Bundle Creation

Commands for creating and validating bundle requests in the LEAP Bundle CLI

create

Create a new bundle request by uploading a model directory.

leap-bundle create <input-path>

Arguments

• input-path (required): Path to the directory to upload and bundle

Options

• --dry-run: Validate the input directory without actually creating a request or uploading files.
• --json: Output result in JSON format for programmatic parsing.
• --parallel: Upload files in parallel with multipart support. This is the default behavior if neither --parallel nor --sequential is specified.
• --sequential: Upload files sequentially. This is the fallback option if parallel upload fails.
  • If neither --parallel nor --sequential is specified, the CLI will attempt parallel upload first, and fall back to sequential if it fails.
  • If both --parallel and --sequential are specified, --parallel takes precedence.
• --executorch (deprecated): Use ExecuTorch bundling instead of GGUF. By default, the CLI uses GGUF bundling. This option may be removed in a future version.
• --quantization <type>: Specify the quantization type for the model bundle. Options include Q4_K_M (default), Q8_0, F16, and other llama.cpp quantization types.
• --mmproj-quantization <type>: (GGUF only) Specify the mmproj quantization type for vision-language or audio models. Valid options: q4, q8 (default), f16.

Behavior

• Calculates a hash of the directory contents
• Creates a bundle request on the LEAP platform
• Uploads the directory contents to cloud storage
• Updates the request status throughout the process
• If a request with the same hash already exists, shows a warning instead of creating a duplicate

Examples

# Create bundle from directory
leap-bundle create ./my-model-directory

# Example output on success
ℹ Calculating directory hash...
ℹ Submitting bundle request...
✓ Bundle request created with ID: 18734
ℹ Starting upload...
Uploading directory... ✓
✓ Upload completed successfully! Request ID: 18734

# Example output when request already exists
⚠ A bundle request with the same input hash already exists: req_xyz789abc123

# Create bundle with JSON output
leap-bundle create ./my-model-directory --json

# Example JSON output on success
{"request_id": "18734", "status": "success"}

# Example JSON output when request already exists
{"error": "A bundle request with the same input hash already exists: req_xyz789abc123", "status": "exists"}

# Create GGUF bundle with specific quantization
leap-bundle create ./my-model-directory --quantization Q8_0

# Create GGUF bundle for VL model with mmproj quantization
leap-bundle create ./my-vl-model-directory --mmproj-quantization f16

Legacy: ExecuTorch Bundling

ExecuTorch bundling is deprecated and may be removed in a future version. Use GGUF bundling (the default) for new projects.

ExecuTorch quantization options: 8da4w_output_8da8w (default), 8da8w_output_8da8w

Examples:

# Create ExecuTorch bundle
leap-bundle create ./my-model-directory --executorch

# Create ExecuTorch bundle with specific quantization
leap-bundle create ./my-model-directory --executorch --quantization 8da8w_output_8da8w

ExecuTorch bundling produces .bundle files instead of .gguf files.

Validation

• Hidden files and subdirectories (starting with .) are ignored
• The directory must have a valid config.json file
• The directory must have one or more safe tensor files with the .safetensors extension
• Individual files must not exceed 10GB
• The total size of the directory must not exceed 10GB
  • This limit will be increased in the future as the bundle service supports more models

Error Cases

• Directory doesn’t exist: Shows file not found error
• Path is not a directory: Shows invalid path error
• Not authenticated: Prompts to login first
• Upload failure: Shows upload error and updates request status to failed
  • Upload will timeout in 10 minutes if not completed. We are working on optimizing upload performance.
  • For transient issues, try resuming the upload: leap-bundle resume <request-id>
  • For persistent issues, cancel the request and create a new one
• Network issues: Shows connection error

Process Flow

1. Validates input directory exists
2. Calculates directory hash for deduplication
3. Creates bundle request via API
4. Uploads directory contents to S3 using signed URLs
5. Updates request status to track progress

Large directories may take significant time to upload. The upload process can be interrupted with Ctrl+C.

