Dontopedia

Inference command

From Dontopedia, the open, paraconsistent wiki. (Last updated 2026-06-06.)

Inference command has 30 facts recorded in Dontopedia across 5 references, with 3 live disagreements.

30 facts·16 predicates·5 sources·3 in dispute

Mostly:has argument(12), includes flag(2), uses flag(2)

Maturity scale raw canonical shape-checked rule-derived certified

Has Argumentin disputehasArgument

  • --load-ckpt /tmp/wave_native_ckpts_25m/step_2000/model.npz[3]sourceall time · 641
  • --size 25m[3]sourceall time · 641
  • --seq-len 256[3]sourceall time · 641
  • --prompt "Your prompt here"[3]sourceall time · 641
  • --generate 256[3]sourceall time · 641
  • --gen-temp 0.8[3]sourceall time · 641
  • Arg Load Ckpt[5]all time · 646
  • Arg Size 50m[5]all time · 646
  • Arg Seq Len 256[5]all time · 646
  • Arg Prompt[5]all time · 646

Inbound mentions (2)

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.

instructsUsageInstructs Usage(1)

providesInferenceInstructionsProvides Inference Instructions(1)

Other facts (17)

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.

17 facts
PredicateValueRef
Includes FlagInteractive Flag[1]
Includes FlagQa Flag[1]
Uses Flag--load-ckpt[4]
Uses Flag--size[4]
Uses ScriptInfer Py[1]
Uses ModelFinetune V7 Model Best Npz[1]
Uses AdapterKant Lora Adapters Final Npz[1]
Uses Size25m[2]
Uses Generate256[2]
Uses Gen Temp0.8[2]
Uses Load Ckpt/tmp/wave_native_ckpts_25m/step_2000/model.npz[2]
Uses PromptYour prompt here[2]
Uses Seq Len256[2]
Has Checkpoint Path/tmp/wave_native_ckpts_25m/step_2000/model.npz[3]
Uses Path/tmp/wave_native_ckpts_50m/step_2000/model.npz[4]
Rdf:typeCommand Line Instruction[5]
Executable./target/release/wave_native_train[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.

includesFlagblah/watt-activation/part-29
ex:interactive-flag
usesScriptblah/watt-activation/part-29
ex:infer-py
usesModelblah/watt-activation/part-29
ex:finetune-v7-model-best-npz
usesAdapterblah/watt-activation/part-29
ex:kant-lora-adapters-final-npz
includesFlagblah/watt-activation/part-29
ex:qa-flag
usesSizeblah/watt-activation/part-644
25m
usesGenerateblah/watt-activation/part-644
256
usesGenTempblah/watt-activation/part-644
0.8
usesLoadCkptblah/watt-activation/part-644
/tmp/wave_native_ckpts_25m/step_2000/model.npz
usesPromptblah/watt-activation/part-644
Your prompt here
usesSeqLenblah/watt-activation/part-644
256
labelblah/watt-activation/641
Inference command
hasArgumentblah/watt-activation/641
--load-ckpt /tmp/wave_native_ckpts_25m/step_2000/model.npz
hasCheckpointPathblah/watt-activation/641
/tmp/wave_native_ckpts_25m/step_2000/model.npz
hasArgumentblah/watt-activation/641
--size 25m
hasArgumentblah/watt-activation/641
--seq-len 256
hasArgumentblah/watt-activation/641
--prompt "Your prompt here"
hasArgumentblah/watt-activation/641
--generate 256
hasArgumentblah/watt-activation/641
--gen-temp 0.8
usesFlagblah/watt-activation/645
--load-ckpt
usesPathblah/watt-activation/645
/tmp/wave_native_ckpts_50m/step_2000/model.npz
usesFlagblah/watt-activation/645
--size
typeblah/watt-activation/646
ex:CommandLineInstruction
executableblah/watt-activation/646
./target/release/wave_native_train
hasArgumentblah/watt-activation/646
ex:arg-load-ckpt
hasArgumentblah/watt-activation/646
ex:arg-size-50m
hasArgumentblah/watt-activation/646
ex:arg-seq-len-256
hasArgumentblah/watt-activation/646
ex:arg-prompt
hasArgumentblah/watt-activation/646
ex:arg-generate-256
hasArgumentblah/watt-activation/646
ex:arg-gen-temp

References (5)

5 references
  1. [1]Part 295 facts
    ctx:discord/blah/watt-activation/part-29
  2. [2]Part 6446 facts
    ctx:discord/blah/watt-activation/part-644
  3. [3]6418 facts
    ctx:discord/blah/watt-activation/641
    • full textwatt-activation-641
      text/plain2 KBdoc:agent/watt-activation-641/f90a56d8-d0f5-46e7-92ee-3c668e2366e4
      Show excerpt
      [2026-04-16 01:42] xenonfun: ```bash ➜ ~/MS/HarmonicRust git:(master) ✗ ./target/release/wave_native_train --size 50m --data /tmp/domain_tinystories.bin --mode full --seq-len 256 --batch-size 32 --total-steps 10 --lr 0.001 --warmup 0 --log
  4. [4]6453 facts
    ctx:discord/blah/watt-activation/645
    • full textwatt-activation-645
      text/plain3 KBdoc:agent/watt-activation-645/5c424827-0967-4bac-bbcc-cb1b4a153e35
      Show excerpt
      [2026-04-16 02:39] lisamegawatts: oh nmd [2026-04-16 02:39] lisamegawatts: its all good [2026-04-16 02:40] lisamegawatts: i thought it still had the books stuff but it doesn't and you can filter it, so that should be plenty for 50m chinchil
  5. [5]6468 facts
    ctx:discord/blah/watt-activation/646
    • full textwatt-activation-646
      text/plain2 KBdoc:agent/watt-activation-646/84b51c92-144f-432e-8454-c9f4c46ef821
      Show excerpt
      [2026-04-16 03:27] xenonfun: ``` ⏺ Step 1,200: BPB 2.90 best, ~1,899 tok/s. Training well — BPB 8.6 → 2.9 in 1.2K steps. Throughput dipped slightly at step 1000 due to the DoReMi reference snapshot (cloning 50M params to disk + loading)

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.