保留正文与补采附件。源页面当前未重新核验；具体限制见交付清单。

Revisit consent button We value your privacy We use cookies to enhance your browsing experience, serve personalised ads or content, and analyse our traffic. By clicking "Accept All", you consent to our use of cookies. Customise Reject All Accept All Customise Consent Preferences We use cookies to help you navigate efficiently and perform certain functions. You will find detailed information about all cookies under each consent category below. The cookies that are categorised as "Necessary" are stored on your browser as they are essential for enabling the basic functionalities of the site. ... Show more NecessaryAlways Active Necessary cookies are required to enable the basic features of this site, such as providing secure log-in or adjusting your consent preferences. These cookies do not store any personally identifiable data.

Functional Functional cookies help perform certain functionalities like sharing the content of the website on social media platforms, collecting feedback, and other third-party features.

Analytics Analytical cookies are used to understand how visitors interact with the website. These cookies help provide information on metrics such as the number of visitors, bounce rate, traffic source, etc.

Performance Performance cookies are used to understand and analyse the key performance indexes of the website which helps in delivering a better user experience for the visitors.

Advertisement Advertisement cookies are used to provide visitors with customised advertisements based on the pages you visited previously and to analyse the effectiveness of the ad campaigns.

Uncategorised Other uncategorised cookies are those that are being analysed and have not been classified into a category as yet. No cookies to display. Reject All Save My Preferences Accept All Toggle SearchToggle Mobile Navigation Publish AI, ML & data-science insights to a global community of data professionals. Sign inSubmit an Article LatestEditor's PicksNewsletterShipAI

Write For TDS LinkedInXToggle Search Machine Learning

How to Fine-Tune an SLM for Emotion Recognition

Python tutorial for fine-tuning a Mistral Small 3.1 on an imbalanced training set to classify 15 emotions in social media communication Petr Koráb June 5, 20269 min read Image created by GPT 5.3

Introduction

Recent small language models (SLMs) fine-tuned for sentiment classification infer sentiment as a single score, capturing the overall emotional tone of the text. For many use cases, the positive-negative classification does not tell the full story a company needs. Emotion recognition models go further, decomposing sentiment into emotion classes (“anger”, “approval”, “disappointment” , etc.) and assigning probabilities to a set of emotions in the text. It is then possible to model emotional content in datasets that a company receives (customer tickets, emails, brand-related discussions), and react swiftly to changing conditions. For one of our recent projects, modeling emotions in online media, we required an emotion recognition model with open weights and a flexible license, maintaining high transparency standards, and, of course, benefiting from the lower costs associated with open models. We subjectively prefer European models, but Hugging Face did not offer a Mistral alternative with a developed model card. One possible reason is that the most detailed training set for emotion recognition, the 28-emotion GoEmotions dataset, is highly class-imbalanced. Fine-tuning an SLM on a high-class-imbalance data set that performs decently on the test requires a deeper focus. We treated the class-imbalance problem by a combination of three techniques: (1) undersampling the most represented emotional category, (2) synthetically expanding the minority classes with Nature’s 2025 ISMOTE algorithm, and (3) weighting the loss function. With this combination of techniques, MistralSmall-3.1.GoEmotions, now released on Hugging Face, infers most target emotions relevant to our project with F1 > 0.7. This article explains in detail how to fine-tune an open-weight SLM. We’ll also figure out:

Learn this step by step with the interactive Machine Learning roadmap.

2. Data

GoEmotions is a human-annotated dataset of 58k Reddit comments extracted from English-language subreddits and labeled with 27 emotion categories and a “neutral” label. It is a multi-label classification dataset in which each comment may be labeled with multiple TRUEs for emotions (e.g., “Hitting me. That just added another funny dynamic to it even though I wasn’t actually trying to hit her” is True for “amusement” , and “annoyance”). The dataset was released on TensorFlow Datasets under the Apache 2.0 License and contains 54,263 labeled texts. Here is what it looks like: Image 1. GoEmotions dataset. Image by author. After a quick check, we can see a high-class imbalance in the data where the neutral category prevails: Image 2. Class imbalance in GoEmotions dataset. Image by author.

