Dobromir Popov 471893c9d5 Skip multimodal/MTP checkpoint tensors absent from the text-only causal LM ...

Qwen3.5/3.6-MoE checkpoints ship vision (model.visual.*) and multi-token-
prediction (mtp.*) weights; the partial shard loader assigned them into the
text-only Qwen3_5MoeForCausalLM and crashed with AttributeError 'mtp'.
Filter selected tensors against the built model's state_dict keys, matching
transformers' _keys_to_ignore_on_load_unexpected behavior.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

2026-07-07 19:16:19 +02:00

.3rdparty

iinit

2026-06-28 23:49:11 +03:00

.agents/skills

skills update; USER ACCOUNT system! Alpha!

2026-07-03 19:22:39 +03:00

.claude

fixing streaming

2026-07-07 16:06:05 +02:00

.scratch

tasks

2026-07-07 15:56:38 +03:00

.vscode

tracher md

2026-07-02 22:03:37 +03:00

deploy

new tasks, devnet topup, cli, new model support

2026-07-06 14:17:36 +03:00

docs

dash QOL

2026-07-07 17:37:38 +03:00

packages

Skip multimodal/MTP checkpoint tensors absent from the text-only causal LM

2026-07-07 19:16:19 +02:00

scripts

feat: USDT reward system — billing ledger, devnet treasury, settlement loop, forfeiture PoW, tracker dashboard (US-030…US-035, ADR-0015)

2026-07-02 22:31:02 +02:00

tests

Skip multimodal/MTP checkpoint tensors absent from the text-only causal LM

2026-07-07 19:16:19 +02:00

_DEV_NOTES.md

notes

2026-07-03 00:55:06 +02:00

.billing.sqlite

billing

2026-07-04 23:58:07 +03:00

.dockerignore

dash works !!! good data. billing seems to work

2026-07-02 23:07:41 +02:00

.env.testnet

env

2026-07-06 18:47:21 +03:00

.gitignore

db

2026-07-07 19:31:29 +03:00

accounts.sqlite

new tasks, devnet topup, cli, new model support

2026-07-06 14:17:36 +03:00

AGENTS.md

memories in git

2026-07-01 14:40:42 +02:00

billing.sqlite

db

2026-07-07 19:31:29 +03:00

conftest.py

feat(us-017): P2P gossip, NAT-traversal relay node, and TLS

2026-06-29 17:58:16 +03:00

CONTEXT.md

feature-gguf-distributed

2026-07-07 15:27:33 +03:00

meshnet_registry.sqlite3

db

2026-07-07 19:31:29 +03:00

pyproject.toml

feat(alpha): complete hardening backlog

2026-07-05 21:47:23 +03:00

QUICKSTART.md

better dash and inference api QOL

2026-07-07 15:51:27 +02:00

skills-lock.json

iinit

2026-06-28 23:49:11 +03:00

uv.lock

feat(alpha): complete hardening backlog

2026-07-05 21:47:23 +03:00

Description

No description provided

1.2 GiB

Languages

Python 94.7%

HTML 4.1%

C++ 0.7%

Shell 0.2%

JavaScript 0.2%

Other 0.1%