SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2403.13372 · ACL · 2024

LLAMAFACTORY: Unified Efficient Fine-Tuning of 100+ Language Models

Junhao Zhang, Richong Zhang, Yaowei Zheng, Zheyan Luo, Yanhan Ye

arXiv · PDF · Open in the Atlas

Code that ran

11 of 17 samples (functions and classes) lifted from this paper's linked repositories executed on a synthesized input in Syntology's jail (0 of them constructed an object rather than computing a result; 0 honoured the contract we drafted, 0 violated it and 8 have no contract verdict; 3 failures of our instrument; 6 no recorded run), measured 2026-11-22; repository links from the Papers with Code archive attributed under CC BY-SA, no affiliation; bodies inline only under a permissive licence as GitHub detects it. Not a reproduction of the paper's results.

"Ran" means the function executed on a synthesized input and returned a value. It is not a reproduction of the paper's results. This record, dated, as JSON: /api/ran/2403.13372.json · badge: /api/ran/2403.13372.svg

Each repository's role is a label: a classifier's output recorded on the link, or, where no role is recorded, what the Papers with Code archive marked. The line under it names its basis. Neither says who wrote the code.

RepositoryRoleRan
Rcrossmeister/RLQG community (classifier label)
classifier basis not recorded on the link
8 ran · 5 no recorded run
BachOzean/TadE community (classifier label)
classifier basis not recorded on the link
3 ran · 0 no recorded run
smelliecat/aaemime community (classifier label)
classifier basis not recorded on the link
0 ran · 1 no recorded run
FunctionStatusWhere it lives
adaptive_ntk_forward Ran
ran on a synthesized input
Rcrossmeister/RLQG/model/src/api/attn_and_long_ctx_patches.py
code served (permissive licence) · get_code("bcfa7233bbbd6056")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned tuple, ndim 3, dtype family float; no output shape is stated. The image it ran in was not recorded. If your copy returns the same type and ndim on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
import torch.nn as nn

torch.manual_seed(0)

class MockRotaryEmbedding(nn.Module):
    def __init__(self):
        super().__init__()
        self.dim = 4
        self.base = 10000.0
        self.max_position_embeddings = 2048
        self.max_seq_len_cached = 512
        self.alpha = 'auto'
        self.scaling_factor = 1.0
        
        # Pre-compute cache for fallback path
        seq_len = 512
        t = torch.arange(seq_len, dtype=torch.float32)
        t = t / self.scaling_factor
        dim = self.dim
        alpha = (seq_len / (self.max_position_embeddings / 2) - 1) * 1.0
        base = self.base * alpha ** (dim / (dim - 2))
        ntk_inv_freq = 1.0 / base ** (torch.arange(0, dim, 2).float() / dim)
        freqs = torch.einsum('i,j->ij', t, ntk_inv_freq)
        emb = torch.cat((freqs, freqs), dim=-1)
        self.register_buffer('cos_cached', emb.cos())
        self.register_buffer('sin_cached', emb.sin())

self_obj = MockRotaryEmbedding()
x = torch.randn(2, 4, dtype=torch.float32)
seq_len = 1024

ARGS = {
    'self': self_obj,
    'x': x,
    'seq_len': seq_len
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type tuple, ndim 3, dtype family float — a tuple output has no array shape of its own; none is stated
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
apply_rotary_pos_emb_single Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
Rcrossmeister/RLQG/model/src/api/attn_and_long_ctx_patches.py
pointer only (licence: MIT); the stored copy get_code serves, from wuyike2000/retrieve-rewrite-answer, is pointer only (licence: NONE) · get_code("5067a5823022c273")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (MIT), but get_code serves the stored copy, first harvested from wuyike2000/retrieve-rewrite-answer, and GitHub detected no licence in that repository when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

calculate_similarity Ran
ran on a synthesized input
Rcrossmeister/RLQG/model/collector.py
code served (permissive licence) · get_code("e03cb2df2255bfe4")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned float, ndim 0, dtype family float, shape () as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
import numpy as np

class MockNLP:
    """Mock spaCy-like NLP model that supports similarity calculation."""
    
    def __call__(self, text):
        """Process text and return a mock doc object."""
        return MockDoc(text)

class MockDoc:
    """Mock spaCy Doc object with similarity method."""
    
