# Tongue Report Generation on MedTCM 1.0 (test) Source: [Tongue Report Generation on MedTCM 1.0 (test)]() Collection index: [llms.txt]() Data updated: 2026-07-03T07:24:18.604Z Evaluation of models generating tongue diagnosis reports using the test split of the MedTCM 1.0 dataset, assessed using metrics including BLEU, ROUGE, and TDEU scores. A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage. | Scope | Recorded value | | --- | --- | | Task | Tongue report generation | | Dataset (including any stated split) | MedTCM 1.0 (test) | | Metric | BLEU-4 | | Unit | PERCENT | | Direction | Higher is better | | Results loaded | 17 of 17 | | Coverage | All results currently returned for this benchmark | Evaluation settings may differ between results. The original papers and recorded configurations are needed to establish comparability. A shared metric name alone does not establish an identical protocol. ## Metrics - [BLEU-4](): Higher is better; unit: PERCENT. - [ROUGE-1](): Higher is better; unit: PERCENT. - [ROUGE-2](): Higher is better; unit: PERCENT. - [ROUGE-L](): Higher is better; unit: PERCENT. - [TDEU Score (Overall)](): Lower is better; unit: PERCENT. - [TDEU Score (Tongue)](): Lower is better; unit: PERCENT. - [TDEU Score (Coat)](): Lower is better; unit: PERCENT. - [TDEU Score (Location)](): Lower is better; unit: PERCENT. ## Recorded results | Method | BLEU-4 | Result date | Report paper | | --- | --- | --- | --- | | MMIR–TCM (Mask train + Mask) | 83.57 | Not recorded | 6a47420d143ebdcacacc5cbb | | MMIR–TCM (Mask train + Raw) | 83.22 | Not recorded | 6a47420d143ebdcacacc5cbb | | MMIR–TCM (Raw train + Raw) | 83.2 | Not recorded | 6a47420d143ebdcacacc5cbb | | MMIR–TCM (Raw train + Mask) | 83.11 | Not recorded | 6a47420d143ebdcacacc5cbb | | MMIR–TCM (Mask train + Mask, 3ep) | 83.09 | Not recorded | 6a47420d143ebdcacacc5cbb | | MMIR–TCM (Raw train + Raw, 3ep) | 81.72 | Not recorded | 6a47420d143ebdcacacc5cbb | | ViTCM–LLM E4 Full (r=64, 20ep) | 44.07 | Not recorded | 6a47420d143ebdcacacc5cbb | | ViTCM–LLM E4 Full (r=64, 10ep) | 43.57 | Not recorded | 6a47420d143ebdcacacc5cbb | | ViTCM–LLM E4 Full (r=64, 3ep) | 39.57 | Not recorded | 6a47420d143ebdcacacc5cbb | | ViTCM–LLM E4 Full (r=16, 3ep) | 37.94 | Not recorded | 6a47420d143ebdcacacc5cbb | | ViTCM–LLM E3 Vision+Projector | 35.82 | Not recorded | 6a47420d143ebdcacacc5cbb | | Original Qwen3–VL 30B (untuned) | 33.98 | Not recorded | 6a47420d143ebdcacacc5cbb | | ViTCM–LLM E2 Language–only | 26.74 | Not recorded | 6a47420d143ebdcacacc5cbb | | Grok–2–Vision–1212 | 26.4 | Not recorded | 6a47420d143ebdcacacc5cbb | | LLaMA4–scout 109B | 25.68 | Not recorded | 6a47420d143ebdcacacc5cbb | | Gemini–2.5–Flash | 24.43 | Not recorded | 6a47420d143ebdcacacc5cbb | | ViTCM–LLM E1 Zero–shot | 24.08 | Not recorded | 6a47420d143ebdcacacc5cbb | ### MMIR–TCM (Mask train + Mask) BLEU-4: 83.57. Result ID: 6a4763a060b88bb33992bcab. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Qwen3–VL 30B","Params":"30B","LoRA r":"64","Epochs":"10","Input Type":"Masked images"}. This record does not assert independent reproduction. ### MMIR–TCM (Mask train + Raw) BLEU-4: 83.22. Result ID: 6a4763a060b88bb33992bcaa. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Qwen3–VL 30B","Params":"30B","LoRA r":"64","Epochs":"10","Input Type":"Masked training, Raw inference"}. This record does not assert independent reproduction. ### MMIR–TCM (Raw train + Raw) BLEU-4: 83.2. Result ID: 6a4763a060b88bb33992bca8. