VizdoomDeathmatch-v1_gradlab-ppo_b0330247

tsilva/VizdoomDeathmatch-v1_gradlab-ppo_b0330247

VizdoomDeathmatch-v1_gradlab-ppo_b0330247 is a reinforcement learning model from tsilva. Use it for the reinforcement learning task on the model card, and read the license before you ship it in a product. It is set up for gradlab. The card lists the license as mit.

An immutable PPO research release for VizdoomDeathmatch-v1 / VizdoomDeathmatch-v1, backed by accepted exact-contract evaluation evidence and a separately labeled representative replay.

Downloads · 30 days

Access

Public

Updated Sep 23, 2026

Repo size

59.1 MB

Likes

Public

Hugging Face

Repo makeup

Click a slice to open those files.

.mp437 MB · 62%

At a glance

Task: Reinforcement Learning
Library: gradlab
License: mit
Access: Public
Created: Aug 7, 2026
Updated: Sep 23, 2026
SHA: 6aeac475

Model card

From the Hugging Face model README

ViZDoom single-player Deathmatch score attack — GradLab PPO @ 464M env steps

An immutable PPO research release for VizdoomDeathmatch-v1 / VizdoomDeathmatch-v1, backed by accepted exact-contract evaluation evidence and a separately labeled representative replay.

Start here

This is the published Policy for the VizdoomDeathmatch-v1 Research Goal. The immutable v3 Release contains the promoted Policy, Acceptance evidence, and a representative replay. Use this exact Release for reproducible Playback:

uvx --from gradlab gradlab play hf://tsilva/VizdoomDeathmatch-v1_gradlab-ppo_b0330247@v3

Browse the ViZDoom environment collection for this Policy and related results. The video is a replay, not Acceptance evidence. The accepted evaluation is documented below and in evaluation_evidence.json.

Identity

Item	Value
Goal	`VizdoomDeathmatch-v1` — ViZDoom single-player Deathmatch score attack
Environment	`VizdoomDeathmatch-v1` (`vizdoom-turbo:VizdoomDeathmatch-v1`)
Trainer	GradLab
Algorithm	PPO
Model class	`gradlab.ppo.GradLabPPO`
Compatibility	SB3-compatible
Checkpoint	Exact step `463970304`
Immutable release	`v3` and `checkpoint-463970304`
Lineage	`b033024760cb83f03f906eb9d49f83a3ba68f8638e2ecaa4879dd17615fc2093`
Environment index	GradLab — VizdoomDeathmatch-v1
Featured	True

Evaluation Status

Status: Accepted under a stochastic full evaluation of 100 episodes.

Metric	Meaning	Observed	Requirement	Outcome
`eval/full/episode/return/shaped/mean`	Full-eval return mean	100.99	`>= 10.0`	Pass

Ranking metrics:

Metric	Meaning	Observed	Direction
`eval/full/episode/return/shaped/mean`	Full-eval return mean	100.99	`max`
`eval/full/episode/return/shaped/max`	Full-eval return max	208	`max`
`leader/checkpoint/step`	Leader checkpoint step	463970304	`min`

The normalized per-episode results, aggregates, rules, contracts, outcomes, and authoritative hashes are in evaluation_evidence.json. Evidence SHA-256: 8a0ae072c5fdcba89a82179878df325a3b975853b426a790d35bf4734deebb7b.

Trainer, Algorithm, and Compatibility

Trainer: GradLab
Algorithm: PPO
Model class: gradlab.ppo.GradLabPPO
Compatibility: SB3-compatible
Hugging Face library: gradlab

Comparable Changes

Comparable: v3 republishes the same checkpoint and exact evaluation contract as v2. Releases are compared only when lineage, run, seed, and evaluation-contract hash all match.

Representative Replay

Watch on YouTube. The root replay.mp4 is one validated training episode. It is representative media and is not evaluation or promotion evidence.

