Add LoRA adapter, model card, and eval artifacts

.gitattributes

@@ -33,3 +33,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.zip filter=lfs diff=lfs merge=lfs -text
 *.zst filter=lfs diff=lfs merge=lfs -text
 *tfevents* filter=lfs diff=lfs merge=lfs -text
+tokenizer.json filter=lfs diff=lfs merge=lfs -text

LICENSE

@@ -0,0 +1,33 @@
+These LoRA adapter weights are released under the MIT License, matching the
+license declared by the base model, zai-org/GLM-4.7-Flash. Note that the base
+model's repository declares `mit` in its metadata but does not itself ship a
+LICENSE file; the standard MIT text is reproduced below and applies to this
+adapter.
+
+The training data is derived from b-mc2/sql-create-context, which is licensed
+CC-BY-4.0. That license is not superseded by this one — downstream use of these
+weights should carry its attribution requirement.
+
+---
+
+MIT License
+
+Copyright (c) 2026 SASVA AI Model Cognition Labs (MCL)
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.

README.md

@@ -1,3 +1,412 @@
 ---
-license: apache-2.0
+# ---- Identity -------------------------------------------------------------
+base_model: zai-org/GLM-4.7-Flash
+base_model_relation: adapter
+library_name: peft
+pipeline_tag: text-generation
+language:
+  - en
+license: mit   # verified via the Hub API: zai-org/GLM-4.7-Flash reports `mit`.
+               # The training data is CC-BY-4.0 — see Licence below.
+
+# ---- Discovery ------------------------------------------------------------
+tags:
+  - lora
+  - qlora
+  - peft
+  - sft
+  - trl
+  - text-to-sql
+  - sql
+
+datasets:
+  - b-mc2/sql-create-context
+
+metrics:
+  - exact_match
+  - bleu
+  - rouge
+
+# ---- Structured evaluation ------------------------------------------------
+model-index:
+  - name: glm-4.7-flash-sql-create-context-lora
+    results:
+      - task:
+          type: text-generation
+          name: Text-to-SQL (natural language + CREATE TABLE schema -> SQL)
+        dataset:
+          type: sql-create-context-val
+          name: sql-create-context derived validation split (400 pairs)
+          split: validation
+        metrics:
+          - type: exact_match
+            name: Exact match
+            value: 0.805
+          - type: bleu
+            name: BLEU
+            value: 0.940082
+          - type: rouge
+            name: ROUGE-L
+            value: 0.986055
+            args:
+              rouge_type: rougeL
 ---
+
+# GLM-4.7-Flash Text-to-SQL (LoRA)
+
+Given a natural-language question and a `CREATE TABLE` schema, emits exactly one
+SQL query answering that question against that schema. For natural-language
+query interfaces over a known relational schema.
+
+This is a **LoRA adapter for**
+[zai-org/GLM-4.7-Flash](https://huggingface.co/zai-org/GLM-4.7-Flash), trained
+with **QLoRA (4-bit NF4 base, bf16 compute)** via
+[TRL](https://github.com/huggingface/trl) SFT.
+
+## Model details
+
+| | |
+|---|---|
+| Developed by | SASVA AI Model Cognition Labs (MCL) Team |
+| Base model | [`zai-org/GLM-4.7-Flash`](https://huggingface.co/zai-org/GLM-4.7-Flash) |
+| Base parameters | 30B total / 3B active (MoE) — 29,943,396,864 in the merged bf16 build |
+| Architecture family | `glm4_moe_lite` (`Glm4MoeLiteForCausalLM`), 47 layers, hidden size 2048, vocab 154,880 |
+| Adaptation | LoRA (`r=64`, `alpha=128`, `dropout=0.05`) |
+| Trainable modules | `q_a_proj`, `q_b_proj`, `kv_a_proj_with_mqa`, `kv_b_proj`, `o_proj`, `gate_proj`, `up_proj`, `down_proj` |
+| Training method | `qlora` (4-bit NF4, double quant, bf16 compute) |
+| Refinement | none |
+| Language | English (questions) / SQL (outputs) |
+| License | MIT (inherited from the base model) |
+
+Trainable parameters: **118,140,928** — **0.3945%** of the base. The adapter file
+is 472,671,000 bytes (752 fp32 tensors: a `lora_A` + `lora_B` pair for each of
+the 8 target modules across all 47 layers).
+
+GLM-4.7-Flash uses Multi-head Latent Attention, so the attention target modules
+are the MLA projections (`q_a_proj`/`q_b_proj`/`kv_a_proj_with_mqa`/`kv_b_proj`),
+not `q_proj`/`k_proj`/`v_proj`. Targeting the conventional names would silently
+adapt nothing.
+
+## Intended use
+
+**Direct use.** Translate one English question plus one `CREATE TABLE` schema
+into one SQL query. The model was trained on a specific prompt shape and that
+shape is part of the contract:
+
+- System prompt (verbatim): *"You are a text-to-SQL engine. Given a
+  natural-language question and a CREATE TABLE schema, output exactly one SQL