3. Training set preprocessing

Our goal is to develop a classifier to identify 15 emotions in general-language texts. Training on class-imbalanced data can introduce bias, as the fine-tuned model tends to favor the majority class and perform worse on the minority ones, so preprocessing is essential. We used a combination of methods for the training set; the validation and test sets remained unchanged to address class imbalance and maximize performance on the target emotions (fear, sadness, disgust, disapproval, annoyance, anger, disappointment, optimism, amusement, surprise, admiration, excitement, confusion, joy, love):

The ISMOTE algorithm extends the common SMOTE technique by (1) expanding the sample generation space and (2) improving sampling distribution. The synthetically generated samples then have more realistic data distributions than those produced by the original method. Image 3. The flowchart of the ISMOTE algorithm. Source: Scientific Reports. By reducing the majority class and synthetically expanding the minority categories to 4000 samples, we constructed a relatively balanced set for fine-tuning. The code for ISMOTE oversampling is here. Image 4. Label relative frequency , train (augmented), validation, and test sets. Image by author.

4. SLM Fine-tuning

Among Mistral’s models, we chose the Small class (Small-3.1-24B-Instruct-2503), which fits our GPU and provides the multilingual capabilities we need for the classifier. The Unsloth framework makes the finetuning steps uncomplicated and faster than with Transformers: 1. Data loading — loading preprocessed training set, validation, and test sets. We use a 60:20:20 split. 2. Loading the base model — loading the Small-3.1–24B-Instruct-2503 locally. _3._Apply LoRA — lowering hardware requirements . 4. Multilabel wrapper with focal loss function — updates the trainer for multilabel classification. Also adds focal loss to weight the loss function for a selected set of emotions, prioritizing their performance. 5. Evaluation metrics and training args — specifying the evaluation metrics and hyperparameters for model training. _6._Model training — trainer formulation and launch. 7. Evaluation — evaluating the best model performance on the test set.

4.1. Coding

Here is the code implementation.

4.1.1. Data loading

python

# Loading augmented train, validation and test sets
BASE = r"augmented"

def load_split(path: str) -> Dataset:
    with open(path, encoding="utf-8") as f:
        d = json.load(f)
    return Dataset.from_dict({"input_embeds": d["X"], "labels": d["y"]})

train_dataset = load_split(f"{BASE}/train.json")
val_dataset   = load_split(f"{BASE}/val.json")
test_dataset  = load_split(f"{BASE}/test.json")

# Formulate embedding dimension
EMBED_DIM = len(train_dataset[0]["input_embeds"])

# Return Pytorch tensors
train_dataset.set_format("torch")
val_dataset.set_format("torch")
test_dataset.set_format("torch")

4.1.2. Loading the base model

python

# Load base model with Unsloth FastLanguageModel
MODEL_NAME = "unsloth/Mistral-Small-3.1-24B-Instruct-2503"

base_model, _ = FastLanguageModel.from_pretrained(
    model_name=MODEL_NAME,
    max_seq_length=2,
    load_in_4bit=True,
    dtype=torch.bfloat16,
)

4.1.3. Apply LoRA

python

# Aply Low-rank adaptation (LoRA) 
base_model = FastLanguageModel.get_peft_model(
    base_model,
    r=16,
    lora_alpha=32,
    lora_dropout=0,
    bias="none",
    target_modules=[
        "q_proj", "k_proj", "v_proj", "o_proj",
        "gate_proj", "up_proj", "down_proj",
    ],
    use_gradient_checkpointing="unsloth",
    random_state = 3407,
    use_rslora = False, 
)

4.1.4. Multilabel wrapper with focal loss function

python

# Focal loss weights for preffered labels  
FOCAL_ALPHA_DEFAULT   = 0.25
FOCAL_ALPHA_PREFERRED = 0.75

PREFERRED_LABELS = {
    "fear", "sadness", "disgust", "disapproval", "annoyance",
    "anger", "disappointment", "optimism", "amusement", "surprise",
    "admiration", "excitement", "confusion","joy","love"
}

FOCAL_ALPHA_PER_LABEL: list[float] = [
    FOCAL_ALPHA_PREFERRED if lbl in PREFERRED_LABELS else FOCAL_ALPHA_DEFAULT
    for lbl in EMOTION_LABELS
]

