← Back to scan
File as GitHub Issue repo: huggingface/accelerate

Push this scan report to huggingface/accelerate

Click the green button below to open GitHub’s new-issue form, pre-filled with the report title, summary table, top findings, and an embedded score-card image. No authentication needed — you review on GitHub before submitting. Repobility is credited as the scanner.

Embedded score card image

This image will render at the top of the issue body. Hosted on Repobility, refreshes automatically after re-scans.

Repobility score card

Issue title

Bare Except Pass

Curate findings to include

Pick exactly which findings appear in the issue body. By default the top 5 are included. Uncheck noise, check what matters.

Top 5 (default)
Severity Rule Title File:line
CRIT MINED030 [MINED030] Python Pickle Loads: pickle.loads() can execute arbitrary code via __reduce__. src/accelerate/utils/operations.py:484
CRIT MINED018 [MINED018] Unsafe Deserialization Pickle: pickle.loads / yaml.load (without Loader=SafeLo… src/accelerate/utils/operations.py:484
CRIT SEC081 [SEC081] Python: pickle.loads / marshal.loads on untrusted data: pickle.load(s) and marsh… src/accelerate/utils/operations.py:484
HIGH MINED001 [MINED001] Bare Except Pass: except: pass or except Exception: pass — silently swallows e… src/accelerate/utils/environment.py:223
HIGH MINED001 [MINED001] Bare Except Pass: except: pass or except Exception: pass — silently swallows e… src/accelerate/test_utils/testing.py:822
HIGH MINED001 [MINED001] Bare Except Pass: except: pass or except Exception: pass — silently swallows e… src/accelerate/commands/menu/selection_…:32
HIGH MINED021 [MINED021] Path Traversal Os Join: os.path.join(user_dir, filename) where filename can co… src/accelerate/utils/fsdp_utils.py:194
HIGH MINED021 [MINED021] Path Traversal Os Join: os.path.join(user_dir, filename) where filename can co… src/accelerate/commands/config/config_a…:30
HIGH MINED021 [MINED021] Path Traversal Os Join: os.path.join(user_dir, filename) where filename can co… src/accelerate/accelerator.py:3787
HIGH MINED004 [MINED004] Weak Crypto: MD5/SHA1/DES/RC4 used for security context (not just checksums). examples/inference/distributed/florence…:88
HIGH MINED020 [MINED020] Logging Credential Via Fstring: logger.error(f"failed for {api_key}") — secret… examples/by_feature/megatron_lm_gpt_pre…:418
HIGH MINED020 [MINED020] Logging Credential Via Fstring: logger.error(f"failed for {api_key}") — secret… examples/by_feature/deepspeed_with_conf…:430
HIGH SEC128 [SEC128] Async function without await — fire-and-forget Promise (AI mistake): Async call … benchmarks/fp8/transformer_engine/distr…:115
HIGH SEC128 [SEC128] Async function without await — fire-and-forget Promise (AI mistake): Async call … benchmarks/fp8/torchao/distrib_deepspee…:128
HIGH SEC128 [SEC128] Async function without await — fire-and-forget Promise (AI mistake): Async call … benchmarks/fp8/ms_amp/distrib_deepspeed…:90
HIGH SEC103 [SEC103] LDAP injection — non-constant search filter: User input concatenated into an LDA… src/accelerate/utils/modeling.py:174
HIGH SEC114 [SEC114] path.join / Path() on user-controlled segment without containment check: filepat… src/accelerate/utils/fsdp_utils.py:194
HIGH SEC114 [SEC114] path.join / Path() on user-controlled segment without containment check: filepat… src/accelerate/tracking.py:237
HIGH SEC114 [SEC114] path.join / Path() on user-controlled segment without containment check: filepat… src/accelerate/accelerator.py:3787
HIGH SEC020 [SEC020] Secret Printed to Logs: Debug or diagnostic code appears to print a credential-b… benchmarks/big_model_inference/big_mode…:129
HIGH SEC013 [SEC013] Path Traversal — User Input in File Path: User-controlled input used in file pat… src/accelerate/utils/fsdp_utils.py:194
HIGH SEC013 [SEC013] Path Traversal — User Input in File Path: User-controlled input used in file pat… src/accelerate/tracking.py:237
HIGH SEC013 [SEC013] Path Traversal — User Input in File Path: User-controlled input used in file pat… src/accelerate/accelerator.py:3787