+  query that answers the question against the provided schema. Output only the
+  raw SQL query on a single line with no explanation, no markdown formatting,
+  and no additional text."*
+- User turn: the question, a blank line, then the schema inside a fenced code
+  block.
+- Applied through the tokenizer's chat template (`chat_template.jinja`, shipped
+  in this repo) with `enable_thinking=False`. Do not concatenate strings by hand.
+- The query is the **first line** of the generation; discard anything after it.
+
+**Out of scope.**
+- **Not validated against a live database.** The model is scored on string
+  similarity to a reference query, never on execution. A syntactically perfect
+  query can still be semantically wrong. Parse and, where you can, dry-run
+  against the real schema before trusting output.
+- **Never interpolate output into a privileged connection.** Treat generated SQL
+  as untrusted input: run it read-only, with least privilege, on a connection
+  that cannot write or drop.
+- Multi-table joins, CTEs, window functions, subqueries, and DDL/DML are largely
+  out of distribution — the training data is dominated by single-table
+  `SELECT`s. Measured join accuracy is poor (see Limitations).
+- Dialect is not controllable. The model reproduces the source corpus's
+  conventions (double-quoted string literals, lower-cased comparison values),
+  which are not portable to every engine.
+- Not a general-purpose assistant. It emits a bare query, never prose.
+
+## How to get started
+
+```python
+import torch
+from transformers import AutoModelForCausalLM, AutoTokenizer, BitsAndBytesConfig
+from peft import PeftModel
+
+BASE = "zai-org/GLM-4.7-Flash"
+ADAPTER = "SASVAAI/GLM-4.7-Flash-sql-create-context"
+
+# 4-bit NF4 matches the numerics the adapter was trained against. A bf16 base
+# also works and scores the same (see Merged-weights equivalence) but needs
+# ~60 GB rather than ~22 GB.
+bnb = BitsAndBytesConfig(
+    load_in_4bit=True,
+    bnb_4bit_quant_type="nf4",
+    bnb_4bit_compute_dtype=torch.bfloat16,
+    bnb_4bit_use_double_quant=True,
+)
+
+tokenizer = AutoTokenizer.from_pretrained(ADAPTER)
+model = AutoModelForCausalLM.from_pretrained(
+    BASE, quantization_config=bnb, dtype=torch.bfloat16, device_map="auto"
+)
+model = PeftModel.from_pretrained(model, ADAPTER)
+model.eval()
+
+SYSTEM = (
+    "You are a text-to-SQL engine. Given a natural-language question and a "
+    "CREATE TABLE schema, output exactly one SQL query that answers the question "
+    "against the provided schema. Output only the raw SQL query on a single line "
+    "with no explanation, no markdown formatting, and no additional text."
+)
+
+question = "Which kingdom has Suin as its capital?"
+schema = "CREATE TABLE table_name_65 (name_of_kingdom VARCHAR, capital VARCHAR)"
+
+messages = [
+    {"role": "system", "content": SYSTEM},
+    {"role": "user", "content": f"{question}\n\n```\n{schema}\n```"},
+]
+prompt = tokenizer.apply_chat_template(
+    messages, tokenize=False, add_generation_prompt=True, enable_thinking=False
+)
+inputs = tokenizer(prompt, return_tensors="pt").to(model.device)
+
+out = model.generate(**inputs, max_new_tokens=128, do_sample=False)
+text = tokenizer.decode(out[0][inputs.input_ids.shape[-1]:], skip_special_tokens=True)
+print(text.strip().splitlines()[0])
+# -> SELECT name_of_kingdom FROM table_name_65 WHERE capital = "suin"
+```
+
+The base model is ~59 GB in bfloat16, or ~22 GB per GPU under 4-bit NF4.
+
+> Decoding matters. This model was evaluated with greedy decoding
+> (`do_sample=False`, `max_new_tokens=128`). Sampling will not reproduce the
+> reported numbers.
+
+**Serving note.** These adapter weights do **not** load as a vLLM LoRA on this
+architecture — vLLM's MLA path asserts inside
+`DeepSeekV2FusedQkvAProjLinear` because `q_a_proj` and `kv_a_proj_with_mqa` are
+fused into one module that a LoRA cannot be attached to. A merged build of these
+weights serves under vLLM without complaint. For vLLM deployment, merge first
+(`peft.merge_and_unload()`).
+
+## Training details
+
+**Data.** A 4,000-pair subset of
+[`b-mc2/sql-create-context`](https://huggingface.co/datasets/b-mc2/sql-create-context)
+(78,577 pairs, itself derived from WikiSQL and Spider), split 90/10 by this
+project's data-generation stage. Each record is
+`{"instruction": <question>, "input": <CREATE TABLE ...>, "output": <SQL>}`.
+The subset selection is a generated artifact, not a published split — the
+`id`/`instruction`/`input`/`gold` tuples in `predictions.jsonl` are the