"Per-label weighted focal binary cross-entropy for multi-label problems"
class FocalLossWithAlpha(nn.Module):
        def __init__(self, alpha: list[float], gamma: float = 2.0):
        super().__init__()
        self.register_buffer("alpha", torch.tensor(alpha, dtype=torch.float32))
        self.gamma = gamma
    def forward(self, logits: torch.Tensor, targets: torch.Tensor) -> torch.Tensor:
        probs   = torch.sigmoid(logits)
        p_t     = probs * targets + (1.0 - probs) * (1.0 - targets)
        alpha_t = self.alpha * targets + (1.0 - self.alpha) * (1.0 - targets)
        focal_w = alpha_t * (1.0 - p_t) ** self.gamma
        bce     = nn.functional.binary_cross_entropy_with_logits(
            logits, targets, reduction="none"
        )
        return (focal_w * bce).mean()

python

# Multilabel classification wrapper with focal loss class weighting
class MistralForMultiLabel(nn.Module):
    is_loaded_in_4bit = True

    def __init__(self, backbone: nn.Module, num_labels: int,
                 hidden_size: int, embed_dim: int):
        super().__init__()
        self.backbone = backbone
        _device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
        self.projection = nn.Sequential(
            nn.Linear(embed_dim, hidden_size // 2),
            nn.GELU(),
            nn.Linear(hidden_size // 2, hidden_size),
        ).to(_device)
        self.dropout    = nn.Dropout(0.1).to(_device)
        self.classifier = nn.Linear(hidden_size, num_labels).to(_device)
        self.focal_loss = FocalLossWithAlpha(FOCAL_ALPHA_PER_LABEL).to(_device)

    def gradient_checkpointing_enable(self, gradient_checkpointing_kwargs=None):
        self.backbone.gradient_checkpointing_enable(gradient_checkpointing_kwargs)

    def gradient_checkpointing_disable(self):
        self.backbone.gradient_checkpointing_disable()

    def forward(
        self,
        input_embeds: torch.Tensor,
        labels: torch.Tensor | None = None,
        **kwargs,
    ):
        B = input_embeds.size(0)
        projected = self.projection(input_embeds).unsqueeze(1)
        attn_mask = torch.ones(B, 1, device=input_embeds.device)

        outputs = self.backbone.base_model.model.model(
            inputs_embeds=projected,
            attention_mask=attn_mask,
            output_hidden_states=True,
        )
        pooled = outputs.hidden_states[-1][:, 0, :]
        logits = self.classifier(self.dropout(pooled))

        loss = self.focal_loss(logits, labels.float()) if labels is not None else None
        return {"loss": loss, "logits": logits}

4.1.5. Evaluation metrics and training args

python

# Specifiy the evaluation function
def compute_metrics(eval_pred):
    logits, labels = eval_pred
    probs = torch.sigmoid(torch.tensor(logits)).numpy()
    preds = (probs >= 0.5).astype(int)
    labels = labels.astype(int)

    from sklearn.metrics import accuracy_score

    exact_accuracy  = accuracy_score(labels, preds)
    macro_f1        = f1_score(labels, preds, average="macro", zero_division=0)
    micro_f1        = f1_score(labels, preds, average="micro", zero_division=0)
    macro_precision = precision_score(labels, preds, average="macro", zero_division=0)
    macro_recall    = recall_score(labels, preds, average="macro", zero_division=0)

    per_class_f1        = f1_score(labels, preds, average=None, zero_division=0)
    per_class_recall    = recall_score(labels, preds, average=None, zero_division=0)
    per_class_precision = precision_score(labels, preds, average=None, zero_division=0)
    per_class_accuracy  = (preds == labels).mean(axis=0)

    per_class_metrics = {}
    for i, emotion in enumerate(EMOTION_LABELS):
        per_class_metrics[f"f1_{emotion}"]        = float(per_class_f1[i])
        per_class_metrics[f"recall_{emotion}"]    = float(per_class_recall[i])
        per_class_metrics[f"precision_{emotion}"] = float(per_class_precision[i])
        per_class_metrics[f"accuracy_{emotion}"]  = float(per_class_accuracy[i])

    return {
        "exact_accuracy":   exact_accuracy,
        "macro_f1":         macro_f1,
        "micro_f1":         micro_f1,
        "macro_precision":  macro_precision,
        "macro_recall":     macro_recall,
        **per_class_metrics,
    }

