Skip to content

Commit

Permalink
Golem-Workers fixes after beta tests (#326)
Browse files Browse the repository at this point in the history 
* verify if fund went well

* better link for cloning (#322)

* article types (#323)

* article types

* article -> better names

* jsons from github in the hello example

* move fund verification higher

* sd example with code from github

* PIL installation

* increase logging

* Invoke-WebRequest

* use logs from examples/docker dir

* docker compose v2

* s/scalepointai/modelserve/

* less error-prone .env code snippets

* reputation not optional anymor
  • Loading branch information
@mateuszsrebrny
mateuszsrebrny authored Oct 8, 2024
1 parent effbc61 commit 129f517
Show file tree
Hide file tree
Showing 4 changed files with 91 additions and 229 deletions.
17 changes: 10 additions & 7 deletions src/pages/docs/creators/golem-workers/getting-started.md
View file
Original file line number Diff line number Diff line change
Expand Up @@ -29,6 +29,8 @@ This doesn't require a Python environment whatsoever. It is enough to have `dock

Make sure that [Docker](https://docker.com) is running on your machine and that your current user has access to it.

The instructions below have been tested with `Docker Compose V2`.

### 2. Clone the repository

Create a folder on your computer, navigate into it from the terminal, and then clone the repository using the command:
Expand Down Expand Up @@ -123,22 +125,23 @@ yagna app-key create <your-token-name>

And put the generated app-key into the `.env` file in the current directory:

```
YAGNA_APPKEY=<your-application-token>
```bash
echo 'YAGNA_APPKEY=<your-application-token>' >> .env
```

This will allow `golem-node` to know your `golem-workers` instance.

### 5. Use Golem Reputation (optional)
### 5. Use Golem Reputation

If you want to use the Golem Reputation service (to use only nodes with a record of good behavior)
put a new entry in the `.env` file in the current directory:
Golem Reputation service (prioritizing nodes with a record of good behavior)
requires putting a new entry in the `.env` file in the current directory:

```
GLOBAL_CONTEXTS=["golem_reputation.ReputationService"]
```bash
echo 'GLOBAL_CONTEXTS=["golem_reputation.ReputationService"]' >> .env
```

Note that, the [examples](/docs/creators/golem-workers/hello-example) in this documentation require the above to work properly.
You can skip this step if you don't use Golem Reputation in your jsons.

### 6. Start Golem-Workers

Expand Down
114 changes: 16 additions & 98 deletions src/pages/docs/creators/golem-workers/hello-example.md
View file
Original file line number Diff line number Diff line change
Expand Up @@ -15,7 +15,8 @@ the [Stable Diffusion tutorial](/docs/creators/golem-workers/sd-example) availab

## Prerequisites

We will be using **`curl`** to interact with the Golem-Workers API, which runs on **localhost:8000**
We will be using **`curl`** (for Windows users consider PowerShell command `Invoke-WebRequest` instead)
to interact with the Golem-Workers API, which runs on **localhost:8000**
(see [Getting started](/docs/creators/golem-workers/getting-started) if it is not).
All API requests will use **JSON data** to configure and manage the Golem-Workers service.

Expand All @@ -36,48 +37,10 @@ curl --location 'http://localhost:8000/create-cluster' \

This is the json file you have just used (also available on
[GitHub](https://github.com/golemfactory/golem-workers/raw/main/examples/hello-example/create-cluster.json)):
```json
{
"cluster_id": "example",
"budget_types": {
"default": {
"budget": {
"golem_workers.budgets.AveragePerCpuUsageLinearModelBudget": {
"average_cpu_load": 1,
"average_duration_hours": 0.5,
"average_max_cost": 1.5
}
},
"scope": "cluster"
}
},
"network_types": {
"default": {
"ip": "192.168.0.1/16"
}
},
"node_types": {
"default": {
"market_config": {
"filters": [
{
"golem_reputation.ProviderBlacklistPlugin": {
"payment_network": "holesky"
}
}
],
"sorters": [
{
"golem_reputation.ReputationScorer": {
"payment_network": "holesky"
}
}
]
}
}
}
}
```
{% codefromgithub url="https://github.com/golemfactory/golem-workers/raw/main/examples/hello-example/create-cluster.json" language="json" %}

{% /codefromgithub %}


### `create-cluster` JSON explained

Expand All @@ -100,54 +63,13 @@ curl --location 'http://localhost:8000/create-node' \