MED SEC127 [SEC127] AI agent stub — TODO: implement / pass placeholder body: Function body left as T… src/accelerate/utils/dataclasses.py:1565
MED SEC007 [SEC007] Unsafe Deserialization: Unsafe deserialization can execute arbitrary code. src/accelerate/utils/operations.py:484
MED ERR001 [ERR001] Silent Exception Swallowing: Silently swallowing all exceptions hides bugs. Even… src/accelerate/utils/environment.py:223
MED SEC123 [SEC123] Production stack trace / debug output exposed: Debug mode left on in production … src/accelerate/launchers.py:308
MED SEC011 [SEC011] Unsafe PyTorch Model Loading: torch.load() uses pickle internally and can execut… src/accelerate/utils/other.py:436
MED SEC011 [SEC011] Unsafe PyTorch Model Loading: torch.load() uses pickle internally and can execut… src/accelerate/accelerator.py:3823
MED SEC034 [SEC034] Log Injection / Log Forging — unsanitized user input in log: User input is logge… src/accelerate/accelerator.py:3796
MED SEC034 [SEC034] Log Injection / Log Forging — unsanitized user input in log: User input is logge… examples/by_feature/megatron_lm_gpt_pre…:425
MED SEC034 [SEC034] Log Injection / Log Forging — unsanitized user input in log: User input is logge… examples/by_feature/deepspeed_with_conf…:437
MED SEC045 [SEC045] eval()/exec() on stored or user-supplied data: eval() and exec() on data — even … src/accelerate/utils/megatron_lm.py:1003
MED SEC045 [SEC045] eval()/exec() on stored or user-supplied data: eval() and exec() on data — even … src/accelerate/optimizer.py:138
MED DKR003 Dockerfile base image uses the latest tag docker/accelerate-gpu-deepspeed/Dockerf…:5
MED DKR003 Dockerfile base image uses the latest tag docker/accelerate-gpu/Dockerfile:5
MED DKR007 Docker build context has no .dockerignore .dockerignore
MED DKR002 Dockerfile base image has no explicit tag benchmarks/fp8/ms_amp/Dockerfile:1
MED DKR009 Dockerfile separates apt update from install docker/accelerate-cpu/Dockerfile:8
MED DKR001 Docker final stage has no non-root USER docker/accelerate-gpu-deepspeed/Dockerf…:32
MED DKR001 Docker final stage has no non-root USER docker/accelerate-gpu/Dockerfile:32
MED DKR001 Docker final stage has no non-root USER benchmarks/fp8/transformer_engine/Docke…:4
MED DKR001 Docker final stage has no non-root USER benchmarks/fp8/torchao/Dockerfile:1
MED DKR001 Docker final stage has no non-root USER benchmarks/fp8/ms_amp/Dockerfile:1
LOW COMP001 [COMP001] High cognitive complexity: Function `train_baseline` has cognitive complexity 9… benchmarks/fp8/torchao/distrib_deepspee…:51
LOW COMP001 [COMP001] High cognitive complexity: Function `main` has cognitive complexity 8 (SonarSou… benchmarks/big_model_inference/big_mode…:79
LOW COMP001 [COMP001] High cognitive complexity: Function `parse_args` has cognitive complexity 11 (S… benchmarks/big_model_inference/big_mode…:43
LOW AIC003 Duplicated implementation block across source files src/accelerate/commands/config/sagemake…:140
LOW AIC003 Duplicated implementation block across source files manim_animations/dataloaders/stage_7.py:120
LOW AIC003 Duplicated implementation block across source files manim_animations/dataloaders/stage_7.py:6
LOW AIC003 Duplicated implementation block across source files manim_animations/dataloaders/stage_6.py:17
LOW AIC003 Duplicated implementation block across source files manim_animations/dataloaders/stage_6.py:3
LOW AIC003 Duplicated implementation block across source files manim_animations/dataloaders/stage_5.py:17
LOW AIC003 Duplicated implementation block across source files manim_animations/dataloaders/stage_1.py:3
LOW AIC003 Duplicated implementation block across source files manim_animations/big_model_inference/st…:17
LOW AIC003 Duplicated implementation block across source files manim_animations/big_model_inference/st…:7
LOW AIC003 Duplicated implementation block across source files manim_animations/big_model_inference/st…:3
LOW AIC003 Duplicated implementation block across source files manim_animations/big_model_inference/st…:28
LOW AIC003 Duplicated implementation block across source files manim_animations/big_model_inference/st…:7