    def __init__(self, text):
        self.text = text
        torch.manual_seed(0)
        self.vector = torch.randn(4)
    
    def similarity(self, other):
        """Calculate cosine similarity between two doc vectors."""
        dot_product = torch.dot(self.vector, other.vector)
        norm_self = torch.norm(self.vector)
        norm_other = torch.norm(other.vector)
        return float(dot_product / (norm_self * norm_other + 1e-8))

torch.manual_seed(0)
nlp = MockNLP()

ARGS = {
    "text1": "This is the first text.",
    "text2": "This is the second text.",
    "nlp": nlp
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type float, ndim 0, dtype family float, shape () — shape () as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
create_score_evaluation_response Ran
ran on a synthesized input
BachOzean/TadE/src/llamafactory/api/chat.py
code served (permissive licence) · get_code("f31c69421a40b0a4")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned coroutine. The image it ran in was not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
import uuid
from typing import List, Optional

torch.manual_seed(0)

# Stand-in classes for the types used in the function
class Message:
    def __init__(self, role: str, content: str):
        self.role = role
        self.content = content

class ScoreEvaluationRequest:
    def __init__(self, messages: List[Message], max_length: int, model: str):
        self.messages = messages
        self.max_length = max_length
        self.model = model

class ScoreEvaluationResponse:
    def __init__(self, id: str, model: str, scores: List[float]):
        self.id = id
        self.model = model
        self.scores = scores

class ChatModel:
    async def aget_scores(self, messages: List[Message], max_length: int) -> List[float]:
        # Return dummy scores based on number of messages
        return [0.5 + 0.1 * i for i in range(len(messages))]

# Construct request with valid messages (non-empty list)
messages = [
    Message(role="user", content="Hello"),
    Message(role="assistant", content="Hi there"),
]

request = ScoreEvaluationRequest(
    messages=messages,
    max_length=512,
    model="gpt-4"
)

# Construct chat model
chat_model = ChatModel()

# Define ARGS dictionary
ARGS = {
    "request": request,
    "chat_model": chat_model,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type coroutine — a coroutine output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a coroutine output the recorded ndim 0 and dtype family float describe that coercion, not the output, and are not stated
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
dictify Ran
ran; our contract draft was wrong, not the code
BachOzean/TadE/src/llamafactory/api/common.py
code served (permissive licence) · get_code("87aa9acaa81294f5")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned dict. The image it ran in was not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
from pydantic import BaseModel

class SampleModel(BaseModel):
    name: str
    value: int
    data: list

# Create a valid Pydantic BaseModel instance
sample_data = SampleModel(name="test", value=42, data=[1, 2, 3])

ARGS = {
    "data": sample_data
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
pydantic, torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type dict — a dict output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a dict output the recorded ndim 0 and dtype family float describe that coercion, not the output, and are not stated
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
exact_match_accuracy Ran
ran on a synthesized input
Rcrossmeister/RLQG/evaluation/utils.py
code served (permissive licence) · get_code("f561697841f2d428")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned float, ndim 0, dtype family float, shape () as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch

torch.manual_seed(0)

# Construct test data with 'response' and 'answer' keys
data = [
    {'response': 'Paris', 'answer': 'paris'},
    {'response': 'London', 'answer': 'LONDON'},
    {'response': 'Berlin', 'answer': 'Madrid'},
]

ARGS = {
    'data': data
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type float, ndim 0, dtype family float, shape () — shape () as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
jsonify Ran
ran; our contract draft was wrong, not the code
BachOzean/TadE/src/llamafactory/api/common.py
code served (permissive licence) · get_code("57583cf880d95444")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned str. The image it ran in was not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import json
from pydantic import BaseModel

class SampleModel(BaseModel):
    name: str
    value: int
    description: str = "default"

# Create an instance with some fields set
data_instance = SampleModel(name="test", value=42)

ARGS = {
    "data": data_instance
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
pydantic — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type str — a str output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a str output the recorded ndim 0 and dtype family float describe that coercion, not the output, and are not stated
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
load_sentences Ran
ran on a synthesized input
Rcrossmeister/RLQG/dataset/utils/loader.py
code served (permissive licence) · get_code("4ea843747b80c800")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned list, ndim 1, dtype family float; no output shape is stated. The image it ran in was not recorded. If your copy returns the same type and ndim on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import json
import tempfile
import os

# Create a temporary file with sample data
temp_dir = tempfile.mkdtemp()
test_file = os.path.join(temp_dir, "test_ace.jsonl")

