{"record":{"id":"9ff16fb1465b14d2","repo":"huggingface/candle","slug":"mismatch-between-deepstack-visual-embeds-and","errorCode":null,"errorMessage":"Mismatch between DeepStack visual embeds ({}) and mask positions ({})","messagePattern":"Mismatch between DeepStack visual embeds \\((.+?)\\) and mask positions \\((.+?)\\)","errorType":"validation","errorClass":null,"httpStatus":null,"severity":"error","filePath":"candle-transformers/src/models/qwen3_vl/text.rs","lineNumber":367,"sourceCode":"\n        let mask = visual_pos_masks.to_device(device)?.to_dtype(DType::F32)?;\n        let mask_flat = mask.flatten_all()?;\n\n        let masked_count = mask_flat.sum_all()?.to_scalar::<f32>()? as usize;\n        let visual_embeds = visual_embeds.to_device(device)?.to_dtype(dtype)?;\n\n        if masked_count == 0 {\n            if visual_embeds.dim(0)? != 0 {\n                candle::bail!(\n                    \"DeepStack visual embeds ({}) provided but mask is empty\",\n                    visual_embeds.dim(0)?\n                );\n            }\n            return Ok(hidden_states);\n        }\n\n        if visual_embeds.dim(0)? != masked_count {\n            candle::bail!(\n                \"Mismatch between DeepStack visual embeds ({}) and mask positions ({})\",\n                visual_embeds.dim(0)?,\n                masked_count\n            );\n        }\n\n        let (batch, seq, hidden) = hidden_states.dims3()?;\n        let total_positions = batch * seq;\n        let mut hidden_flat = hidden_states.reshape((total_positions, hidden))?;\n\n        let prefix = mask_flat.cumsum(0)?;\n        let rank = (prefix - &mask_flat)?.mul(&mask_flat)?;\n        let rank_u32 = rank.to_dtype(DType::U32)?;\n\n        let positions = Tensor::arange(0u32, total_positions as u32, device)?;\n        let positions_f32 = positions.to_dtype(DType::F32)?;\n        let masked_positions = positions_f32.mul(&mask_flat)?;\n","sourceCodeStart":349,"sourceCodeEnd":385,"githubUrl":"https://github.com/huggingface/candle/blob/d5fee525bfde3273eb7c9b75fd2bc4937be867ca/candle-transformers/src/models/qwen3_vl/text.rs#L349-L385","documentation":"Thrown when the number of DeepStack visual embeddings does not match the number of positions flagged in the attention mask for insertion.","triggerScenarios":"Thrown at candle-transformers/src/models/qwen3_vl/text.rs:367 when the library encounters an invalid state.","commonSituations":"See trigger scenarios.","solutions":[],"exampleFix":null,"handlingStrategy":null,"validationCode":null,"typeGuard":null,"tryCatchPattern":null,"preventionTips":[],"tags":[],"backgroundTag":null,"analyzedSha":"d5fee525bfde3273eb7c9b75fd2bc4937be867ca","analyzedAt":"2026-09-02T00:15:47.023Z","contentChangedAt":null,"schemaVersion":2},"datasetVersion":"2026-09-09T06:17:21.866Z"}