python

# Specify hyperparameters
training_args = TrainingArguments(
    output_dir=OUTPUT_DIR,            # where checkpoints and logs are written
    eval_strategy="epoch",            # run evaluation once per epoch
    save_strategy="epoch",            # save checkpoint once per epoch
    per_device_train_batch_size=8,    # samples per GPU per step
    per_device_eval_batch_size=16,    # larger batch is fine — no gradients
    gradient_accumulation_steps=4,    # effective batch = 8 × 4 = 32
    num_train_epochs=15,              # total passes over the training data
    learning_rate=1e-4,               # peak LR after warmup
    bf16=True,                        # bfloat16 mixed precision
    optim="adamw_8bit",               # 8-bit AdamW
    warmup_ratio=0.05,                # first 5 % of steps ramp LR from 0 to peak
    lr_scheduler_type="cosine",       # cosine decay from peak LR to ~0
    logging_steps=25,                 # print loss/LR to console every 25 steps
    logging_first_step=True,          # also log step 1 to catch early instability
    load_best_model_at_end=True,      # restore best checkpoint after training ends
    metric_for_best_model="macro_f1", # criterion used to select the best checkpoint
    greater_is_better=True,           # higher macro_f1 is better in evaluation
    gradient_checkpointing=False,    
    remove_unused_columns=False,      # keep input_embeds column
    save_total_limit=15,              # keep all checkpoints on disk to load the best model
    weight_decay=0.01,                # L2 regularisation on all trainable parameters
)

4.1.6. Model training

python

# Set-up the trainer for multilabel finetuning
class MultiLabelTrainer(Trainer):
    def compute_loss(self, model, inputs, return_outputs=False, **kwargs):
        labels = inputs.pop("labels")
        outputs = model(**inputs, labels=labels)
        loss = outputs["loss"]
        return (loss, outputs) if return_outputs else loss

    def _save_checkpoint(self, model, trial, metrics=None):
        super()._save_checkpoint(model, trial)
        ckpt_dir = self._get_output_dir(trial)
        # Save head
        torch.save({
            "projection": model.projection.state_dict(),
            "classifier":  model.classifier.state_dict(),
        }, os.path.join(ckpt_dir, "head_weights.pt"))
        # Save LoRA adapter explicitly (bypasses bitsandbytes serialization issues)
        model.backbone.save_pretrained(os.path.join(ckpt_dir, "lora_adapter"))

    def _load_best_model(self):
        best_ckpt = self.state.best_model_checkpoint
        if not best_ckpt:
            return
        # Restore head
        head_path = os.path.join(best_ckpt, "head_weights.pt")
        if os.path.exists(head_path):
            head = torch.load(head_path, map_location="cpu")
            self.model.projection.load_state_dict(head["projection"])
            self.model.classifier.load_state_dict(head["classifier"])
            print(f"Head restored from: {best_ckpt}")
        else:
            print(f"WARNING: head_weights.pt not found in {best_ckpt}")
        # Restore LoRA adapter
        lora_path = os.path.join(best_ckpt, "lora_adapter")
        if os.path.exists(lora_path):
            from peft import PeftModel
            self.model.backbone.load_adapter(lora_path, adapter_name="default")
            print(f"LoRA restored from: {best_ckpt}")
        else:
            print(f"WARNING: lora_adapter/ not found in {best_ckpt}")

# Launch the trainer
trainer = MultiLabelTrainer(
    model=model,
    args=training_args,
    train_dataset=train_dataset,
    eval_dataset=val_dataset,
    compute_metrics=compute_metrics,
)

# Launch training
trainer.train()

Fine-tuning for 15 epochs took 9 hours and 30 minutes on a machine with an NVIDIA RTX 6000 GPU and 192 GB of VRAM, with the best model loaded at the end.

4.1.7. Model evaluation

Let's show the performance on the test dataset. The standard statistics for model evaluation per class are F1 , Precision , and Recall. We can see relatively good performance on the target emotions, with F1 scores over 0.7, for most categories. Full performance is on the model card.
| Emotion | Precision | Recall | F1 | N |
| --- | --- | --- | --- | --- |
| admiration | 0.7415 | 0.6354 | 0.6844 | 993 |
| amusement | 0.7810 | 0.7422 | 0.7611 | 543 |
| anger | 0.7423 | 0.7367 | 0.7395 | 395 |
| annoyance | 0.7049 | 0.5452 | 0.6148 | 609 |
| confusion | 0.7576 | 0.8251 | 0.7899 | 303 |
| disappointment | 0.8487 | 0.8459 | 0.8473 | 305 |
| disapproval | 0.7208 | 0.5841 | 0.6453 | 517 |
| disgust | 0.8396 | 0.9368 | 0.8856 | 190 |
| excitement | 0.8240 | 0.9366 | 0.8767 | 205 |
| fear | 0.9112 | 0.9686 | 0.9390 | 159 |
| joy | 0.7577 | 0.8024 | 0.7794 | 339 |
| love | 0.7424 | 0.7903 | 0.7656 | 496 |
| optimism | 0.8145 | 0.7636 | 0.7882 | 368 |
| sadness | 0.8534 | 0.8899 | 0.8713 | 327 |
| surprise | 0.8456 | 0.8555 | 0.8505 | 256 |
| Macro precision | 0.8295 | | | |
| Macro recall | 0.8184 | | | |
| Micro F1 | 0.7527 | | | |
| Macro F1 | 0.8215 | | | |