# Write sample JSON lines to the file
sample_data = [
    {
        "sentence": "The company announced a major acquisition",
        "s_start": 0,
        "event": [[[4, 10], "ACQUISITION"]],
        "tokens": ["The", "company", "announced", "a", "major", "acquisition"],
        "trigger_label": [0, 0, 0, 0, 0, 1],
        "char_offset": [[0, 3], [4, 11], [12, 20], [21, 22], [23, 28], [29, 40]]
    },
    {
        "sentence": "The CEO resigned from the board",
        "s_start": 0,
        "event": [[[4, 7], "PERSONNEL"]],
        "tokens": ["The", "CEO", "resigned", "from", "the", "board"],
        "trigger_label": [0, 0, 1, 0, 0, 0],
        "char_offset": [[0, 3], [4, 7], [8, 15], [16, 20], [21, 24], [25, 30]]
    }
]

with open(test_file, 'w') as f:
    for item in sample_data:
        f.write(json.dumps(item) + '\n')

ARGS = {
    "path": test_file
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
none outside the standard library — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type list, ndim 1, dtype family float — a list output has no array shape of its own; none is stated
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
read_json Ran
ran on a synthesized input
Rcrossmeister/RLQG/model/converter.py
code served (permissive licence) · get_code("147f30501936b7f3")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned dict. The image it ran in was not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import json
import tempfile
import os

# Create a temporary JSON file with valid content
temp_dir = tempfile.mkdtemp()
temp_file = os.path.join(temp_dir, "test.json")

test_data = {
    "key1": "value1",
    "key2": [1, 2, 3],
    "key3": {"nested": "object"}
}

with open(temp_file, 'w') as f:
    json.dump(test_data, f)

ARGS = {
    "file_path": temp_file
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
none outside the standard library — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type dict — a dict output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a dict output the recorded ndim 0 and dtype family float describe that coercion, not the output, and are not stated
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
read_jsonl Ran
ran on a synthesized input
Rcrossmeister/RLQG/model/converter.py
code served (permissive licence) · get_code("2a7c632d17841b02")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned list, ndim 1, dtype family float; no output shape is stated. The image it ran in was not recorded. If your copy returns the same type and ndim on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import tempfile
import json
import os

# Create a temporary JSONL file with valid JSON lines
temp_dir = tempfile.mkdtemp()
temp_file = os.path.join(temp_dir, "test.jsonl")

with open(temp_file, 'w') as f:
    f.write(json.dumps({"key1": "value1", "number": 42}) + "\n")
    f.write(json.dumps({"key2": "value2", "list": [1, 2, 3]}) + "\n")
    f.write(json.dumps({"key3": "value3", "nested": {"inner": "data"}}) + "\n")

ARGS = {
    "file_path": temp_file
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
none outside the standard library — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type list, ndim 1, dtype family float — a list output has no array shape of its own; none is stated
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
rmANS Ran
ran on a synthesized input
Rcrossmeister/RLQG/evaluation/llama2_qa.py
code served (permissive licence) · get_code("05cd11daf8842801")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned str. The image it ran in was not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch

torch.manual_seed(0)

def rmANS(response):
    if '[ANS]' in response:
        context = response.split('[ANS] ')[1].split(' [/ANS]')[0]
        if '[ANS]' in context:
            context = response.split(' [ANS]')[0]
            return context
        else:
            return context
    else:
        return response

ARGS = {
    'response': 'This is a test [ANS] the answer here [/ANS] and more text'
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type str — a str output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a str output the recorded ndim 0 and dtype family float describe that coercion, not the output, and are not stated
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
content_overlap_ratio Not yet run
unverified (harvested, no recorded run)
Rcrossmeister/RLQG/evaluation/utils.py
code served (permissive licence) · get_code("b508d81907f540d0")
How it ran

no invocation recorded.

convert_examples_to_features_sequence Not yet run
unverified (harvested, no recorded run)
Rcrossmeister/RLQG/dataset/utils/loader.py
code served (permissive licence) · get_code("6f15166ed679f5d3")
How it ran

no invocation recorded.

load_chat Not yet run
unverified (harvested, no recorded run)
smelliecat/aaemime/new/src/gen_for_coraal.py
code served (permissive licence) · get_code("dd1bd4b20b76b8bd")
How it ran

no invocation recorded.

load_template Not yet run
unverified (harvested, no recorded run)
Rcrossmeister/RLQG/dataset/utils/loader.py
code served (permissive licence) · get_code("0c5d5a49e28081e7")
How it ran

no invocation recorded.

process_sentence Not yet run
unverified (harvested, no recorded run)
Rcrossmeister/RLQG/dataset/RAMS/rams_load.py
code served (permissive licence) · get_code("e9a8c69747eb64a7")
How it ran

no invocation recorded.

xformers_forward Not yet run
unverified (harvested, no recorded run)
Rcrossmeister/RLQG/model/src/api/attn_and_long_ctx_patches.py
code served (permissive licence) · get_code("7385dc9c056d4dc9")
How it ran

no invocation recorded.