This is the json file you have just used (also available on
[GitHub](https://github.com/golemfactory/golem-workers/raw/main/examples/hello-example/create-node.json)):
```json
{
"cluster_id": "example",
"node_networks": {
"default": {}
},
"node_config": {
"market_config": {
"demand": {
"payloads": [
{
"golem_workers.payloads.ClusterNodePayload": {
"image_tag": "scalepointai/echo-test:2"
}
}
]
}
},
"on_start_commands": [
{
"golem_workers.work.deploy_and_start_activity": {
"deploy_timeout_minutes": 60
}
},
{
"golem_workers.work.run_in_shell": [
[
"nginx"
]
]
}
],
"sidecars": [
{
"golem_workers.sidecars.WebsocatPortTunnelSidecar": {
"network_name": "default",
"local_port": "8080",
"remote_port": "80"
}
}
]
}
}
```
{% codefromgithub url="https://github.com/golemfactory/golem-workers/raw/main/examples/hello-example/create-node.json" language="json" %}

{% /codefromgithub %}

### `create-node` JSON explained

- **Image**: The node runs the **Hello Service** image `scalepointai/echo-test:2`,
- **Image**: The node runs the **Hello Service** image `modelserve/echo-test:2`,
which is a simple NGINX-based service that returns "Hello" on port **80**.
- **Commands**: We deploy the node and start the service using NGINX.
- **Port Forwarding**: We forward port **80** on the node to **localhost:8080** on your machine,
Expand All @@ -166,11 +88,9 @@ curl --location 'http://localhost:8000/get-cluster' \

This is the json file you have just used (also available on
[GitHub](https://github.com/golemfactory/golem-workers/raw/main/examples/hello-example/get-cluster.json)):
```json
{
"cluster_id": "example",
}
```
{% codefromgithub url="https://github.com/golemfactory/golem-workers/raw/main/examples/hello-example/get-cluster.json" language="json" %}

{% /codefromgithub %}

## 4. Access the Hello Service

Expand All @@ -189,11 +109,9 @@ curl --location 'http://localhost:8000/delete-cluster' \

This is the json file you have just used (also available on
[GitHub](https://github.com/golemfactory/golem-workers/raw/main/examples/hello-example/delete-cluster.json)):
```json
{
"cluster_id": "example"
}
```
{% codefromgithub url="https://github.com/golemfactory/golem-workers/raw/main/examples/hello-example/delete-cluster.json" language="json" %}

{% /codefromgithub %}

---

Expand Down
139 changes: 20 additions & 119 deletions src/pages/docs/creators/golem-workers/sd-example.md
View file
Original file line number Diff line number Diff line change
Expand Up @@ -12,7 +12,8 @@ In this example, we’ll show you how to run a sample [Automatic1111](https://gi

## Prerequisites

We will be using **`curl`** to interact with the Golem-Workers API, which runs on **localhost:8000**
We will be using **`curl`** (for Windows users consider PowerShell command `Invoke-WebRequest` instead)
to interact with the Golem-Workers API, which runs on **localhost:8000**
(see [Getting started](/docs/creators/golem-workers/getting-started) if it is not).
All API requests will use **JSON data** to configure and manage the Golem-Workers service.

Expand Down Expand Up @@ -53,51 +54,9 @@ curl --location 'http://localhost:8000/create-cluster' \

This is the json file you have just used (also available on
[GitHub](https://github.com/golemfactory/golem-workers/raw/main/examples/sd-example/create-cluster.json)):
```json
{
"cluster_id": "example",
"payment_config": {
"network": "polygon"
},
"budget_types": {
"default": {
"budget": {
"golem_workers.budgets.AveragePerCpuUsageLinearModelBudget": {
"average_cpu_load": 1,
"average_duration_hours": 0.5,
"average_max_cost": 1.5
}
},
"scope": "cluster"
}
},
"network_types": {
"default": {
"ip": "192.168.0.1/16"
}
},
"node_types": {
"default": {
"market_config": {
"filters": [
{
"golem_reputation.ProviderBlacklistPlugin": {
"payment_network": "polygon"
}
}
],
"sorters": [
{
"golem_reputation.ReputationScorer": {
"payment_network": "polygon"
}
}
]
}
}
}
}
```
{% codefromgithub url="https://github.com/golemfactory/golem-workers/raw/main/examples/sd-example/create-cluster.json" language="json" %}

{% /codefromgithub %}

### `create-cluster` JSON explained

Expand Down Expand Up @@ -177,7 +136,7 @@ will be considered when selecting a provider.
## 3. Create a node

Next, create a node based on the image
[scalepointai/automatic1111:4](https://registry.golem.network/explore/scalepointai/automatic1111)
[modelserve/automatic1111:4](https://registry.golem.network/explore/modelserve/automatic1111)
and specify the path to the model
(e.g. a precached on Golem Network proxy SD model `https://gpu-provider.dev.golem.network/models/v1-5-pruned-emaonly.safetensors` ).

