Commit Graph

187 Commits

Author SHA1 Message Date
Dobromir Popov
2868fc0d56 fix: tracker must not advertise incomplete snapshot as a full model source
files_for_layer_range() silently dropped required weight files that were
missing from the tracker's local snapshot directory, so a snapshot with
zero (or partial) shard files downloaded still reported back a "complete"
file list (metadata only, or a partial subset). Nodes then trusted that
subset as fully cached and skipped the real download, only discovering
the missing weight shards at model-load time.

Now, if any weight file required for the requested layer range is absent
on disk, the tracker returns an empty file list for that range so it
isn't offered as a source at all, letting the node fall through to
peers/HuggingFace for the real data.

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
2026-07-21 13:45:27 +03:00
Dobromir Popov
6aced6a005 fix: reconcile legacy branch runtime with current GGUF 2026-07-17 14:21:02 +03:00
Dobromir Popov
c758106a42 Merge branch 'archived_ralph/proxy-stream-cancellation' into merge/all-branches-into-master 2026-07-17 13:45:04 +03:00
Dobromir Popov
f0ddb69d33 Merge branch 'archived_ralph/dgr-001-performance-contract' into merge/all-branches-into-master
# Conflicts:
#	.claude/memory/MEMORY.md
#	.scratch/distributed-gguf-runtime/PRD.md
#	.scratch/distributed-gguf-runtime/RALPH-CONTEXT.md
#	.scratch/distributed-gguf-runtime/README.md
#	.scratch/distributed-gguf-runtime/architecture.md
#	.scratch/distributed-gguf-runtime/evidence/DGR-017/README.md
#	.scratch/distributed-gguf-runtime/implementation-strategy.md
#	.scratch/distributed-gguf-runtime/issues/07-add-isolated-concurrent-local-hot-kv-state.md
#	.scratch/distributed-gguf-runtime/issues/13-harden-failure-cancellation-and-restart-semantics.md
#	.scratch/distributed-gguf-runtime/milestones.md
#	.scratch/distributed-gguf-runtime/prd.json
#	docs/issues/distributed-gguf-runtime/01-lock-the-safetensors-versus-gguf-performance-contract.md
#	docs/issues/distributed-gguf-runtime/02-adopt-the-versioned-grpc-shard-protocol.md
#	docs/issues/distributed-gguf-runtime/03-define-exact-artifact-and-runtime-recipe-identity.md
#	docs/issues/distributed-gguf-runtime/05-implement-dense-llama-range-aware-gguf-ownership.md
#	docs/issues/distributed-gguf-runtime/06-implement-architecture-defined-boundary-input-output.md
2026-07-17 13:44:52 +03:00
Dobromir Popov
9b257d9a1b feat: define shard lifecycle status contract 2026-07-17 11:58:50 +03:00
Dobromir Popov
ab466ce6b6 feat: add activation stream envelope 2026-07-17 02:36:24 +03:00
Dobromir Popov
9db036f91a feat: DGR-018 canonical Ralph metadata schema 2026-07-16 23:01:55 +03:00
Dobromir Popov
369b2072cc chore: clean superseded GGUF scaffolding 2026-07-16 22:32:37 +03:00
Dobromir Popov
994546f78e Merge remote-tracking branch 'origin/master' into ralph/proxy-stream-cancellation 2026-07-16 18:06:54 +03:00
Dobromir Popov
6fd9d93e4b fix: flush direct SSE proxy frames before completion 2026-07-16 18:06:50 +03:00
Dobromir Popov
02b3709311 feat: checkpoint batching and release-gate stories 2026-07-16 17:24:56 +03:00
Dobromir Popov
254627629b Merge commit '47b243cd98fd94da7918cacf5725373b099208e5' into ralph/distributed-gguf-runtime 2026-07-15 23:04:52 +02:00
Dobromir Popov
1fe31ef38d feat: checkpoint distributed gguf runtime stories 2026-07-15 23:42:58 +03:00
Dobromir Popov
47b243cd98 model loading, dash 2026-07-15 13:55:38 +02:00