Also present in other papers' code

Of 17 samples lifted from this paper's linked repositories, 17 carry a function identity and 0 do not. Of the 17 distinct functions with an identity, 4 are also present in other papers whose code we hold (1 with at least one paper dated earlier, 0 the same month, 4 later) and 13 are only here among the 52,977 papers whose code Syntology holds (as of 2026-09-26).

Only here means: among the 52,977 papers whose code Syntology holds (51,887 with lifted samples, 1,090 through their assemblies only), as of 2026-09-26, no other paper's lifted samples or current assembly bricks contain an identical function body after normalisation.

FunctionAlso present inEarliest-datedOthers named
jsonify 28 other papers
28 later
2405.04760 (2024-05)
Large Language Models for Cyber Security: A Systematic Literature Review
2406.04496 (2024-06)
Time Sensitive Knowledge Editing through Efficient Finetuning
2407.08044 (2024-07)
RoLoRA: Fine-tuning Rotated Outlier-free LLMs for Effective Weight-Activation Quantization
and 25 more not named here
dictify 14 other papers
14 later
2407.08044 (2024-07)
RoLoRA: Fine-tuning Rotated Outlier-free LLMs for Effective Weight-Activation Quantization
2409.13183 (2024-09)
$\textit{SKIntern}$: Internalizing Symbolic Knowledge for Distilling Better CoT Capabilities into Small Language Models
2409.13203 (2024-09)
Neural-Symbolic Collaborative Distillation: Advancing Small Language Models for Complex Reasoning Tasks
and 11 more not named here
apply_rotary_pos_emb_single 2 other papers
1 dated earlier, 1 later
2309.11206 (2023-09)
Retrieve-Rewrite-Answer: A KG-to-Text Enhanced LLMs Framework for Knowledge Graph Question Answering
2407.03157 (2024-07)
Let the Code LLM Edit Itself When You Edit the Code
create_score_evaluation_response 2 other papers
2 later
2410.10209 (2024-10)
EFFICODER: Enhancing Code Generation in Large Language Models through Efficiency-Aware Fine-tuning
2505.16869 (2025-05)
MPO: Multilingual Safety Alignment via Reward Gap Optimization

Presence, not direction: an identical function body in two papers' repositories does not say who wrote it, which came from which, or that either paper's method rests on it. Matching is exact after normalisation, so a renamed or edited function is not matched. Dates are when papers appeared, not when code was written. Only .py files the harvester lifted are covered; notebooks are not read. A new-style arXiv id YYMM.NNNN dates a paper to the month 20YY-MM; otherwise Paper.year dates it to the year; otherwise it is undated. Dates are when the paper appeared, not when the code was written. Read 2026-11-22.

Repositories linked to this paper

Some links come from the archived Papers with Code dataset (CC BY-SA 4.0): attribution and licence.

Abstract

Efficient fine-tuning is vital for adapting large language models (LLMs) to downstream tasks. However, it requires non-trivial efforts to implement these methods on different models. We present LLAMAFACTORY, a unified framework that integrates a suite of cutting-edge efficient training methods. It provides a solution for flexibly customizing the fine-tuning of 100+ LLMs without the need for coding through the built-in web UI LLAMABOARD. We empirically validate the efficiency and effectiveness of our framework on language modeling and text generation tasks. It has been released at https://github.com/hiyouga/LLaMA-Factory and received over 25,000 stars and 3,000 forks.

For agents

The same record, over MCP at https://syntology.ai/mcp:

get_harvested_code_for_paper("2403.13372")
get_code_for_paper("2403.13372")
have("2403.13372")

The run record, dated, one paper per request, free:

curl https://syntology.ai/api/ran/2403.13372.json

A badge for a README (the split and the date, never a ratio):

[![Syntology run record](https://syntology.ai/api/ran/2403.13372.svg)](https://syntology.ai/paper/2403.13372)

Connect an agent — have() is free.