
| 32K | 1.8 GB | 7.4 GB | 11 GB | 18 GB | |
| 64K | 3.5 GB | 9.1 GB | 13 GB | 20 GB |
Totals include the weights, the KV cache and about 1 GB of runtime overhead. The KV cache column shows the memory the context adds. Real use varies with the app and settings.
Total mencakup bobot, KV cache, dan sekitar 1 GB beban runtime. Kolom KV cache menunjukkan memori tambahan dari konteks. Pemakaian nyata bergantung pada aplikasi dan pengaturan.
Good to know
Perlu diketahui
- Reads images as well as text.
- Bisa membaca gambar selain teks.
Source: Hugging Face model card · checked 2025-04-06 (model files), page built from the model's configdicek 2025-04-06 (berkas model), halaman dibuat dari konfigurasi model