Item	Value
Outcome	`terminated`
Goal success	`False`
Seed	`123`
Start	`default`
Policy steps	`1331`
Episode return	`96.0`
Player runtime	{"asset":null,"contract_mode":"training","device_type":"mps","environment_hash":"sha256:5ce5419576605f9e2f1f532b147067952c8a5876a7bc08f4ef96ccf6eaec412a","execution_target":"local_player","overrides":[],"provider_id":"vizdoom-turbo","provider_version":"1.3.0.post23","qualified_environment_id":"vizdoom-turbo:VizdoomDeathmatch-v1","runtime_image_digest":"docker:ghcr.io/tsilva/gradlab/gradlab-train@sha256:5c8d1e2e09e9781942cc3038344b5ec4cc902ad9e3cf95f1219dd3951104743e","runtime_versions":{"breakout_turbo_env":"0.5.2","stable_baselines3":"2.8.0","stable_retro_turbo":"1.0.1.post37","supermariobrosnes_turbo":"0.6.2","vizdoom_turbo":"1.3.0.post23"},"seed":123,"source":{"distribution":"gradlab","git_commit":"1713894536640a6a27e30bddb787eb6b6a902187","kind":"checkout","source_tree_sha256":"533b692e2afe698221133f1f92cf1086df5723ffb2d0670cb059ffb1635081c5","version":"0.1.1","working_tree_dirty":true}}

Quick Start

uvx --from gradlab gradlab play hf://tsilva/VizdoomDeathmatch-v1_gradlab-ppo_b0330247@v3

The immutable model reference is hf://tsilva/VizdoomDeathmatch-v1_gradlab-ppo_b0330247@v3.

Full Evaluation

Evaluation used stochastic action selection for all 100 declared episodes. See evaluation_evidence.json for every normalized episode and all materialized acceptance and ranking outcomes.