+authoritative record of what was evaluated.
+
+| | |
+|---|---|
+| Train samples | 3,600 |
+| Validation samples | 400 |
+| Prompt format | chat template + system prompt (see Intended use) |
+
+### Method
+
+| | |
+|---|---|
+| SFT method | `qlora` |
+| Base quantisation during training | 4-bit NF4, double quant, bf16 compute |
+| Refinement stage | none |
+| Hardware | 4x NVIDIA H200 (141 GB), `torchrun --nproc_per_node=4` |
+
+`qlora` is one of three methods considered for this model, alongside 8-bit LoRA
+and attention-only 4-bit LoRA. All three were tried; `qlora` scored highest
+every time it was compared against the other two on the same data.
+
+No refinement stage ran; the published weights are the SFT adapter.
+
+### Final hyperparameters
+
+| Hyperparameter | Value | Source |
+|---|---|---|
+| `learning_rate` | 0.0002 | `training_args.bin` |
+| `lr_scheduler_type` | cosine | `training_args.bin` |
+| `num_train_epochs` | 3 | `training_args.bin` |
+| `per_device_train_batch_size` | 1 | `training_args.bin` |
+| `gradient_accumulation_steps` | 4 | `training_args.bin` |
+| `max_length` | 2048 | `training_args.bin` |
+| `warmup_ratio` | 0.05 | `training_args.bin` |
+| `weight_decay` | 0.01 | `training_args.bin` |
+| `optim` / `max_grad_norm` / `seed` | `adamw_torch` / 1.0 / 42 | `training_args.bin` |
+| `bf16` / `gradient_checkpointing` | true / true (`use_reentrant=False`) | `training_args.bin` |
+| `neftune_noise_alpha` / `packing` | None / false | `training_args.bin` |
+| `lora_r` / `lora_alpha` / `lora_dropout` | 64 / 128 / 0.05 | `adapter_config.json` |
+| `target_modules` | the 8 listed in Model details | `adapter_config.json` |
+
+**Effective batch size: 16** (`1 x 4 x 4`). Optimizer steps: 675.
+
+KD parameters are omitted deliberately — this is a `qlora` run, not
+`bf16_lora_kd`, so `KD_ALPHA`/`KD_BETA`/`KD_TEMPERATURE` carry inert defaults
+that would imply distillation that did not happen.
+
+This configuration is **not a unique optimum**. Other configurations reached the
+same `exact_match`; this one was published for being the simplest and cheapest
+of them — fewest epochs, shortest training time — and because it scored
+marginally higher on BLEU and ROUGE-L. Treat the values as a good working point,
+not a tuned maximum.
+
+**Observed training metrics.**
+
+| | |
+|---|---|
+| Final train loss | 0.4594 |
+| Mean train loss | 0.5646523337894016 |
+| Train runtime | 6392.6844s |
+| Total FLOPs | 2.3441226939026637e+17 |
+| Throughput | 1.689 samples/s, 0.106 steps/s |
+
+No eval loss was computed during training; the loop scores on generation, not
+perplexity. Loss falls from 2.3882 at step 10 to 1.1077 at step 20 and 0.5940 by
+step 170, then improves slowly to ~0.45 by step 660. The task is essentially
+learned within the first quarter of epoch 1; epochs 2 and 3 together buy roughly
+0.14 of training loss.
+
+## Evaluation
+
+**Protocol.** All 400 validation pairs, no sampling. Predictions generated
+greedily (`do_sample=False`, `max_new_tokens=128`) through the same chat template
+used in training, against a 4-bit NF4 base to match training numerics. The
+predicted query is the **first line** of the generation, stripped. No constrained
+decoding and no SQL grammar were applied. Exact match is byte equality against
+the reference query; BLEU and ROUGE-L are computed over the same strings.
+
+| Metric | Value |
+|---|---|
+| Exact match | 0.805000 (322 / 400) |
+| BLEU | 0.940082 |
+| ROUGE-L | 0.986055 |
+| Samples | 400 |
+
+**Baseline for comparison.** **Not measured.** The untuned
+`zai-org/GLM-4.7-Flash` was never scored on this split, so these numbers
+quantify the fine-tuned model's performance but do not establish how much of it
+the fine-tuning is responsible for.
+
+**This is a validation split, not a held-out test set.** Hyperparameters were
+selected against it, so expect optimistic bias. A clean estimate needs a third
+split that was never used for selection.
+
+## Limitations and bias
+
+**Exact match understates the model; BLEU and ROUGE-L overstate it.** The
+0.805 / 0.940 / 0.986 spread is the story of this model. Exact match is byte
+equality, so a semantically identical query loses the point on quoting or a
+missing `DISTINCT`:
+
+```
+question: Find the states where have some college students in tryout and their decisions are yes.
+schema:   CREATE TABLE tryout (cName VARCHAR, decision VARCHAR);
+          CREATE TABLE college (state VARCHAR, cName VARCHAR)
