tiny_world_model/Makefile at main · msalvaris/tiny_world_model · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
.DEFAULT_GOAL := help # Sets default action to be help


define PRINT_HELP_PYSCRIPT # start of Python section
import re, sys

output = []
# Loop through the lines in this file
for line in sys.stdin:
    # if the line has a command and a comment start with
    #   two pound signs, add it to the output
    match = re.match(r'^([a-zA-Z_-]+):.*?## (.*)$$', line)
    if match:
        target, help = match.groups()
        output.append("%-10s %s" % (target, help))
# Sort the output in alphanumeric order
output.sort()
# Print the help result
print('\n'.join(output))
endef
export PRINT_HELP_PYSCRIPT # End of python section


# Set the Docker image name and tag
IMAGE_NAME := tiny_world_model
IMAGE_TAG := latest
SAVE_NAME := tworldm.tar

# Set the Docker container name
CONTAINER_NAME := tinyworld

# Set any environment variables needed for the container
ENV_VARS := -e ENV_VAR_1=value1 -e ENV_VAR_2=value2

# Set the Docker build context (the directory containing the Dockerfile)
BUILD_CONTEXT := .

# Set any additional flags for the Docker build command
BUILD_FLAGS :=

# Set any additional flags for the Docker run command
RUN_FLAGS := -p 8080:80 -v /home/mat/repos/tiny_world_model:/src -v /home/mat/data:/data
#GPU_RUN_FLAGS := -p 8080:80 -v /home/ubuntu/repos/tiny_world_model:/src -v /home/ubuntu/_data:/data -v /home/ubuntu/data1:/cloud_data
GPU_RUN_FLAGS := -p 8080:80 -v /home/ubuntu/repos/tiny_world_model:/src -v /home/ubuntu/_data:/data

help:
	@python -c "$$PRINT_HELP_PYSCRIPT" < $(MAKEFILE_LIST)

prepare: ## Prepare lambda box for use
	sudo usermod -aG docker $(USER)
	newgrp docker
	mkdir /home/ubuntu/_data

## Docker commands
build: ## Build docker repo
	docker build $(BUILD_FLAGS) -t $(IMAGE_NAME):$(IMAGE_TAG) $(BUILD_CONTEXT)

run-local: ## Target to run the Docker container locally to test
	docker run --rm -it $(ENV_VARS) $(RUN_FLAGS) --name $(CONTAINER_NAME) $(IMAGE_NAME):$(IMAGE_TAG)

run-gpu: ## Target to run the Docker container on GPU
	docker run --rm -it --gpus all $(ENV_VARS) $(GPU_RUN_FLAGS) --name $(CONTAINER_NAME) $(IMAGE_NAME):$(IMAGE_TAG)

stop: ## Target to stop and remove the Docker container
	docker stop $(CONTAINER_NAME) || true
	docker rm $(CONTAINER_NAME) || true


clean: stop ## Target to remove the Docker image
	docker rmi $(IMAGE_NAME):$(IMAGE_TAG)

save: ## Save the image locally
	docker save --output ${SAVE_NAME} $(IMAGE_NAME):$(IMAGE_TAG)

load: ## Load the local image
	docker load --input ${SAVE_NAME}


install: ## Install the tiny world model repo. Run from the src directory
	pip3 install -e .


# Called inside container
dataset: ## Create the ball dataset
	create-dataset ball "/data/ball_dataset" 1000

train-example: ## Simple training example
	train --data.dataset_dir="/data/ball_dataset" --system.work_dir="/data/experiments"

train: ## Train model
	train --data.dataset_dir="/data/ball_dataset" --system.work_dir="/data/experiments" --trainer.batch_size=128 --trainer.num_epochs=200

movie: ## Generate world model movie using trained model
	generate movie "/data/experiments/experiment_ball_2024-06-01_15:40:45/checkpoints/trained_model.pt" "/data/ball_dataset" "/data/movies"

.PHONY: help build run stop clean save load