Contracts

Item	Value
Environment hash	`sha256:95a7666af8d6c9bf03df393875fed8bf90d81981a45a647c8ab68c3e0a9eaff8`
Preprocessing and model inputs	{"model_inputs":{"context":{"armor":{"encoding":{"clip":true,"high":1.0,"kind":"continuous","low":0.0,"offset":0.0,"scale":0.005},"signal":"armor","update":"transition"},"health":{"encoding":{"clip":true,"high":2.0,"kind":"continuous","low":0.0,"offset":0.0,"scale":0.01},"signal":"health","update":"transition"},"selected_weapon":{"encoding":{"kind":"categorical","values":[1,2,3,4,5,6]},"signal":"selected_weapon","update":"transition"},"selected_weapon_ammo":{"encoding":{"clip":true,"high":1.0,"kind":"continuous","low":0.0,"offset":0.0,"scale":0.0033333333333333335},"signal":"selected_weapon_ammo","update":"transition"},"weapon_ammo":{"encoding":{"clip":true,"high":1.0,"kind":"continuous","low":0.0,"offset":0.0,"scale":[1.0,0.005,0.02,0.005,0.02,0.0033333333333333335]},"signal":"weapon_ammo","update":"transition"},"weapons_owned":{"encoding":{"clip":true,"high":1.0,"kind":"continuous","low":0.0,"offset":0.0,"scale":1.0},"signal":"weapons_owned","update":"transition"}},"schema_version":1},"preprocessing":{"frame_skip":2,"frame_stack":4,"max_pool_frames":false,"obs_copy":"safe_view","obs_crop":[0,32,0,0],"obs_crop_fill":0,"obs_crop_mode":"mask","obs_grayscale":true,"obs_resize":[84,84],"obs_resize_algorithm":"area","pipeline":"vizdoom_turbo_native_vec_env","policy_observation_layout":"dict_observation_context_v1","sticky_action_prob":0.0}}
Action semantics	{"contract_hash":"f1aa454eeec78c4a1f5ad4f056d1ceec7881d397649cfa0cb505b3b7cd790634","execution_hash":"abfee888d741c472cbab5c9925937b12c34f99f381e243c5264bcb00c689b437","policy":{"codec":{"type":"identity"},"semantics":{"encoding":"explicit","entries":[{"controls":[{"atoms":[],"inputs":[],"player":1}],"label":"noop","semantic_id":"noop","value":0},{"controls":[{"atoms":["attack"],"inputs":["a"],"player":1}],"label":"attack","semantic_id":"attack","value":1},{"controls":[{"atoms":["move_forward"],"inputs":["up"],"player":1}],"label":"move forward","semantic_id":"move_forward","value":2},{"controls":[{"atoms":["move_backward"],"inputs":["down"],"player":1}],"label":"move backward","semantic_id":"move_backward","value":3},{"controls":[{"atoms":["move_left"],"inputs":["left"],"player":1}],"label":"move left","semantic_id":"move_left","value":4},{"controls":[{"atoms":["move_right"],"inputs":["right"],"player":1}],"label":"move right","semantic_id":"move_right","value":5},{"controls":[{"atoms":["turn_left"],"inputs":["turn_left"],"player":1}],"label":"turn left","semantic_id":"turn_left","value":6},{"controls":[{"atoms":["turn_right"],"inputs":["turn_right"],"player":1}],"label":"turn right","semantic_id":"turn_right","value":7},{"controls":[{"atoms":["speed","move_forward"],"inputs":["x","up"],"player":1}],"label":"speed move forward","semantic_id":"speed_move_forward","value":8},{"controls":[{"atoms":["attack","move_forward"],"inputs":["a","up"],"player":1}],"label":"attack move forward","semantic_id":"attack_move_forward","value":9},{"controls":[{"atoms":["attack","move_backward"],"inputs":["a","down"],"player":1}],"label":"attack move backward","semantic_id":"attack_move_backward","value":10},{"controls":[{"atoms":["attack","move_left"],"inputs":["a","left"],"player":1}],"label":"attack move left","semantic_id":"attack_move_left","value":11},{"controls":[{"atoms":["attack","move_right"],"inputs":["a","right"],"player":1}],"label":"attack move right","semantic_id":"attack_move_right","value":12},{"controls":[{"atoms":["attack","turn_left"],"inputs":["a","turn_left"],"player":1}],"label":"attack turn left","semantic_id":"attack_turn_left","value":13},{"controls":[{"atoms":["attack","turn_right"],"inputs":["a","turn_right"],"player":1}],"label":"attack turn right","semantic_id":"attack_turn_right","value":14},{"controls":[{"atoms":["select_next_weapon"],"inputs":["select_next_weapon"],"player":1}],"label":"select next weapon","semantic_id":"select_next_weapon","value":15},{"controls":[{"atoms":["select_prev_weapon"],"inputs":["select_prev_weapon"],"player":1}],"label":"select prev weapon","semantic_id":"select_prev_weapon","value":16}],"status":"available"},"space":{"dtype":"int64","n":17,"start":0,"type":"discrete"}},"provider":{"mode":"custom_discrete","preset":null,"provider_id":"vizdoom-turbo","semantics":{"encoding":"explicit","entries":[{"controls":[{"atoms":[],"inputs":[],"player":1}],"label":"noop","semantic_id":"noop","value":0},{"controls":[{"atoms":["attack"],"inputs":["a"],"player":1}],"label":"attack","semantic_id":"attack","value":1},{"controls":[{"atoms":["move_forward"],"inputs":["up"],"player":1}],"label":"move forward","semantic_id":"move_forward","value":2},{"controls":[{"atoms":["move_backward"],"inputs":["down"],"player":1}],"label":"move backward","semantic_id":"move_backward","value":3},{"controls":[{"atoms":["move_left"],"inputs":["left"],"player":1}],"label":"move left","semantic_id":"move_left","value":4},{"controls":[{"atoms":["move_right"],"inputs":["right"],"player":1}],"label":"move right","semantic_id":"move_right","value":5},{"controls":[{"atoms":["turn_left"],"inputs":["turn_left"],"player":1}],"label":"turn left","semantic_id":"turn_left","value":6},{"controls":[{"atoms":["turn_right"],"inputs":["turn_right"],"player":1}],"label":"turn right","semantic_id":"turn_right","value":7},{"controls":[{"atoms":["speed","move_forward"],"inputs":["x","up"],"player":1}],"label":"speed move forward","semantic_id":"speed_move_forward","value":8},{"controls":[{"atoms":["attack","move_forward"],"inputs":["a","up"],"player":1}],"label":"attack move forward","semantic_id":"attack_move_forward","value":9},{"controls":[{"atoms":["attack","move_backward"],"inputs":["a","down"],"player":1}],"label":"attack move backward","semantic_id":"attack_move_backward","value":10},{"controls":[{"atoms":["attack","move_left"],"inputs":["a","left"],"player":1}],"label":"attack move left","semantic_id":"attack_move_left","value":11},{"controls":[{"atoms":["attack","move_right"],"inputs":["a","right"],"player":1}],"label":"attack move right","semantic_id":"attack_move_right","value":12},{"controls":[{"atoms":["attack","turn_left"],"inputs":["a","turn_left"],"player":1}],"label":"attack turn left","semantic_id":"attack_turn_left","value":13},{"controls":[{"atoms":["attack","turn_right"],"inputs":["a","turn_right"],"player":1}],"label":"attack turn right","semantic_id":"attack_turn_right","value":14},{"controls":[{"atoms":["select_next_weapon"],"inputs":["select_next_weapon"],"player":1}],"label":"select next weapon","semantic_id":"select_next_weapon","value":15},{"controls":[{"atoms":["select_prev_weapon"],"inputs":["select_prev_weapon"],"player":1}],"label":"select prev weapon","semantic_id":"select_prev_weapon","value":16}],"status":"available"},"space":{"dtype":"int64","n":17,"start":0,"type":"discrete"},"table_hash":"0bd9dd28d67a88ef6bc54734f53d55bc4af597e672665a7f20d4b204098036af"},"requested":{"meanings":["noop","attack","move_forward","move_backward","move_left","move_right","turn_left","turn_right","speed_move_forward","attack_move_forward","attack_move_backward","attack_move_left","attack_move_right","attack_turn_left","attack_turn_right","select_next_weapon","select_prev_weapon"],"mode":"custom_discrete","preset":null,"table":[[],["ATTACK"],["MOVE_FORWARD"],["MOVE_BACKWARD"],["MOVE_LEFT"],["MOVE_RIGHT"],["TURN_LEFT"],["TURN_RIGHT"],["SPEED","MOVE_FORWARD"],["ATTACK","MOVE_FORWARD"],["ATTACK","MOVE_BACKWARD"],["ATTACK","MOVE_LEFT"],["ATTACK","MOVE_RIGHT"],["ATTACK","TURN_LEFT"],["ATTACK","TURN_RIGHT"],["SELECT_NEXT_WEAPON"],["SELECT_PREV_WEAPON"]],"table_hash":"0bd9dd28d67a88ef6bc54734f53d55bc4af597e672665a7f20d4b204098036af"},"schema_version":1,"semantic_hash":"7a5a73ecc7f770d673626e5aaec26aa1ad3c0db7c87a1b715c79019b066009d2"}