+gold:     SELECT DISTINCT T1.state FROM college AS T1 JOIN tryout AS T2 ON T1.cName = T2.cName WHERE T2.decision = 'yes'
+pred:     SELECT T1.state FROM college AS T1 JOIN tryout AS T2 ON T1.cName = T2.cName WHERE T2.decision = "yes"
+```
+
+Relaxing to case- and whitespace-insensitive comparison moves exact match from
+0.805 to **0.8125** (325/400) — so only 3 of the 78 misses are pure formatting.
+The other 75 are real semantic or structural errors. ROUGE-L at 0.986 mostly
+measures that both strings are short SQL over the same table names; it is not
+evidence of correctness.
+
+**Output is never degenerate.** All 400 golds and all 400 predictions begin with
+`SELECT`; the model never emitted prose, markdown, or an empty string. Format
+compliance is not the failure mode.
+
+**Joins are the failure mode.** Miss rate by gold-query feature:
+
+| Gold query contains | n | misses | miss rate |
+|---|---|---|---|
+| `JOIN` | 13 | 10 | **76.9%** |
+| `ORDER BY` | 3 | 1 | 33.3% |
+| aggregate (`COUNT`/`SUM`/`AVG`/`MIN`/`MAX`) | 138 | 37 | 26.8% |
+| `GROUP BY` | 8 | 2 | 25.0% |
+| multi-predicate `WHERE` (`AND`/`OR`) | 123 | 27 | 22.0% |
+| single-predicate `SELECT..WHERE` | 181 | 22 | 12.2% |
+
+The model is reliable on the shape it saw constantly (one table, one predicate)
+and unreliable on the shape it barely saw. **Do not deploy this on a
+multi-table schema.** Note the join, `ORDER BY`, and `GROUP BY` rows rest on
+13, 3, and 8 examples respectively — read them as a strong warning, not a precise
+rate.
+
+Complexity tracks length: correct predictions have a mean gold length of 10.6
+tokens, misses 12.8.
+
+**Dialect is baked in.** The model emits the source corpus's double-quoted
+string literals and lower-cases comparison values. On engines where `"x"` is an
+identifier rather than a string (PostgreSQL, ANSI mode), output will not run
+unmodified.
+
+**No execution or injection safety.** Correctness is measured only as string
+similarity to a reference. Nothing here prevents a generated query from being
+expensive, wrong, or destructive against a real database.
+
+**Inherits all biases and limitations of the base model.** This adapter changes
+0.3945% of the parameters and was not evaluated for social bias, safety, or
+fairness.
+
+## Merged-weights equivalence
+
+A merged build of these weights (base + adapter folded into one standalone bf16
+model, `W + (alpha/r) * B @ A`) was evaluated on the identical split:
+
+| Metric | Adapter (4-bit base) | Merged (bf16) | Delta |
+|---|---|---|---|
+| Exact match | 0.805000 | 0.805000 | 0.000000 |
+| BLEU | 0.940082 | 0.942079 | +0.001997 |
+| ROUGE-L | 0.986055 | 0.987663 | +0.001608 |
+
+**39 of 400 predictions differ** (9.75%) — far more churn than a bf16-trained
+adapter would show, because merging into an unquantised base genuinely changes
+the numerics the adapter was fitted against. The changes cancel exactly:
+**10 predictions flip correct→incorrect and 10 flip incorrect→correct**, leaving
+exact match identical and BLEU/ROUGE-L marginally higher.
+
+So a merged distribution is behaviourally equivalent in aggregate but not
+prediction-for-prediction. Merging is also the practical route to vLLM serving
+(see Serving note). MIT permits distributing derivative works, so publishing a
+merged build is allowed.
+
+## Environmental impact
+
+| | |
+|---|---|
+| Hardware | 4x NVIDIA H200 (141 GB) |
+| Training time | 106.5 minutes (6392.6844s) |
+| Cloud provider / region | on-premise |
+
+Covers the training of these published weights only. It excludes the wider
+hyperparameter search that selected them, which cost substantially more.
+
+## Framework versions
+
+- PEFT 0.18.1
+- TRL: 1.0.0
+- Transformers: 5.7.0.dev0
+- Pytorch: 2.5.1+cu121
+- Datasets: 4.8.4
+- Tokenizers: 0.22.2
+- bitsandbytes: 0.49.2
+
+`transformers` is a git-main build: GLM-4.7-Flash's `Glm4MoeLite` architecture is
+not in the stable PyPI release.
+
+## Licence
+
+Adapter weights: **MIT**, inherited from
+[`zai-org/GLM-4.7-Flash`](https://huggingface.co/zai-org/GLM-4.7-Flash)
+(verified via the Hub API). Training data:
+[`b-mc2/sql-create-context`](https://huggingface.co/datasets/b-mc2/sql-create-context),
+licensed **CC-BY-4.0** — downstream use should carry that attribution.
+
+## Citation
+
+```bibtex
+@misc{glm47flash_sql_create_context_lora_2026,
+  title  = {GLM-4.7-Flash Text-to-SQL (LoRA)},
+  author = {{SASVA AI Model Cognition Labs (MCL) Team}},
+  year   = {2026},
+  url    = {https://huggingface.co/SASVAAI/GLM-4.7-Flash-sql-create-context}
+}
+```