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Qwen3–VL 30B","Params":"30B","LoRA r":"64","Epochs":"10","Input Type":"Raw images"}. This record does not assert independent reproduction. ### MMIR–TCM (Raw train + Mask) BLEU-4: 83.11. Result ID: 6a4763a060b88bb33992bca9. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Qwen3–VL 30B","Params":"30B","LoRA r":"64","Epochs":"10","Input Type":"Raw training, Masked inference"}. This record does not assert independent reproduction. ### MMIR–TCM (Mask train + Mask, 3ep) BLEU-4: 83.09. Result ID: 6a4763a060b88bb33992bcad. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Qwen3–VL 30B","Params":"30B","LoRA r":"64","Epochs":"3","Input Type":"Masked images"}. This record does not assert independent reproduction. ### MMIR–TCM (Raw train + Raw, 3ep) BLEU-4: 81.72. Result ID: 6a4763a060b88bb33992bcac. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Qwen3–VL 30B","Params":"30B","LoRA r":"64","Epochs":"3","Input Type":"Raw images"}. This record does not assert independent reproduction. ### ViTCM–LLM E4 Full (r=64, 20ep) BLEU-4: 44.07. Result ID: 6a4763a060b88bb33992bca6. [Reporting paper]() Recorded evaluation configuration: {"Backbone":"Qwen2.5–VL 32B","Params":"32B","LoRA r":"64","Epochs":"20"}. This record does not assert independent reproduction. ### ViTCM–LLM E4 Full (r=64, 10ep) BLEU-4: 43.57. Result ID: 6a4763a060b88bb33992bca5. [Reporting paper]() Recorded evaluation configuration: {"Backbone":"Qwen2.5–VL 32B","Params":"32B","LoRA r":"64","Epochs":"10"}. This record does not assert independent reproduction. ### ViTCM–LLM E4 Full (r=64, 3ep) BLEU-4: 39.57. Result ID: 6a4763a060b88bb33992bca4. [Reporting paper]() Recorded evaluation configuration: {"Backbone":"Qwen2.5–VL 32B","Params":"32B","LoRA r":"64","Epochs":"3"}. This record does not assert independent reproduction. ### ViTCM–LLM E4 Full (r=16, 3ep) BLEU-4: 37.94. Result ID: 6a4763a060b88bb33992bca3. [Reporting paper]() Recorded evaluation configuration: {"Backbone":"Qwen2.5–VL 32B","Params":"32B","LoRA r":"16","Epochs":"3"}. This record does not assert independent reproduction. ### ViTCM–LLM E3 Vision+Projector BLEU-4: 35.82. Result ID: 6a4763a060b88bb33992bca2. [Reporting paper]() Recorded evaluation configuration: {"Backbone":"Qwen2.5–VL 32B","Params":"32B","LoRA r":"16","Epochs":"3"}. This record does not assert independent reproduction. ### Original Qwen3–VL 30B (untuned) BLEU-4: 33.98. Result ID: 6a4763a060b88bb33992bca7. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Params":"30B"}. This record does not assert independent reproduction. ### ViTCM–LLM E2 Language–only BLEU-4: 26.74. Result ID: 6a4763a060b88bb33992bca1. [Reporting paper]() Recorded evaluation configuration: {"Backbone":"Qwen2.5–VL 32B","Params":"32B","LoRA r":"16","Epochs":"3"}. This record does not assert independent reproduction. ### Grok–2–Vision–1212 BLEU-4: 26.4. Result ID: 6a4763a060b88bb33992bc9d. [Reporting paper]() Recorded evaluation configuration: {"Setting":"zero-shot"}. This record does not assert independent reproduction. ### LLaMA4–scout 109B BLEU-4: 25.68. Result ID: 6a4763a060b88bb33992bc9e. [Reporting paper]() Recorded evaluation configuration: {"Params":"109B","Setting":"zero-shot"}. This record does not assert independent reproduction. ### Gemini–2.5–Flash BLEU-4: 24.43. Result ID: 6a4763a060b88bb33992bc9f. [Reporting paper]() Recorded evaluation configuration: {"Setting":"zero-shot"}. This record does not assert independent reproduction. ### ViTCM–LLM E1 Zero–shot BLEU-4: 24.08. Result ID: 6a4763a060b88bb33992bca0. [Reporting paper]() Recorded evaluation configuration: {"Backbone":"Qwen2.5–VL 32B","Params":"32B"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.