Spaces:
Sleeping
Sleeping
Update requirements
Browse files- requirements +0 -10
requirements
CHANGED
|
@@ -1,15 +1,5 @@
|
|
| 1 |
-
# KaLM-Reranker-V1-Nano needs transformers >= 5.3.0 because the
|
| 2 |
-
# T5Gemma2ForConditionalGeneration architecture is built into transformers v5+.
|
| 3 |
-
# The model card validates against 5.6.2 — pin to that to match upstream tests.
|
| 4 |
transformers==5.6.2
|
| 5 |
-
|
| 6 |
-
# huggingface_hub is needed to fetch kalm_reranker.py from the model repo
|
| 7 |
huggingface_hub>=0.25.0
|
| 8 |
-
|
| 9 |
-
# gradio + spaces are pre-installed on HF Spaces, but pinning avoids surprises
|
| 10 |
gradio==5.9.0
|
| 11 |
spaces>=0.0.34
|
| 12 |
-
|
| 13 |
-
# torch is also pre-installed on HF Spaces GPU images; pinning is optional.
|
| 14 |
-
# Leave it unpinned so HF can resolve a CUDA-compatible wheel automatically.
|
| 15 |
torch
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
transformers==5.6.2
|
|
|
|
|
|
|
| 2 |
huggingface_hub>=0.25.0
|
|
|
|
|
|
|
| 3 |
gradio==5.9.0
|
| 4 |
spaces>=0.0.34
|
|
|
|
|
|
|
|
|
|
| 5 |
torch
|