You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Command line interface and SDKs for managing Prime Intellect GPU resources, sandboxes, and environments.
23
+
Command line interface and SDKs for Prime Lab, Hosted Training, GPU resources, sandboxes, and environments.
24
24
</div>
25
25
26
26
## Quick Start
@@ -35,6 +35,16 @@ uv tool install prime
35
35
# Authenticate
36
36
prime login
37
37
38
+
# Set up a Lab workspace for environments, evals, GEPA, and Hosted Training
39
+
prime lab setup
40
+
41
+
# See available Hosted Training models, capacity, and pricing
42
+
prime train models
43
+
44
+
# Generate and launch a Hosted Training config
45
+
prime train init
46
+
prime train rl.toml
47
+
38
48
# Browse verified environments
39
49
prime env list
40
50
@@ -44,6 +54,8 @@ prime availability list
44
54
45
55
## Features
46
56
57
+
-**Lab Workspaces** - Set up local verifiers workspaces for environments, evals, GEPA, and training
58
+
-**Hosted Training** - Train models against verifiers environments and inspect runs, logs, metrics, and checkpoints
47
59
-**Environments** - Access hundreds of verified environments on our community hub
48
60
-**Evaluations** - Push and manage evaluation results
49
61
-**GPU Resource Management** - Query and filter available GPU resources
@@ -124,6 +136,9 @@ prime env list
124
136
# View environment details
125
137
prime env info <environment-name>
126
138
139
+
# Inspect environment source without downloading the archive
140
+
prime env inspect <environment-name>
141
+
127
142
# Install an environment locally
128
143
prime env install <environment-name>
129
144
@@ -132,6 +147,30 @@ prime env init my-environment
132
147
prime env push my-environment
133
148
```
134
149
150
+
### Lab and Hosted Training
151
+
152
+
Prime Lab connects verifiers environments to evaluations, GEPA prompt optimization, and Hosted Training. Start with `prime lab setup` to create a local workspace with starter configs, then use `prime train models` to choose a Hosted Training model with current capacity and pricing.
153
+
154
+
```bash
155
+
# Set up a Lab workspace
156
+
prime lab setup
157
+
158
+
# List trainable models, capacity, and token pricing
0 commit comments