Skip to content

Box6 cartpole binding pr #199

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

Merged
jsuarez5341 merged 2 commits into from
Apr 21, 2025
Merged

Box6 cartpole binding pr #199

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
a4ffe21
cartpole newbind pr.
Apr 19, 2025
c3f340e
final revisions. note: current state of dev branch has broken continu…
Apr 19, 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
20 changes: 20 additions & 0 deletions pufferlib/ocean/cartpole/binding.c
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,20 @@
#include "cartpole.h"
#define Env Cartpole
#include "../env_binding.h"

static int my_init(Env* env, PyObject* args, PyObject* kwargs) {
env->continuous = unpack(kwargs, "continuous");
init(env);
return 0;
}

static int my_log(PyObject* dict, Log* log) {
assign_to_dict(dict, "score", log->score);
assign_to_dict(dict, "perf", log->perf);
assign_to_dict(dict, "episode_length", log->episode_length);
assign_to_dict(dict, "x_threshold_termination", log->x_threshold_termination);
assign_to_dict(dict, "pole_angle_termination", log->pole_angle_termination);
assign_to_dict(dict, "max_steps_termination", log->max_steps_termination);
assign_to_dict(dict, "n", log->n);
return 0;
}
44 changes: 24 additions & 20 deletions pufferlib/ocean/cartpole/cartpole.c
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,4 +1,4 @@
// local compile/eval not implemented
// local compile/eval implemented for discrete actions only
// eval with python demo.py --mode eval --env puffer_cartpole --eval-mode-path <path to model>

#include <math.h>
Expand All @@ -11,8 +11,8 @@
#define NUM_WEIGHTS 133123
#define OBSERVATIONS_SIZE 4
#define ACTIONS_SIZE 2
#define CONTINUOUS 1
const char* WEIGHTS_PATH = "/puffertank/pufferlib/pufferlib/resources/cartpole/cartpole_weights.bin";
#define CONTINUOUS 0
const char* WEIGHTS_PATH = "/puffertank/test_newbind/pufferlib/pufferlib/resources/cartpole/cartpole_weights.bin";

float movement(int discrete_action, int userControlMode) {
if (userControlMode) {
Expand All @@ -26,14 +26,15 @@ void demo() {
Weights* weights = load_weights(WEIGHTS_PATH, NUM_WEIGHTS);
LinearLSTM* net;

if (CONTINUOUS) {
net = make_linearlstm_float(weights, 1, OBSERVATIONS_SIZE, ACTIONS_SIZE, ACTION_TYPE_FLOAT);
} else {
net = make_linearlstm(weights, 1, OBSERVATIONS_SIZE, ACTIONS_SIZE, ACTION_TYPE_INT);
}
// if (CONTINUOUS) {
// net = make_linearlstm_float(weights, 1, OBSERVATIONS_SIZE, ACTIONS_SIZE);
// } else {
// net = make_linearlstm_int(weights, 1, OBSERVATIONS_SIZE, ACTIONS_SIZE);
// }

CartPole env = {0};
env.continuous = CONTINUOUS;
net = make_linearlstm(weights, 1, OBSERVATIONS_SIZE, ACTIONS_SIZE);
Cartpole env = {0};
env.is_continuous = CONTINUOUS;
allocate(&env);
Client* client = make_client(&env);
c_reset(&env);
Expand All @@ -46,14 +47,17 @@ void demo() {
int userControlMode = IsKeyDown(KEY_LEFT_SHIFT);

if (!userControlMode) {
if (CONTINUOUS) {
forward_linearlstm_float(net, env.observations, env.actions);
env.actions[0] = tanhf(env.actions[0]);
} else {
int action_value;
forward_linearlstm_int(net, env.observations, &action_value);
env.actions[0] = movement(action_value, 0);
}
// if (CONTINUOUS) {
// forward_linearlstm_float(net, env.observations, env.actions);
// env.actions[0] = tanhf(env.actions[0]);
// } else {
// int action_value;
// forward_linearlstm_int(net, env.observations, &action_value);
// env.actions[0] = movement(action_value, 0);
// }
int action_value;
forward_linearlstm(net, env.observations, &action_value);
env.actions[0] = movement(action_value, 0);
} else {
env.actions[0] = movement(env.actions[0], userControlMode);
}
Expand All @@ -64,11 +68,11 @@ void demo() {

BeginDrawing();
ClearBackground(RAYWHITE);
c_render(client, &env);
c_render(&env);
DrawText("Evaluating policy...", 10, 160, 20, DARKGRAY);
EndDrawing();

if (env.dones[0]) {
if (env.terminals[0]) {
printf("Episode done. Steps: %d, Return: %.2f\n\n", episode_steps, episode_return);
episode_steps = 0;
episode_return = 0.0f;
Expand Down
Loading