LOW AIC003 Duplicated implementation block across source files manim_animations/big_model_inference/st…:6
LOW AIC003 Duplicated implementation block across source files manim_animations/big_model_inference/st…:3
LOW DKR011 Dockerfile installs recommended OS packages docker/accelerate-gpu-deepspeed/Dockerf…:37
LOW DKR011 Dockerfile installs recommended OS packages docker/accelerate-gpu-deepspeed/Dockerf…:9
LOW DKR011 Dockerfile installs recommended OS packages docker/accelerate-gpu/Dockerfile:37
LOW DKR011 Dockerfile installs recommended OS packages docker/accelerate-gpu/Dockerfile:9
LOW DKR012 Dockerfile keeps pip download cache benchmarks/fp8/transformer_engine/Docke…:9
LOW DKR012 Dockerfile keeps pip download cache benchmarks/fp8/transformer_engine/Docke…:6
LOW DKR012 Dockerfile keeps pip download cache benchmarks/fp8/torchao/Dockerfile:6
LOW DKR012 Dockerfile keeps pip download cache benchmarks/fp8/torchao/Dockerfile:3
LOW DKR012 Dockerfile keeps pip download cache benchmarks/fp8/ms_amp/Dockerfile:6
LOW DKR012 Dockerfile keeps pip download cache benchmarks/fp8/ms_amp/Dockerfile:3
INFO MINED063 [MINED063] Toctou Os Path Exists: if os.path.exists(p): open(p) — file can be replaced/de… src/accelerate/utils/deepspeed.py:141
INFO MINED072 [MINED072] Python Pass Only Class: class Foo: pass — stub waiting to be filled in. src/accelerate/utils/operations.py:739
INFO MINED072 [MINED072] Python Pass Only Class: class Foo: pass — stub waiting to be filled in. src/accelerate/test_utils/testing.py:822
INFO MINED062 [MINED062] Python Dataclass No Fields: @dataclass over an empty class — unfinished model. src/accelerate/utils/environment.py:250
INFO MINED062 [MINED062] Python Dataclass No Fields: @dataclass over an empty class — unfinished model. src/accelerate/utils/dataclasses.py:113
INFO MINED062 [MINED062] Python Dataclass No Fields: @dataclass over an empty class — unfinished model. src/accelerate/parallelism_config.py:33
INFO MINED069 [MINED069] Debug True Prod: Django/Flask DEBUG=True or app.debug=True in non-test files. src/accelerate/launchers.py:308
INFO MINED079 [MINED079] Off By One Slice: range(len(x)+1), arr[i+1:i+n+1], or while i<=len(arr) — off-… src/accelerate/utils/modeling.py:690
INFO MINED079 [MINED079] Off By One Slice: range(len(x)+1), arr[i+1:i+n+1], or while i<=len(arr) — off-… src/accelerate/commands/menu/selection_…:141
INFO MINED064 [MINED064] Python Input Call: input() blocks for stdin. Inappropriate in services. src/accelerate/commands/menu/selection_…:134
INFO MINED064 [MINED064] Python Input Call: input() blocks for stdin. Inappropriate in services. src/accelerate/commands/config/config_u…:50
INFO MINED064 [MINED064] Python Input Call: input() blocks for stdin. Inappropriate in services. src/accelerate/accelerator.py:3074
INFO MINED077 [MINED077] Python Open No Context: fp = open(path) outside with-block leaks file handles. setup.py:61
INFO MINED055 [MINED055] Npm Install No Lockfile: Production image runs npm install (resolves new versi… examples/inference/distributed/florence…:36
INFO MINED055 [MINED055] Npm Install No Lockfile: Production image runs npm install (resolves new versi… examples/inference/distributed/distribu…:35
INFO MINED050 [MINED050] Stub Only Function: Function declared but body is just pass, return None, rais… examples/by_feature/gradient_accumulati…:163
INFO MINED050 [MINED050] Stub Only Function: Function declared but body is just pass, return None, rais… examples/by_feature/gradient_accumulati…:129
INFO MINED050 [MINED050] Stub Only Function: Function declared but body is just pass, return None, rais… examples/by_feature/fsdp_with_peak_mem_…:311
INFO MINED049 [MINED049] Print Pii: Logging password/token/email/ssn directly to stdout. src/accelerate/commands/config/sagemake…:108
INFO MINED049 [MINED049] Print Pii: Logging password/token/email/ssn directly to stdout. examples/inference/pippy/llama.py:56
INFO MINED049 [MINED049] Print Pii: Logging password/token/email/ssn directly to stdout. benchmarks/big_model_inference/big_mode…:129
Reset to top 5 92 findings available (after auto-suppression of test files + won't-fix)