Provenance

Item	Value
Source	gradlab
Run	gradlab-294ea8ebb1c7c049caedfb29221eca2f
Recipe	`ppo`
Seed	`123`
Source commit	`bfad768f66d68d3ddc01f69691ef1e5dbab9a931`
Evaluated artifact	`https://pub-fc35c0b186ce4aad8eea5e93d38c99db.r2.dev/runs/gradlab-294ea8ebb1c7c049caedfb29221eca2f/checkpoints/463970304-e5596d939cef0b6d34e75874b4d917660e7b6efab672ad8d7bea85445a7bb100/model.zip`

Limitations

Evaluation establishes performance only for the published environment, policy, action-selection, start, boundary, and evaluation contracts. It does not establish generalization to other tasks or contracts. The representative replay cannot establish or change acceptance, ranking, promotion, or featured status.

Files

File	Purpose
`model.zip`	Immutable policy checkpoint
`model.json`	Trainer, algorithm, model class, checkpoint identity, and provenance
`recipe.json`	Materialized training, playback, and evaluation contracts
`evaluation_evidence.json`	Normalized authoritative evaluation evidence and rule outcomes
`release_manifest.json`	Release identity, evaluation evidence, and artifact hashes
`replay.mp4`	Browser-safe representative completed episode
`LICENSE`	License for GradLab-authored weights and publication material

Licensing

GradLab-authored policy weights and publication material are licensed under the MIT License in LICENSE. Emulator/runtime software and game assets remain governed by their own licenses and terms. This repository does not redistribute a game ROM.

Task: Reinforcement Learning
Library: gradlab
License: mit
Created: Aug 7, 2026
Updated: Sep 23, 2026