Currently it just sees black square instead of actual image

We have a start of this on the explorer page, but it would be nice to have something specific to validators which:

• showed all active validators and how many solutions and claims they are solving
• how much validators are claiming
• if there has been any contestations by or against a validators solutions

This could be implemented similarly to explorer page with just an event listener updating the stats

If tx reverts due to min stake, automatically restake immediately

Arbius has a $150 AIUS bounty for this issue!

Please read carefully:

To begin work on a bounty, reply by saying “I claim this bounty” - you will have 48 hours to submit your PR before someone else may attempt to claim this bounty.

To complete the bounty, within 48 hours of claiming, reply with a link to your PR referencing this issue and an Ethereum address. You must comply with reviewers comments and have the PR merged to receive the bounty reward. Please be sure to focus on quality submissions to minimize the amount of time reviewers must take.

Currently the model only support A100(40G/80G)

Need support for more financially affordable GPUs.

We can support by either:

• all kind of GPUs output the same image.
• same output for several different GPUs.

all time periods should be evaluated based on real world and adjusted if necessary

Implement and test a cog container built for svd and deploy it

Requires some tracking of wallet balance

in future update, these should be removed and we can use contract upgrade to set these values. this will save opcodes and allow us to do more in core contract.

How can we mine arbius using GPU?

A program which indexes all contestations, including all relevant details such as voting, tasks, and ideally can provide alerts of these too.

in submitTask, since getIPFSCID is called for each input on same type of input, we could cache and re-use this if bulk generation all used same input.

not an issue now, but as staked amounts get to be more significant % matters

[email protected]:~/kubo/arbius/miner$ yarn start:dev MiningConfig.json
yarn run v1.22.21
$ npx nodemon MiningConfig.json
[nodemon] 2.0.22
[nodemon] to restart at any time, enter rs
[nodemon] watching path(s): src/**/*
[nodemon] watching extensions: ts,js
[nodemon] starting npx ts-node ./src/start.ts MiningConfig.json
/root/kubo/arbius/miner/node_modules/ts-node/src/index.ts:859
return new TSError(diagnosticText, diagnosticCodes, diagnostics);
^
TSError: _ Unable to compile TypeScript:
src/blockchain.ts:7:28 - error TS2307: Cannot find module './artifacts/contracts/EngineV2.sol/EngineV2.json' or its corresponding type declarations.

7 import EngineArtifact from './artifacts/contracts/EngineV2.sol/EngineV2.json';
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

at createTSError (/root/kubo/arbius/miner/node_modules/ts-node/src/index.ts:859:12)
at reportTSError (/root/kubo/arbius/miner/node_modules/ts-node/src/index.ts:863:19)
at getOutput (/root/kubo/arbius/miner/node_modules/ts-node/src/index.ts:1077:36)
at Object.compile (/root/kubo/arbius/miner/node_modules/ts-node/src/index.ts:1433:41)
at Module.m._compile (/root/kubo/arbius/miner/node_modules/ts-node/src/index.ts:1617:30)
at Module._extensions..js (node:internal/modules/cjs/loader:1213:10)
at Object.require.extensions.<computed> [as .ts] (/root/kubo/arbius/miner/node_modules/ts-node/src/index.ts:1621:12)
at Module.load (node:internal/modules/cjs/loader:1037:32)
at Function.Module._load (node:internal/modules/cjs/loader:878:12)
at Module.require (node:internal/modules/cjs/loader:1061:19) {

diagnosticCodes: [ 2307 ]
}
[nodemon] app crashed - waiting for file changes before starting...

For at least the cooldown time

Hi, I'm trying to get a miner running on an A100 40gb instance, everything is set up properly but I'm getting this cryptic error.
Thanks!

2024-02-14 18:28:13.396 INFO /src/start.ts:31 Arbius Miner 32415dc1 starting 2024-02-14 18:28:13.405 DEBUG /src/start.ts:36 Logging to log.txt 2024-02-14 18:28:13.410 DEBUG /src/start.ts:39 Database loaded from db.sqlite 2024-02-14 18:28:13.411 DEBUG /src/start.ts:42 ML initialized 2024-02-14 18:28:13.484 DEBUG /src/start.ts:45 Loaded wallet (0xF7d6D37DFeB5fAF8014b6bBDA42c192c1e92B07e) 2024-02-14 18:28:13.503 DEBUG /src/rpc.ts:100 RPC server listening on localhost:4000 2024-02-14 18:28:13.506 DEBUG /src/start.ts:48 RPC initialized 2024-02-14 18:28:13.507 DEBUG /src/index.ts:973 Setting max file listeners to 100 for ethers 2024-02-14 18:28:13.511 DEBUG /src/index.ts:977 Clearing old automatically added retry jobs 2024-02-14 18:28:13.512 DEBUG /src/index.ts:981 Bootup check 2024-02-14 18:28:13.764 INFO /src/index.ts:964 Arbius version (0) fits miner version (0) 2024-02-14 18:28:17.379 ERROR /src/index.ts:996 Model (0x82ae0d19f32b6912204160f8d744de08265b7775d07c23b4171c94c8e2892c62) CID (0x12202fb3b8bea893a490d2c8b7c329c6c935a748c7fdde81dc94a1fca61a9f2060b7) does not match expected CID (0x12201bdab4164320cc8621282982c55eb76e14427aa5793278b37b6108f63fb5d577) 2024-02-14 18:28:17.380 INFO /src/index.ts:997 If you are running a100 this is a bug, please report with system details at https://github.com/semperai/arbius 2024-02-14 18:28:17.380 INFO /src/index.ts:998 Join our telegram https://t.me/arbius_ai [nodemon] app crashed - waiting for file changes before starting...

Miners running Kandinsky-2 generate output image at ./cache/out-1.png which is later pinned to IPFS to generate cid. If a second running miner generates output for a different task between these steps a potentially wrong file and cid is generated for the first task.

Solutions:

1. Create unique output files such as {taskid}.png, I believe changing the filename results in a different cid? I don't know IPFS well but if so this change requires all miners to update naming consensus. This is probably more future-proof once multiple models are running etc. but does require cleaning up generated outputs occasionally.
2. Add a cache path field to MiningConfig.json that users can define on a per-miner basis. I've already done this locally as it doesn't break my IPFS cid generation so I'll create a pr

Hi, I have problems running A100 GPU locally, it gave me an error of this. I did make sure if my docker container is running A100, and it did. Can someone please help me?

2024-02-26 00:37:24.199 ERROR /src/index.ts:1196 Model (0x98617a8cd4a11db63100ad44bea4e5e296aecfd78b2ef06aee3e364c7307f212) CID (0x12202fb3b8bea893a490d2c8b7c329c6c935a748c7fdde81dc94a1fca61a9f2060b7) does not match expected CID (0x12201bdab4164320cc8621282982c55eb76e14427aa5793278b37b6108f63fb5d577)

A user reported that mintime filter was not functioning properly.

Needs better filter design

llm and vlm should be set in .env or similar

needs more investigation, but currently people incentivized to vote at last second. perhaps good to add some decaying time if voting near end. or, maybe not needed if people vote earlier.