cortex

fix(gateway): full observability + stop leaking upstream bodies

#161:Commit aa88d37509 pushed by grenade

main

2026-05-22 04:22:28 +00:00

4m51s

fix(router,handlers): strip trailing slash from rewritten URL + log upstream failures

#159:Commit 0f00f72b47 pushed by grenade

main

2026-05-22 04:15:43 +00:00

4m51s

fix(rpm): migrate legacy helexa-cortex firewalld service to `cortex`

#155:Commit dc2a803266 pushed by grenade

main

2026-05-22 03:17:53 +00:00

4m45s

feat(stage-8e-3): quantize lm_head in TP Qwen3-Next

#153:Commit e71181499e pushed by grenade

main

2026-05-21 20:29:18 +00:00

9m18s

feat(stage-8e-2d): route quantized matmul by M (prefill vs decode)

#149:Commit 34f9b77d9d pushed by grenade

main

2026-05-21 18:20:51 +00:00

5m4s

fix(stage-8e-2c): cast bf16/f16 activations to f32 around QMatMul

#147:Commit f084aaab8e pushed by grenade

main

2026-05-21 17:10:04 +00:00

4m35s

fix(stage-8e-2b): allow quant on the TP load path

#145:Commit 68a606a79c pushed by grenade

main

2026-05-21 16:22:05 +00:00

4m40s

feat(stage-8e-2): plumb quant config from ModelSpec to TP load path

#143:Commit 4aa71902d0 pushed by grenade

main

2026-05-21 15:09:27 +00:00

5m36s

feat(stage-8d-7): direct safetensors fused-region loader

#139:Commit 8d7b099b36 pushed by grenade

main

2026-05-21 14:54:35 +00:00

4m45s

diag(stage-8d-6): per-layer VRAM logging in TP load path

#137:Commit 89d98d1fb2 pushed by grenade

main

2026-05-21 09:59:25 +00:00

5m10s

feat(stage-8d-5b): wire fused_gdn_gating CUDA kernel

#135:Commit cc95fe28d9 pushed by grenade

main

2026-05-21 09:18:16 +00:00

10m44s

feat(stage-8d-5): wire gated_delta_rule_recurrence kernel into tp_qwen3_5

#129:Commit 10c151efa5 pushed by grenade

main

2026-05-21 08:49:15 +00:00

4m51s

feat(tp): cancellation-safe inference + structured tracing

#123:Commit 70eb6af42b pushed by grenade

main

2026-05-21 05:27:36 +00:00

5m27s

fix(tp): always drain worker responses on leader failure

#121:Commit d1a4aad91d pushed by grenade

main

2026-05-21 04:45:12 +00:00

5m17s

feat(stage-8c): TP-aware Qwen3-Next (tp_qwen3_5)

#119:Commit 95dc8745eb pushed by grenade

main

2026-05-20 19:07:53 +00:00

5m0s

fix(qwen3_5): promote beta to F32 alongside q/k/v in delta rule

#117:Commit 495d3f7c05 pushed by grenade

main

2026-05-20 18:18:32 +00:00

4m59s

fix(qwen3_5): tensor names are under `model.language_model.*`, not `model.*`

#115:Commit 5c4c8e0eba pushed by grenade

main

2026-05-20 13:54:14 +00:00

5m44s

fix(qwen3_5): nested rope_parameters + partial_rotary_factor=0.25

#111:Commit 07c44d5db1 pushed by grenade

main

2026-05-20 13:23:53 +00:00

4m50s

feat(stage-8c): full-attention layer + decoder + Model + ForCausalLM for qwen3_5

#109:Commit e7eb3dab6a pushed by grenade

main

2026-05-20 12:57:56 +00:00

5m8s

feat(stage-8c): linear-attention layer (Qwen3-Next GatedDeltaNet)

#107:Commit 180274548d pushed by grenade

main

2026-05-20 06:35:13 +00:00

5m10s

feat(stage-8c): scaffold qwen3_5 (Qwen3.6) — dispatch + stubs + TP gate

#105:Commit a70f317729 pushed by grenade

main

2026-05-20 06:03:19 +00:00

5m8s

feat(stage-8b): Llama + Qwen3 MoE families on the candle harness

#103:Commit c6022aa6b9 pushed by grenade

main

2026-05-20 05:42:19 +00:00

5m47s

feat(stage-8a): pre-flight architecture check for dense model loads

#101:Commit 9e31d8deca pushed by grenade

main

2026-05-20 05:32:17 +00:00

4m35s

chore: keep models.example.toml generic; deploy.sh sync's local models.toml

#97:Commit 62ca125a68 pushed by grenade

main

2026-05-20 04:52:02 +00:00

4m40s

feat(cortex): unified /v1/models — catalogue × topology feasibility + cold-load

#95:Commit 735945ee81 pushed by grenade

main

2026-05-20 04:44:13 +00:00

4m57s

feat(tp): Stage 7c-i — streaming SSE through TP

#93:Commit f72dee094f pushed by grenade

main

2026-05-20 04:38:18 +00:00

5m21s

feat(tp): Stage 7b-iv — RPC + orchestration for TP load/inference

#91:Commit d46d8d4f6c pushed by grenade

main

2026-05-20 04:01:34 +00:00

7m4s

feat(tp): --tp-smoke CLI subcommand + remote validation script

#89:Commit 9b8bd146f6 pushed by grenade

main

2026-05-19 16:45:21 +00:00

4m46s

fix(tp): add half dep + drop double-wrapped .w() on CudaDevice::alloc

#87:Commit 96d8755245 pushed by grenade

main

2026-05-19 16:17:11 +00:00

5m1s

fix(tp): import BackendStorage trait for CudaStorage methods

#85:Commit 12549c9aed pushed by grenade

main

2026-05-19 15:37:00 +00:00

4m40s