5. Summary

Let’s now summarize the key points of the article. The requirements and full code are in this repo.

Petr Koráb is the founder of Text Mining Stories, a Prague-based development & consultancy company. Learn more about the cutting-edge NLP on our blog. AI statement. Some parts of the code were reviewed by Sonnet 4.6 (Cursor). No text was generated using AI. Acknowledgements . The National Bank of Slovakia Foundation supported this development. I thank Martin Feldkircher, Václav Jež, and Michala Moravcová for comments and suggestions.

References

[1] Ying Li, Yali Yang, Peihua Song, Lian Duan, Rui Ren. 2025. An improved SMOTE algorithm for enhanced imbalanced data classification by expanding sample generation space.Scientific Reports, 15(23521). [2] Yinhan Liu , Jiatao Gu , Naman Goyal, Xian Li, Sergey Edunov
Marjan Ghazvininejad, Mike Lewis, Luke Zettlemoyer. 2020. Multilingual Denoising Pre-training for Neural Machine Translation. Transactions of the Association for Computational Linguistics, 8, pp. 726–742.

Register Now to Continue Reading

Unlock Towards Data Science with a Free Account

All articles on Towards Data Science are available to registered users. Create a free account today to access this article and thousands more. Register for freeSign in to your account

Written By

Petr Koráb

See all from Petr Koráb Emotion RecognitionFine TuningLlmSentiment ClassificationSmall Language Model Share This Article

Towards Data Science is a community publication. Submit your insights to reach our global audience and earn through the TDS Author Payment Program. Write for TDS

Related Articles

Stop Choosing Between Local and Cloud LLMs: A Field Guide to Hybrid Patterns

LLM Applications A hands-on walkthrough of a hybrid local-cloud workflow using Gemma 4 and GPT-5.4, with reasoning and structured outputs Shuai Guo June 30, 202616 min read

Transformer-XL Review: Beyond Fixed-Length Contexts

Artificial Intelligence Review of "Transformer-XL: Attentive Language Models Beyond a Fixed-Length Context" Jiajin Li October 12, 20207 min read

Building a simple Agent with Tools and Toolkits in LangChain

Get familiar with the building blocks of Agents in LangChain Sami Maameri April 10, 202413 min read

Google's MusicLM: from text description to music

Artificial Intelligence A new model is generating impressive music from just text prompt Salvatore Raieli February 1, 20237 min read

State of Video Generation

Artificial Intelligence Overview of latest developments in the Video Generation area: From LLMs to MultyDiffusion. Paper's insights and Outstanding questions. Nikita Kiselov February 16, 202416 min read

Understanding ARIMA Models using PyCaret's Time Series Module - Part 1

Machine Learning 📚 Introduction Nikhil Gupta November 14, 202110 min read

How to Build Graph Transformers with O(N) Complexity

Machine Learning Tutorial on Large-Graph Transformers Qitian Wu April 19, 20238 min read

LLM Routing - Intuitively and Exhaustively Explained

Artificial Intelligence Dynamically Choosing the Right LLM Daniel Warfield November 22, 202459 min read

6 RAG question-parsing mistakes we keep seeing in production

Large Language Model Enterprise Document Intelligence [Vol.1 #6ter] - Six positions on the question-parsing brick that contradict the mainstream RAG playbook Angela and Kezhan Shi July 2, 202611 min read

Your home for data science and AI. The world's leading publication for data science, data analytics, data engineering, machine learning, and artificial intelligence professionals. Subscribe to Our NewsletterWrite For TDS•About•Advertise•Privacy Policy•Terms of Use © Insight Media Group, LLC 2026•The New Stack•roadmap.sh•Creator Network