Skip to content

Add onlyfish bin #406

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
BoxingBytes wants to merge 18 commits into
base: 3.0
Choose a base branch
from
Open

Add onlyfish bin #406

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
18 commits
Select commit Hold shift + click to select a range
455f65c
Add onlyfish bin
BoxingBytes Oct 28, 2025
3411ff3
add predprey env test
BoxingBytes Oct 29, 2025
654e003
better logs
BoxingBytes Oct 29, 2025
e33d40b
fix data mismatch predprey
BoxingBytes Oct 30, 2025
a6a868f
fix max food to avoid infinite loop on spawning agents
BoxingBytes Oct 30, 2025
4b52316
put back 1000 max timesteps
BoxingBytes Oct 30, 2025
632a888
rework predprey env
BoxingBytes Nov 1, 2025
5c991de
add diff poses for rendering
BoxingBytes Nov 1, 2025
758eb52
improve logs & agents spawn in house
BoxingBytes Nov 3, 2025
326788a
add security check in step
BoxingBytes Nov 3, 2025
b598e7a
update predprey conf
BoxingBytes Nov 3, 2025
ad1b04f
fix obs
BoxingBytes Nov 3, 2025
2dd5c66
rm collect reward
BoxingBytes Nov 3, 2025
4283bf3
add randomization on start and teleportation
BoxingBytes Nov 4, 2025
0a8e134
add reward parameters for sweep
BoxingBytes Nov 7, 2025
0fb6067
use best sweeps in conf
BoxingBytes Nov 9, 2025
d4efa13
upgrade predprey to new version
BoxingBytes Nov 21, 2025
1d63553
update init
BoxingBytes Nov 21, 2025
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
74 changes: 74 additions & 0 deletions pufferlib/config/ocean/predprey.ini
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,74 @@
[base]
package = ocean
env_name = puffer_predprey
rnn_name = Recurrent


[vec]
num_envs = 8

[env]
num_envs = 64
vision = 3
num_agents = 4
report_interval = 1
reward_death_scale = 0.14320154190448353
reward_eat = 0.6301939255961027
reward_collect = 1
timestep_reward = -0.001012632066429986
hp_reward_scale = 0.07674633247055918
held_food_reward_scale = 0.2643837513970884
reward_fireplace_lit = 0.7
reward_store_chest = 0.5

[train]
total_timesteps = 100_000_000
checkpoint_interval = 100
adam_beta1 = 0.9925640021442416
adam_beta2 = 0.9
adam_eps = 6.225983651908837e-10
bptt_horizon = 64
clip_coef = 1
ent_coef = 0.000999567018772538
gae_lambda = 0.9948939854010467
gamma = 0.9997469057538332
learning_rate = 0.002574057351505564
max_grad_norm = 1.8944338753964156
max_minibatch_size = 32768
minibatch_size = 65536
prio_alpha = 0.9328200510590207
prio_beta0 = 0.9225942853355249
vf_clip_coef = 0.1
vf_coef = 0.6935921910790133
vtrace_c_clip = 2.819851610841173
vtrace_rho_clip = 5


[sweep]
method = Protein
metric = score
goal = maximize
downsample = 5
use_gpu = True
prune_pareto = True

[sweep.env.reward_fireplace_lit]
distribution = uniform
min = -1.0
max = 1.0
mean = 0.5
scale = auto

[sweep.env.reward_store_chest]
distribution = uniform
min = -1.0
max = 1.0
mean = 0.5
scale = auto

[sweep.train.total_timesteps]
distribution = log_normal
min = 3e7
max = 1e10
mean = 2e8
scale = time
1 change: 1 addition & 0 deletions pufferlib/ocean/environment.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -163,6 +163,7 @@ def make_multiagent(buf=None, **kwargs):
'spaces': make_spaces,
'multiagent': make_multiagent,
'slimevolley': 'SlimeVolley',
'predprey': 'PredPrey',
}

def env_creator(name='squared', *args, **kwargs):
Expand Down
30 changes: 30 additions & 0 deletions pufferlib/ocean/predprey/binding.c
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,30 @@
#include "predprey.h"