adapter_config.json

@@ -0,0 +1,47 @@
+{
+  "alora_invocation_tokens": null,
+  "alpha_pattern": {},
+  "arrow_config": null,
+  "auto_mapping": null,
+  "base_model_name_or_path": "zai-org/GLM-4.7-Flash",
+  "bias": "none",
+  "corda_config": null,
+  "ensure_weight_tying": false,
+  "eva_config": null,
+  "exclude_modules": null,
+  "fan_in_fan_out": false,
+  "inference_mode": true,
+  "init_lora_weights": true,
+  "layer_replication": null,
+  "layers_pattern": null,
+  "layers_to_transform": null,
+  "loftq_config": {},
+  "lora_alpha": 128,
+  "lora_bias": false,
+  "lora_dropout": 0.05,
+  "megatron_config": null,
+  "megatron_core": "megatron.core",
+  "modules_to_save": null,
+  "peft_type": "LORA",
+  "peft_version": "0.18.1",
+  "qalora_group_size": 16,
+  "r": 64,
+  "rank_pattern": {},
+  "revision": null,
+  "target_modules": [
+    "up_proj",
+    "down_proj",
+    "gate_proj",
+    "q_a_proj",
+    "kv_a_proj_with_mqa",
+    "o_proj",
+    "kv_b_proj",
+    "q_b_proj"
+  ],
+  "target_parameters": null,
+  "task_type": "CAUSAL_LM",
+  "trainable_token_indices": null,
+  "use_dora": false,
+  "use_qalora": false,
+  "use_rslora": false
+}

adapter_model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f7c4b6367c7ac51b2700e675946c6e0dffb7d037d46dadf8096fecd3757b9d41
+size 472671000