If the upload fails due to transient issue (e.g. network instability), you can use the resume command to continue from where it left off.

Rate Limits

Account type	Max requests per 24 hours
Free tier	5
Enterprise tier	Coming soon

Processing Time

If the request is not rate limited, it will usually be completed within 20 minutes for small model checkpoints.

resume

Resume uploading for a failed or interrupted bundle request.

leap-bundle resume [<request-id>]

Arguments

• request-id (optional): ID of the bundle request to resume. If not provided, resumes the latest request.

Options

• --json: Output result in JSON format for programmatic parsing
• --sequential: Upload files sequentially (fallback option if parallel upload fails)
• --parallel: Upload files in parallel with multipart support (default behavior if neither option is specified)

Behavior

• Retrieves the specified bundle request (or latest request if no ID provided)
• Validates that the request is in a resumable status (“Uploading” or “Uploading Failed”)
• Checks that the original input directory still exists
• Resumes the upload process from where it left off
• Updates the request status throughout the process
• Supports both parallel and sequential upload mechanisms

Examples

# Resume the latest failed upload
leap-bundle resume

# Example output on success
ℹ Retrieving latest bundle request...
✓ Resuming upload for request 18734...
ℹ Input path: /path/to/model/directory
ℹ Using parallel upload with multipart support...
✓ Upload completed successfully! Request ID: 18734

# Resume a specific request by ID
leap-bundle resume 18734

# Example output
ℹ Retrieving bundle request 18734...
✓ Resuming upload for request 18734...
ℹ Input path: /path/to/model/directory
ℹ Using parallel upload with multipart support...
✓ Upload completed successfully! Request ID: 18734

# Resume with JSON output
leap-bundle resume 18734 --json

# Example JSON output on success
{"request_id": "18734", "status": "success"}

# Example JSON output on error
{"error": "Resume only works for model uploading. Current status: completed"}

Resumable Statuses

The resume command only works for requests in the following statuses:

• Uploading: Upload was interrupted before the status was updated
• Uploading Failed: Upload encountered an error

Upload Mechanisms

• Parallel (default): Uses multipart upload with better performance for large files
• Sequential: Uploads files one by one, used as fallback if parallel upload fails
• If parallel upload fails, the command automatically falls back to sequential upload
• You can force sequential upload using the --sequential flag

Error Cases

• Not authenticated: Prompts to login first
• Request not found: Shows request not found error
• Request not resumable: Shows status error (only “Uploading” and “Uploading Failed” requests can be resumed)
• Original directory missing: Shows directory not found error
• Upload failure: Shows upload error and updates request status
• Network issues: Shows connection error

Use resume to continue interrupted uploads without losing progress. The command will pick up where the upload left off, saving time on large model directories.

The original input directory must still exist at the same path for resume to work. If you’ve moved or deleted the directory, the resume will fail.

validate

Validate a directory for bundle creation without uploading or creating a request. This command is equivalent to the create command with --dry-run option, but does not require authentication.

leap-bundle validate <input-path>

Arguments

• input-path (required): Path to the directory to validate

Behavior

• Performs the same validation checks as the create command
• Does not require authentication
• Does not upload files or create bundle requests
• Useful for checking if a directory is ready for bundling before actually creating a request

Examples

# Validate a model directory
leap-bundle validate ./my-model-directory

# Example output on success
✓ Directory validation passed
✓ Directory is ready for bundle creation

# Example output on failure
✗ Validation failed: No config.json found in directory.

Validation

The same validation rules apply as for the create command:

• Hidden files and subdirectories (starting with .) are ignored
• The directory must have a valid config.json file
• The directory must have one or more safe tensor files with the .safetensors extension
• Individual files must not exceed 10GB
• The total size of the directory must not exceed 10GB

Error Cases

• Directory doesn’t exist: Shows file not found error
• Path is not a directory: Shows invalid path error
• Validation failure: Shows specific validation error message

Use validate to check your directory before running create to catch validation issues early without using your bundle request quota.