Eval
From Dontopedia, the open, paraconsistent wiki. (Last updated 2026-06-10.)
Eval has 10 facts recorded in Dontopedia across 5 references.
Mostly:cuts time to(1), cuts time from(1), involves(1)
Maturity scale
raw canonical shape-checked rule-derived certifiedInbound mentions (9)
Other subjects in dontopedia point AT this entity as a value. These are inverse relationships — e.g. "X motherOf this subject" — and answer questions the forward facts can't. Grouped by predicate.
allowsReadingByCallerAfterAllows Reading by Caller After(1)
- Option a
ex:option-a
callsCalls(1)
- Model Eval
ex:model-eval
enablesSpeedupEnables Speedup(1)
- Metal Sgemm
ex:metal-sgemm
isCustomIs Custom(1)
- Omega and Ajax Custom Eval
ex:omega-and-ajax-custom-eval
noIssueInNo Issue in(1)
- Randy Models
ex:randy-models
requiresMetalForRequires Metal for(1)
- Wavenativelm
ex:wavenativelm
scalesSameWayAsScales Same Way As(1)
- Training
ex:training
setsSets(1)
- Model Eval Mode
ex:model-eval-mode
setStateSet State(1)
- Model
ex:model
Other facts (10)
The long tail: predicates that appear too rarely to warrant their own section. Filter or scroll to find a specific one. Each row links to its source.
| Predicate | Value | Ref |
|---|---|---|
| Cuts Time to | ~10–18 seconds | [1] |
| Cuts Time From | 90 seconds | [1] |
| Involves | 448 prefills × 256 tokens | [1] |
| Has Sequences | 448 | [2] |
| Has Total Dispatches | 185000 | [2] |
| Uses Salon Holdout Mix | true | [3] |
| Is | Salon Holdout Mix | [3] |
| Is Method | Model Instance | [4] |
| Is Method of | Model | [5] |
| Rdf:type | Model State | [5] |
Timeline
Timeline axis is valid_time — when each source says the fact was true in the world, not when Dontopedia learned about it. Retracted rows are kept for provenance; coloured stripes indicate the context kind.
References (5)
ctx:discord/blah/watt-activation/part-638ctx:discord/blah/watt-activation/part-641ctx:discord/blah/watt-activation/part-680ctx:claims/beam/b729dc6d-53ff-42db-95a2-0b4b64111a65- full textbeam-chunktext/plain1 KB
doc:beam/b729dc6d-53ff-42db-95a2-0b4b64111a65Show excerpt
self.fc3 = nn.Linear(32, 1) self.dropout = nn.Dropout(0.5) def forward(self, x): x = torch.relu(self.fc1(x)) x = self.dropout(x) x = torch.relu(self.fc2(x)) x = self.dropout(x) x …
ctx:claims/beam/af924c4f-8579-4b2a-85d1-c042076b09c7- full textbeam-chunktext/plain1 KB
doc:beam/af924c4f-8579-4b2a-85d1-c042076b09c7Show excerpt
loss = loss / accumulation_steps # Backward pass scaler.scale(loss).backward() # Update weights if (i + 1) % accumulation_steps == 0: scaler.step(optimizer) …
See also
Keep researching
Missing something or suspicious of what's here? Kick off a research session — a Claude agent will investigate, cite its sources, and file new facts into a dedicated context you can review before accepting into the shared view.