Downloads · 30 days
9
12% of all-time downloads
MaximeM/Qwen3-Embedding-4B-recsys-challenge-2026-retriever-ctx1024
Qwen3-Embedding-4B-recsys-challenge-2026-retriever-ctx1024 is a feature extraction model from MaximeM. Use it when you need embeddings to search or compare text. It is set up for sentence-transformers. The card lists the license as apache-2.0.
Dense dual-encoder retriever for conversational music recommendation, from team FPMsUMONS's entry to the RecSys Challenge 2026. It is the dense channel of the four-channel retrieval pool (BM25, dense, item-based colla…
Downloads · 30 days
9
12% of all-time downloads
All-time downloads
77
Public
Parameters
4B
8.1 GB on disk
Likes
0
Public
Click a slice to open those files.
.safetensors8 GB · 100%
From the Hugging Face model README
Dense dual-encoder retriever for conversational music recommendation, from team FPMs_UMONS's entry to the RecSys Challenge 2026. It is the dense channel of the four-channel retrieval pool (BM25, dense, item-based collaborative filtering, artist expansion) that feeds our scoring-head reranker.
Fine-tuned from Qwen/Qwen3-Embedding-4B.
This is a pipeline component, not a standalone model. It is one channel of a retrieval-and-reranking system and is only meaningful inside it. There is no standalone usage recipe here on purpose: how to download it, what text to encode, and how its output is fused with the other three channels are all documented — and executable end to end — in the repository.
MaximeM/Qwen3-8B-recsys-challenge-2026-scorehead-top200Retrieval recall figures and ablations are reported in the paper and the repository; they are not duplicated here.
The model must be used with max_seq_length = 1024 and
truncation_side = 'left', matching training. The current user request is placed
last in the query text, and most dev queries exceed 256 tokens — with default
right truncation the request is silently thrown away on every long conversation,
which costs most of the model's value. The pipeline sets both; anything loading
these weights directly has to do the same.
LoRA on attention, MultipleNegativesRankingLoss with cross-device gather, no
hard negatives, at sequence length 1024 with left truncation. Merged after
training and exported in sentence-transformers format.
Released under the Apache License 2.0, inherited from
Qwen/Qwen3-Embedding-4B. A copy
of the licence is included as LICENSE.
Modifications: these weights are a fine-tuned derivative of Qwen3-Embedding-4B, adapted with LoRA on the attention projections and merged, with the sequence length and truncation side changed as described above.