#define Env PredPrey
#include "../env_binding.h"

static int my_init(Env* env, PyObject* args, PyObject* kwargs) {
env->width = unpack(kwargs, "width");
env->height = unpack(kwargs, "height");
env->num_agents = unpack(kwargs, "num_agents");
env->vision = unpack(kwargs, "vision");
env->reward_death_scale = unpack(kwargs, "reward_death_scale");
env->reward_eat = unpack(kwargs, "reward_eat");
env->reward_collect = unpack(kwargs, "reward_collect");
env->timestep_reward = unpack(kwargs, "timestep_reward");
env->hp_reward_scale = unpack(kwargs, "hp_reward_scale");
env->held_food_reward_scale = unpack(kwargs, "held_food_reward_scale");
env->reward_fireplace_lit = unpack(kwargs, "reward_fireplace_lit");
env->reward_store_chest = unpack(kwargs, "reward_store_chest");
init_cenv(env);
return 0;
}

static int my_log(PyObject* dict, Log* log) {
assign_to_dict(dict, "perf", log->perf);
assign_to_dict(dict, "score", log->score);
assign_to_dict(dict, "episode_return", log->episode_return);
assign_to_dict(dict, "steals", log->steals);
assign_to_dict(dict, "collects", log->collects);
return 0;
}
42 changes: 42 additions & 0 deletions pufferlib/ocean/predprey/grid.h
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,42 @@
#ifndef GRID_H
#define GRID_H

#define GRID_HEIGHT 32
#define GRID_WIDTH 32

static const unsigned char grid[GRID_HEIGHT][GRID_WIDTH] = {
{0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02},
{0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02},
{0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02},
{0x02, 0x02, 0x02, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x02, 0x02, 0x02},
{0x02, 0x02, 0x02, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x02, 0x02, 0x02},
{0x02, 0x02, 0x02, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x02, 0x02, 0x02},
{0x02, 0x02, 0x02, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x02, 0x02, 0x02},
{0x02, 0x02, 0x02, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x02, 0x02, 0x02},
{0x02, 0x02, 0x02, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x02, 0x02, 0x02},
{0x02, 0x02, 0x02, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x02, 0x02, 0x02},
{0x02, 0x02, 0x02, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x02, 0x02, 0x02},
{0x02, 0x02, 0x02, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x02, 0x02, 0x02},
{0x02, 0x02, 0x02, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x02, 0x02, 0x02},
{0x02, 0x02, 0x02, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x02, 0x02, 0x02},
{0x02, 0x02, 0x02, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x02, 0x02, 0x02},
{0x02, 0x02, 0x02, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x02, 0x02, 0x02},
{0x02, 0x02, 0x02, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x02, 0x02, 0x02},
{0x02, 0x02, 0x02, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x02, 0x02, 0x02},
{0x02, 0x02, 0x02, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x02, 0x02, 0x02},
{0x02, 0x02, 0x02, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x02, 0x02, 0x02},
{0x02, 0x02, 0x02, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x02, 0x02, 0x02},
{0x02, 0x02, 0x02, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x02, 0x02, 0x02},
{0x02, 0x02, 0x02, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x02, 0x02, 0x02},
{0x02, 0x02, 0x02, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x02, 0x02, 0x02},
{0x02, 0x02, 0x02, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x02, 0x02, 0x02},
{0x02, 0x02, 0x02, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x02, 0x02, 0x02},
{0x02, 0x02, 0x02, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x02, 0x02, 0x02},
{0x02, 0x02, 0x02, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x02, 0x02, 0x02},
{0x02, 0x02, 0x02, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x02, 0x02, 0x02},
{0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02},
{0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02},
{0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02}
};

#endif // GRID_H
175 changes: 175 additions & 0 deletions pufferlib/ocean/predprey/predprey.c
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,175 @@
#include <raylib.h>
#include <unistd.h>
#include <stdlib.h>
// #include "puffernet.h"
#include "predprey.h"

int main() {
srand(43);
PredPrey env = {
.num_agents = 4,
.width = 32,
.height = 32,
.vision = 3,
.reward_death_scale = 1.0f,
.reward_eat = 0.0f,
.reward_collect = 0.0f,
.timestep_reward = 0.0f,
.hp_reward_scale = 0.0f,
.held_food_reward_scale = 0.0f,
};
allocate_cenv(&env);
c_reset(&env);
c_render(&env);