Expand All @@ -189,63 +148,9 @@ curl --location 'http://localhost:8000/create-node' \

This is the json file you have just used (also available on
[GitHub](https://github.com/golemfactory/golem-workers/raw/main/examples/sd-example/create-node.json)):
```json
{
"cluster_id": "example",
"node_networks": {
"default": {}
},
"node_config": {
"market_config": {
"demand": {
"payloads": [
{
"golem_workers.payloads.ClusterNodePayload": {
"runtime": "vm-nvidia",
"image_tag": "scalepointai/automatic1111:4",
"outbound_urls": [
"https://gpu-provider.dev.golem.network"
]
}
}
]
}
},
"on_start_commands": [
{
"golem_workers.work.deploy_and_start_activity": {
"deploy_timeout_minutes": 60
}
},
{
"golem_workers.work.prepare_and_run_ssh_server": {
"ssh_private_key_path": "/tmp/ssh_key"
}
},
{
"golem_workers.work.run_in_shell": [
"cd /usr/src/app/ && ./start.sh --model_url https://gpu-provider.dev.golem.network/models/v1-5-pruned-emaonly.safetensors > /usr/src/app/output/log 2>&1 &"
]
}
],
"sidecars": [
{
"golem_workers.sidecars.WebsocatPortTunnelSidecar": {
"network_name": "default",
"local_port": "8080",
"remote_port": "8000"
}
},
{
"golem_workers.sidecars.WebsocatPortTunnelSidecar": {
"network_name": "default",
"local_port": "8081",
"remote_port": "8001"
}
}
]
}
}
{% codefromgithub url="https://github.com/golemfactory/golem-workers/raw/main/examples/sd-example/create-node.json" language="json" %}

{% /codefromgithub %}

### `create-node` JSON explained

Expand All @@ -257,7 +162,7 @@ This is the json file you have just used (also available on
{
"golem_workers.payloads.ClusterNodePayload": {
"runtime": "vm-nvidia",
"image_tag": "scalepointai/automatic1111:4",
"image_tag": "modelserve/automatic1111:4",
"subnet_tag": "gpu-test",
"outbound_urls": [
"https://gpu-provider.dev.golem.network"
Expand All @@ -272,7 +177,7 @@ This is the json file you have just used (also available on
In this example, we are using the `vm-nvidia` runtime to obtain nodes with GPUs.
Additionally, we configure the node to allow downloading images from **Hugging Face** (which is available only on the testnet)
and our own **gpu-provider.dev.golem.network** proxy. The node will use the image
published to Golem Registry as [scalepointai/automatic1111](https://registry.golem.network/explore/scalepointai/automatic1111)
published to Golem Registry as [modelserve/automatic1111](https://registry.golem.network/explore/modelserve/automatic1111)
in version `4`.

This image includes a `start.sh` script that automatically downloads a model and starts the **Automatic1111 Web UI**.
Expand Down Expand Up @@ -356,13 +261,10 @@ curl --location 'http://localhost:8000/get-node' \

This is the json file you have just used (also available on
[GitHub](https://github.com/golemfactory/golem-workers/raw/main/examples/sd-example/get-node.json)):
```json
{
"cluster_id": "example",
"node_id": "node0"
}
```

{% codefromgithub url="https://github.com/golemfactory/golem-workers/raw/main/examples/sd-example/get-node.json" language="json" %}

{% /codefromgithub %}
It returns the node’s state:

- `provisioning` means you have to wait a bit more for the negotiations with the available nodes to finish
Expand All @@ -385,8 +287,10 @@ To generate some images using the Automatic1111 GUI, you can either enter your p

#### Alternatively, use the automatic1111 API directly

Copy and run the following Python code (directly in the terminal or using a Jupyter Notebook).
The code is also availabla on
Copy and run the following Python code (directly in the terminal or using a Jupyter Notebook).
Note it requires `PIL` library (get it with `pip install Pillow`).

The code is also available on
[GitHub](https://github.com/golemfactory/golem-workers/raw/main/examples/sd-example/sd-inference.py).

```python
Expand Down Expand Up @@ -434,9 +338,6 @@ curl --location 'http://localhost:8000/delete-cluster' \

This is the json file you have just used (also available on
[GitHub](https://github.com/golemfactory/golem-workers/raw/main/examples/hello-example/delete-cluster.json)):
```json
{
"cluster_id": "example"
}
```
{% codefromgithub url="https://github.com/golemfactory/golem-workers/raw/main/examples/hello-example/delete-cluster.json" language="json" %}

{% /codefromgithub %}
Loading

0 comments on commit 129f517

Please sign in to comment.