all_results.json

@@ -0,0 +1,7 @@
+{
+    "total_flos": 2.3441226939026637e+17,
+    "train_loss": 0.5646523337894016,
+    "train_runtime": 6392.6844,
+    "train_samples_per_second": 1.689,
+    "train_steps_per_second": 0.106
+}

chat_template.jinja

@@ -0,0 +1,86 @@
+[gMASK]<sop>
+{%- if tools -%}
+<|system|>
+# Tools
+
+You may call one or more functions to assist with the user query.
+
+You are provided with function signatures within <tools></tools> XML tags:
+<tools>
+{% for tool in tools %}
+{{ tool | tojson(ensure_ascii=False) }}
+{% endfor %}
+</tools>
+
+For each function call, output the function name and arguments within the following XML format:
+<tool_call>{function-name}<arg_key>{arg-key-1}</arg_key><arg_value>{arg-value-1}</arg_value><arg_key>{arg-key-2}</arg_key><arg_value>{arg-value-2}</arg_value>...</tool_call>{%- endif -%}
+{%- macro visible_text(content) -%}
+    {%- if content is string -%}
+        {{- content }}
+    {%- elif content is iterable and content is not mapping -%}
+        {%- for item in content -%}
+            {%- if item is mapping and item.type == 'text' -%}
+                {{- item.text }}
+            {%- elif item is string -%}
+                {{- item }}
+            {%- endif -%}
+        {%- endfor -%}
+    {%- else -%}
+        {{- content }}
+    {%- endif -%}
+{%- endmacro -%}
+{%- set ns = namespace(last_user_index=-1) %}
+{%- for m in messages %}
+    {%- if m.role == 'user' %}
+        {% set ns.last_user_index = loop.index0 -%}
+    {%- endif %}
+{%- endfor %}
+{% for m in messages %}
+{%- if m.role == 'user' -%}<|user|>{{ visible_text(m.content) }}
+{%- elif m.role == 'assistant' -%}
+<|assistant|>
+{%- set reasoning_content = '' %}
+{%- set content = visible_text(m.content) %}
+{%- if m.reasoning_content is string %}
+    {%- set reasoning_content = m.reasoning_content %}
+{%- else %}
+    {%- if '</think>' in content %}
+        {%- set reasoning_content = content.split('</think>')[0].rstrip('\n').split('<think>')[-1].lstrip('\n') %}
+        {%- set content = content.split('</think>')[-1].lstrip('\n') %}
+    {%- endif %}
+{%- endif %}
+{%- if ((clear_thinking is defined and not clear_thinking) or loop.index0 > ns.last_user_index) and reasoning_content -%}
+{{ '<think>' + reasoning_content.strip() +  '</think>'}}
+{%- else -%}
+{{ '</think>' }}
+{%- endif -%}
+{%- if content.strip() -%}
+{{ content.strip() }}
+{%- endif -%}
+{% if m.tool_calls %}
+{% for tc in m.tool_calls %}
+{%- if tc.function %}
+    {%- set tc = tc.function %}
+{%- endif %}
+{{- '<tool_call>' + tc.name -}}
+{% set _args = tc.arguments %}{% for k, v in _args.items() %}<arg_key>{{ k }}</arg_key><arg_value>{{ v | tojson(ensure_ascii=False) if v is not string else v }}</arg_value>{% endfor %}</tool_call>{% endfor %}
+{% endif %}
+{%- elif m.role == 'tool' -%}
+{%- if m.content is string -%}
+{%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+    {{- '<|observation|>' }}
+{%- endif %}
+{{- '<tool_response>' }}
+{{- m.content }}
+{{- '</tool_response>' }}
+{%- else -%}
+<|observation|>{% for tr in m.content %}
+<tool_response>{{ tr.output if tr.output is defined else tr }}</tool_response>{% endfor -%}
+{% endif -%}
+{%- elif m.role == 'system' -%}
+<|system|>{{ visible_text(m.content) }}
+{%- endif -%}
+{%- endfor -%}
+{%- if add_generation_prompt -%}
+    <|assistant|>{{- '</think>' if (enable_thinking is defined and not enable_thinking) else '<think>' -}}
+{%- endif -%}

eval_results.json

@@ -0,0 +1,8 @@
+{
+  "metrics": {
+    "exact_match": 0.805,
+    "bleu": 0.9400823275452048,
+    "rouge_l": 0.9860549005631489
+  },
+  "num_samples": 400
+}