// Weights* weights = load_weights("resources/cpr/cpr_weights.bin", 139270);
// int logit_sizes[] = {5};
// LinearLSTM* net = make_linearlstm(weights, env.num_agents, 49, logit_sizes, 1);
while (!WindowShouldClose()) {

for (int i = 0; i < env.num_agents; i++) {
env.actions[i] = rand() % 7;
}

// User can take control of the first puffer
if (IsKeyDown(KEY_LEFT_SHIFT)) {
sleep(1);
env.actions[0] = NO_MOVE;
if (IsKeyDown(KEY_UP) || IsKeyDown(KEY_W))
env.actions[0] = UP;
if (IsKeyDown(KEY_DOWN) || IsKeyDown(KEY_S))
env.actions[0] = DOWN;
if (IsKeyDown(KEY_LEFT) || IsKeyDown(KEY_A))
env.actions[0] = LEFT;
if (IsKeyDown(KEY_RIGHT) || IsKeyDown(KEY_D))
env.actions[0] = RIGHT;
if (IsKeyDown(KEY_C))
env.actions[0] = INTERACT;
if (IsKeyDown(KEY_E))
env.actions[0] = EAT;

printf("Getting user input %d\n", env.actions[0]);
} else {
// for (int i = 0; i < env.num_agents*49; i++) {
// net->obs[i] = env.observations[i];
// }
// forward_linearlstm(net, net->obs, env.actions);
}

c_step(&env);
c_render(&env);

}
free_CEnv(&env);
close_renderer(env.client);

return 0;
}
/////////////////////
// Reset stress test
/////////////////////

// int main() {
// srand(43);
// PredPrey env = {
// .num_agents = 4,
// .width = 32,
// .height = 32,
// .vision = 3,
// .reward_death_scale = 1.0f,
// .reward_eat = 0.0f,
// .reward_collect = 0.0f,
// .timestep_reward = 0.0f,
// .reward_steal = 0.0f,
// .hp_reward_scale = 0.0f,
// .held_food_reward_scale = 0.0f,
// .food_base_spawn_rate = 1e-1,
// };
// for (int i = 0; i < 1000; i++) {
// allocate_cenv(&env);
// c_reset(&env);
// c_render(&env);

// // Weights* weights = load_weights("resources/cpr/cpr_weights.bin", 139270);
// // int logit_sizes[] = {5};
// // LinearLSTM* net = make_linearlstm(weights, env.num_agents, 49, logit_sizes, 1);
// for (int i = 0; i < 10; i++) {

// for (int i = 0; i < env.num_agents; i++) {
// env.actions[i] = rand() % 7;
// }

// // User can take control of the first puffer
// if (IsKeyDown(KEY_LEFT_SHIFT)) {
// sleep(1);
// env.actions[0] = NO_MOVE;
// if (IsKeyDown(KEY_UP) || IsKeyDown(KEY_W))
// env.actions[0] = UP;
// if (IsKeyDown(KEY_DOWN) || IsKeyDown(KEY_S))
// env.actions[0] = DOWN;
// if (IsKeyDown(KEY_LEFT) || IsKeyDown(KEY_A))
// env.actions[0] = LEFT;
// if (IsKeyDown(KEY_RIGHT) || IsKeyDown(KEY_D))
// env.actions[0] = RIGHT;
// if (IsKeyDown(KEY_C))
// env.actions[0] = INTERACT;
// if (IsKeyDown(KEY_E))
// env.actions[0] = EAT;

// printf("Getting user input %d\n", env.actions[0]);
// } else {
// // for (int i = 0; i < env.num_agents*49; i++) {
// // net->obs[i] = env.observations[i];
// // }
// // forward_linearlstm(net, net->obs, env.actions);
// }

// c_step(&env);
// c_render(&env);

// }
// close_renderer(env.client);
// free_CEnv(&env);
// }
// return 0;
// }

////////////////
// For profile
////////////////
// #include <raylib.h>
// #include <unistd.h>
// #include <stdlib.h>
// #include "predprey.h"

// int main() {
// PredPrey env = {
// .num_agents = 4,
// .width = 32,
// .height = 32,
// .vision = 3,
// .reward_food = 0.0f,
// .food_base_spawn_rate = 1e-1,
// };
// allocate_cenv(&env);
// c_reset(&env);

// long i = 0;
// while (true) {

// for (int i = 0; i < env.num_agents; i++) {
// env.actions[i] = rand() % 7;
// }

// c_step(&env);

// i++;
// if (i > 10000000) {
// printf("breaking");
// break;
// }
// }
// free_CEnv(&env);

// return 0;
// }
Loading
Loading