pplx-embed-context-v1.2-4B / configuration.py
seslami-pplx's picture
Upload SLERP-merged checkpoint (alpha=0.5) from two adversarial-FT runs at step-1500
43188fa verified
Raw
History Blame Contribute Delete
152 Bytes
from transformers.models.qwen3.configuration_qwen3 import Qwen3Config
class PPLXQwen3Config(Qwen3Config):
model_type = "bidirectional_pplx_qwen3"