tokenizer.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:19e773648cb4e65de8660ea6365e10acca112d42a854923df93db4a6f333a82d
+size 20217442

tokenizer_config.json

@@ -0,0 +1,13 @@
+{
+  "backend": "tokenizers",
+  "clean_up_tokenization_spaces": false,
+  "do_lower_case": false,
+  "eos_token": "<|endoftext|>",
+  "is_local": false,
+  "local_files_only": false,
+  "model_max_length": 128000,
+  "pad_token": "<|endoftext|>",
+  "padding_side": "right",
+  "remove_space": false,
+  "tokenizer_class": "TokenizersBackend"
+}

train_results.json

@@ -0,0 +1,7 @@
+{
+    "total_flos": 2.3441226939026637e+17,
+    "train_loss": 0.5646523337894016,
+    "train_runtime": 6392.6844,
+    "train_samples_per_second": 1.689,
+    "train_steps_per_second": 0.106
+}

trainer_state.json

@@ -0,0 +1,724 @@
+{
+  "best_global_step": 675,
+  "best_metric": 0.5975516438484192,
+  "best_model_checkpoint": "/home/svadouser1/pooja/llm_autotuning/checkpoints/exp_20260827_221617/checkpoint-675",
+  "epoch": 3.0,
+  "eval_steps": 9999,
+  "global_step": 675,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "entropy": 1.461071628332138,
+      "epoch": 0.044444444444444446,
+      "grad_norm": 1.6171875,
+      "learning_rate": 5.294117647058824e-05,
+      "loss": 2.388203811645508,
+      "mean_token_accuracy": 0.6123090386390686,
+      "num_tokens": 19486.0,
+      "step": 10
+    },
+    {
+      "entropy": 0.9562640145421029,
+      "epoch": 0.08888888888888889,
+      "grad_norm": 0.6875,
+      "learning_rate": 0.00011176470588235294,
+      "loss": 1.1077078819274901,
+      "mean_token_accuracy": 0.8430786401033401,
+      "num_tokens": 38826.0,
+      "step": 20
+    },
+    {
+      "entropy": 0.77465368360281,
+      "epoch": 0.13333333333333333,
+      "grad_norm": 0.455078125,
+      "learning_rate": 0.00017058823529411766,
+      "loss": 0.8623821258544921,
+      "mean_token_accuracy": 0.864073084294796,
+      "num_tokens": 58244.0,
+      "step": 30
+    },
+    {
+      "entropy": 0.7483784720301628,
+      "epoch": 0.17777777777777778,
+      "grad_norm": 0.423828125,
+      "learning_rate": 0.00019996997576394573,
+      "loss": 0.7713714599609375,
+      "mean_token_accuracy": 0.8697837173938752,
+      "num_tokens": 77711.0,
+      "step": 40
+    },
+    {
+      "entropy": 0.7391687169671058,
+      "epoch": 0.2222222222222222,
+      "grad_norm": 0.376953125,
+      "learning_rate": 0.00019972989003925543,
+      "loss": 0.7115508556365967,
+      "mean_token_accuracy": 0.8751530349254608,
+      "num_tokens": 97104.0,
+      "step": 50
+    },
+    {
+      "entropy": 0.7042164742946625,
+      "epoch": 0.26666666666666666,
+      "grad_norm": 0.330078125,
+      "learning_rate": 0.00019925029517454195,
+      "loss": 0.6672511577606202,
+      "mean_token_accuracy": 0.8725973218679428,
+      "num_tokens": 116498.0,
+      "step": 60
+    },
+    {
+      "entropy": 0.6745826601982117,
+      "epoch": 0.3111111111111111,
+      "grad_norm": 0.361328125,
+      "learning_rate": 0.00019853234295442638,
+      "loss": 0.6365277290344238,
+      "mean_token_accuracy": 0.872233435511589,
+      "num_tokens": 136089.0,
+      "step": 70
+    },
+    {
+      "entropy": 0.6878040120005607,
+      "epoch": 0.35555555555555557,
+      "grad_norm": 0.37109375,
+      "learning_rate": 0.00019757775759738272,
+      "loss": 0.6395976066589355,
+      "mean_token_accuracy": 0.8741081401705741,
+      "num_tokens": 155486.0,
+      "step": 80
+    },
+    {
+      "entropy": 0.6898319900035859,
+      "epoch": 0.4,
