Views
No views yet
log1p(ReLU(logits)).max() document pooling)static_query_weights.pt)
initialized from IDF and learned during training.| File | Description |
|---|---|
config.json | Model config (HF format) |
model.safetensors / pytorch_model.bin | Document-encoder weights |
tokenizer* | Tokenizer files |
static_query_weights.pt | Learned static query token weights |
trainer_state.pt | Optimizer / scheduler state + training metrics at best step |
1import torch
2from transformers import AutoTokenizer, AutoModelForMaskedLM
3
4repo = "Cdn13/splade-multi-static-doc"
5tokenizer = AutoTokenizer.from_pretrained(repo)
6model = AutoModelForMaskedLM.from_pretrained(repo)
7
8# Load static query weights
9sqw = torch.load("static_query_weights.pt", map_location="cpu")
10query_weights = sqw["query_weights"] # shape: [vocab_size]Note: The query representation ispresence(token) * query_weights[token], computed without any forward pass through the model.