Issue body (markdown)

## Code-quality scan: `huggingface/accelerate`

**Score: 70/100 (B-)**  ·  92 findings  ·  scanned 2026-05-18 14:47 UTC  ·  63,090 LOC

| Severity | Count |
|---|---|
| CRITICAL | 3 |
| HIGH | 20 |
| MEDIUM | 21 |
| LOW | 27 |

📊 [Full filterable report](https://repobility.com/scan/997bbbe3-2683-4e32-a73e-484d75ceba8d/)  ·  ![scorecard](https://repobility.com/scan/997bbbe3-2683-4e32-a73e-484d75ceba8d/report.png?v=1779115643-s2)

### Top findings

1. **CRITICAL** `MINED030` — Python Pickle Loads
   `src/accelerate/utils/operations.py:484` · CWE-502 · ✓ Repobility
2. **CRITICAL** `MINED018` — Unsafe Deserialization Pickle
   `src/accelerate/utils/operations.py:484` · CWE-502 · ✓ Repobility
3. **CRITICAL** `SEC081` — Python: pickle.loads / marshal.loads on untrusted data
   `src/accelerate/utils/operations.py:484` · A05:2021 Security Misconfiguration
4. **HIGH** `MINED001` — Bare Except Pass
   `src/accelerate/utils/environment.py:223` · CWE-755 · ✓ Repobility
5. **HIGH** `MINED001` — Bare Except Pass
   `src/accelerate/test_utils/testing.py:822` · CWE-755 · ✓ Repobility

---

**Security note**: this issue is public. If any flagged finding is a real, exploitable vulnerability, please redirect to your `SECURITY.md` policy or open a [private security advisory](https://docs.github.com/en/code-security/security-advisories/guidance-on-reporting-and-writing-information-about-vulnerabilities/privately-reporting-a-security-vulnerability) instead. We're happy to close this and re-submit privately.

---

_Filed automatically. Close this issue if not useful — we won't refile. Full report: https://repobility.com/scan/997bbbe3-2683-4e32-a73e-484d75ceba8d/_
Already filed
'huggingface' is on the known-megaproject org list. These projects use auto-triage bots and established security disclosure channels. Unsolicited automated issues from Repobility would be perceived as AI-generated spam. For security findings, follow the project's SECURITY.md policy. For non-security findings, open a focused PR or community discussion instead.
Megaproject â high spam risk
Could not determine 'huggingface/accelerate' star count (GitHub API rate-limited or unreachable). When in doubt about repo size, prefer opening a focused PR or a discussion rather than an issue.
Already filed
32/106 findings (30%) on this scan are already flagged as test-file, won't-fix, or suppressed. The scan is too noisy to file as a single issue. Curate down to specific actionable findings, or address the FP source first.
I read the warnings â override Cancel

The button opens GitHubâs new-issue page in a new tab. You will see the title + body pre-filled â review, edit if you want, then click GitHubâs "Submit new issue" button. Repobility never posts anything on your behalf.

For real security findings on big repos: use the project's SECURITY.md or private advisory flow instead of a public issue.