Instructions to use jatshi/StreamSense-Serve-v4-Router with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- sentence-transformers
How to use jatshi/StreamSense-Serve-v4-Router with sentence-transformers:
from sentence_transformers import SentenceTransformer model = SentenceTransformer("jatshi/StreamSense-Serve-v4-Router") sentences = [ "The weather is lovely today.", "It's so sunny outside!", "He drove to the stadium." ] embeddings = model.encode(sentences) similarities = model.similarity(embeddings, embeddings) print(similarities.shape) # [3, 3] - Notebooks
- Google Colab
- Kaggle
Download e2e_summary.json from jatshi/StreamSense-Serve-v4-Router: direct link, hf CLI and curl.
- Browser
- Download file 1.91 kB
-
https://huggingface.co/jatshi/StreamSense-Serve-v4-Router/resolve/main/e2e_summary.json
- Command line
-
hf download hf://jatshi/StreamSense-Serve-v4-Router/e2e_summary.json
-
curl -L -o e2e_summary.json https://huggingface.co/jatshi/StreamSense-Serve-v4-Router/resolve/main/e2e_summary.json
1.91 kB
| { | |
| "schema_version": 1, | |
| "physical_cases_per_system": 32, | |
| "independent_test_groups": 8, | |
| "latency_comparable_across_systems": false, | |
| "latency_note": "Observed wall latency is retained for execution audit only. ASR/OCR observations are memoized across systems in requested order, so later systems exclude analyzer work and their latency must not be compared with the first system.", | |
| "systems": { | |
| "always_vlm": { | |
| "cases": 32, | |
| "quality_pass_rate": 0.6875, | |
| "state_accuracy": 0.71875, | |
| "exact_state_accuracy": 0.6875, | |
| "answer_fact_coverage": 0.96875, | |
| "citation_validity": 0.875, | |
| "mean_answer_char_f1": 0.6598538614163614, | |
| "visual_backend_call_rate": 1.0, | |
| "text_answer_backend_call_rate": 1.0, | |
| "backend_calls_per_case": 2.0, | |
| "mean_observed_wall_latency_ms": 2576.251625374425, | |
| "backend_call_savings_vs_always": 0.0 | |
| }, | |
| "learned": { | |
| "cases": 32, | |
| "quality_pass_rate": 0.65625, | |
| "state_accuracy": 0.6875, | |
| "exact_state_accuracy": 0.6875, | |
| "answer_fact_coverage": 0.96875, | |
| "citation_validity": 0.84375, | |
| "mean_answer_char_f1": 0.6769566939286767, | |
| "visual_backend_call_rate": 0.90625, | |
| "text_answer_backend_call_rate": 1.0, | |
| "backend_calls_per_case": 1.90625, | |
| "mean_observed_wall_latency_ms": 12376.805822888855, | |
| "backend_call_savings_vs_always": 0.046875 | |
| }, | |
| "never_vlm": { | |
| "cases": 32, | |
| "quality_pass_rate": 0.78125, | |
| "state_accuracy": 0.8125, | |
| "exact_state_accuracy": 0.8125, | |
| "answer_fact_coverage": 0.96875, | |
| "citation_validity": 0.84375, | |
| "mean_answer_char_f1": 0.6423656732759995, | |
| "visual_backend_call_rate": 0.0, | |
| "text_answer_backend_call_rate": 1.0, | |
| "backend_calls_per_case": 1.0, | |
| "mean_observed_wall_latency_ms": 1510.641771601513, | |
| "backend_call_savings_vs_always": 0.5 | |
| } | |
| } | |
| } |