Dobromir Popov
2852b1f80b loading more 2026-07-15 12:54:51 +02:00
Dobromir Popov
22f28bd69a fix model load/unload 2026-07-15 12:35:32 +02:00
Dobromir Popov
97e2784b37 node registration fixes 2026-07-15 10:34:41 +02:00
Dobromir Popov
c035bad5b7 feat: wire live benchmark CLI endpoints 2026-07-15 10:34:20 +03:00
Dobromir Popov
a508768e8a feat: add live endpoint benchmark runner 2026-07-14 22:46:11 +03:00
Dobromir Popov
e6f6782995 feat: add deterministic CPU/GPU benchmark runner slice 2026-07-14 21:39:13 +03:00
Dobromir Popov
ba7c656364 node metrics 2026-07-14 20:33:02 +02:00
Dobromir Popov
b661590ac7 log window bigger 2026-07-14 17:47:20 +02:00
Dobromir Popov
5b33bf8b99 feat: compare safetensors and gguf on cpu and gpu 2026-07-14 18:45:12 +03:00
Dobromir Popov
c7554ef7d8 feat: add DGR-001 performance contract 2026-07-14 18:13:54 +03:00
Dobromir Popov
8cb00e951f feat: show admin node pool capacity 2026-07-14 16:11:18 +02:00
Dobromir Popov
22467f145c merge: distributed performance baseline benchmark 2026-07-14 17:01:08 +03:00
Dobromir Popov
35af1e21de fix: make model placement controls observable 2026-07-14 16:00:37 +02:00
Dobromir Popov
905ea16ce0 feat: complete route session baseline benchmark 2026-07-14 16:55:52 +03:00
Dobromir Popov
348b003d6e fix: restore responsive dashboard panel grid 2026-07-14 15:55:24 +02:00
Dobromir Popov
1e64a5b2b9 new dash update 2026-07-14 15:29:11 +02:00
Dobromir Popov
e2f3ae32b8 feat: let admins manage model placement 2026-07-14 15:16:23 +02:00
Dobromir Popov
5c9a2f6c97 dash style fix 2026-07-14 13:29:51 +02:00
Dobromir Popov
4eeec7fa7f feat: MAINT-001 - Fix Ruff violations across all Python source 2026-07-14 14:17:23 +03:00
Dobromir Popov
7925e5253d feat: implement DGR-006 tensor bundle boundary 2026-07-14 13:52:57 +03:00
Dobromir Popov
13d82f8032 dash, tests 2026-07-14 12:26:10 +02:00
Dobromir Popov
31065c0e12 feat: distributed GGUF shard load integration test with TinyLlama 1.1B 2026-07-14 13:01:51 +03:00
Dobromir Popov
ec36290863 feat: emit native DGR-003 shard identity 2026-07-14 11:31:10 +03:00
Dobromir Popov
252d131e7d feat: DGR-005A dense Llama owned range loader 2026-07-14 11:01:28 +03:00
Dobromir Popov
f9722e7b57 feat: DGR-004-CHAIN - Execute chained DGR-004/005/003-emission with anchor 2026-07-14 10:34:38 +03:00
Dobromir Popov
7b8e467c6b fix: harden DGR-003 identity trust boundary 2026-07-14 09:48:42 +03:00
Dobromir Popov
7364ed6731 fix: harden DGR-017 contract continuity 2026-07-14 01:10:33 +03:00
Dobromir Popov
ad2d17541c feat: DGR-003 - Define exact Artifact and runtime recipe identity 2026-07-14 01:10:07 +03:00
Dobromir Popov
e7c780a623 feat: DGR-017 - Lock the GLM-5.2 Max target and alpha contract 2026-07-14 00:19:16 +03:00
Dobromir Popov
5d87e81bc9 feat: harden node placement and partial model loading 2026-07-13 21:58:08 +02:00
Dobromir Popov
ef2a9e67e8 feat: add signed ROCm diagnostic lane 2026-07-13 21:24:43 +03:00
Dobromir Popov
a6bcc69288 sol mainnet payouts tasks 2026-07-13 18:51:40 +02:00
Dobromir Popov
b1c9deeb01 fix: cryptographically bind DGR-001 evidence 2026-07-13 19:38:14 +03:00
Dobromir Popov
9e67b829e3 fix: harden DGR-001 performance contract evidence 2026-07-13 19:10:24 +03:00
Dobromir Popov
e24db7854f feat: DGR-001 - Lock the safetensors-versus-GGUF performance contract 2026-07-13 17:55:55 +03:00
Dobromir Popov
59f2486bf2 feat: DGR-001 - Lock the safetensors-versus-GGUF performance contract 2026-07-13 17:49:09 +03:00