+      "grad_norm": 0.330078125,
+      "learning_rate": 0.00019638883161489224,
+      "loss": 0.6368544578552247,
+      "mean_token_accuracy": 0.8765213042497635,
+      "num_tokens": 175116.0,
+      "step": 90
+    },
+    {
+      "entropy": 0.656438185274601,
+      "epoch": 0.4444444444444444,
+      "grad_norm": 0.33203125,
+      "learning_rate": 0.0001949684203057978,
+      "loss": 0.6183670043945313,
+      "mean_token_accuracy": 0.8779006212949753,
+      "num_tokens": 194495.0,
+      "step": 100
+    },
+    {
+      "entropy": 0.6711793914437294,
+      "epoch": 0.4888888888888889,
+      "grad_norm": 0.388671875,
+      "learning_rate": 0.00019331993489907977,
+      "loss": 0.633416748046875,
+      "mean_token_accuracy": 0.8727847129106522,
+      "num_tokens": 214311.0,
+      "step": 110
+    },
+    {
+      "entropy": 0.6290019288659096,
+      "epoch": 0.5333333333333333,
+      "grad_norm": 0.33203125,
+      "learning_rate": 0.00019144733436152284,
+      "loss": 0.5889303684234619,
+      "mean_token_accuracy": 0.8814716726541519,
+      "num_tokens": 233676.0,
+      "step": 120
+    },
+    {
+      "entropy": 0.6458568304777146,
+      "epoch": 0.5777777777777777,
+      "grad_norm": 0.318359375,
+      "learning_rate": 0.00018935511588994715,
+      "loss": 0.6043062686920166,
+      "mean_token_accuracy": 0.8794952735304833,
+      "num_tokens": 253188.0,
+      "step": 130
+    },
+    {
+      "entropy": 0.6591948792338371,
+      "epoch": 0.6222222222222222,
+      "grad_norm": 0.32421875,
+      "learning_rate": 0.000187048304110838,
+      "loss": 0.6233312129974365,
+      "mean_token_accuracy": 0.873698017001152,
+      "num_tokens": 273185.0,
+      "step": 140
+    },
+    {
+      "entropy": 0.6109217047691345,
+      "epoch": 0.6666666666666666,
+      "grad_norm": 0.291015625,
+      "learning_rate": 0.00018453243901331195,
+      "loss": 0.5925732135772706,
+      "mean_token_accuracy": 0.884858050942421,
+      "num_tokens": 292431.0,
+      "step": 150
+    },
+    {
+      "entropy": 0.6377103522419929,
+      "epoch": 0.7111111111111111,
+      "grad_norm": 0.3671875,
+      "learning_rate": 0.00018181356264439905,
+      "loss": 0.5884737968444824,
+      "mean_token_accuracy": 0.8802034169435501,
+      "num_tokens": 311687.0,
+      "step": 160
+    },
+    {
+      "entropy": 0.6447647094726563,
+      "epoch": 0.7555555555555555,
+      "grad_norm": 0.3046875,
+      "learning_rate": 0.0001788982045985939,
+      "loss": 0.5939778327941895,
+      "mean_token_accuracy": 0.8817022785544395,
+      "num_tokens": 330813.0,
+      "step": 170
+    },
+    {
+      "entropy": 0.6029744669795036,
+      "epoch": 0.8,
+      "grad_norm": 0.3125,
+      "learning_rate": 0.00017579336633652317,
+      "loss": 0.5752908706665039,
+      "mean_token_accuracy": 0.8855120778083801,
+      "num_tokens": 350308.0,
+      "step": 180
+    },
+    {
+      "entropy": 0.6441277623176574,
+      "epoch": 0.8444444444444444,
+      "grad_norm": 0.259765625,
+      "learning_rate": 0.00017250650437038964,
+      "loss": 0.5958236217498779,
+      "mean_token_accuracy": 0.8808144956827164,
+      "num_tokens": 370083.0,
+      "step": 190
+    },
+    {
+      "entropy": 0.5766214028000831,
+      "epoch": 0.8888888888888888,
+      "grad_norm": 0.2734375,
+      "learning_rate": 0.0001690455123565743,
+      "loss": 0.5667627811431885,
+      "mean_token_accuracy": 0.8866458177566529,
+      "num_tokens": 389516.0,
+      "step": 200
+    },
+    {
+      "entropy": 0.6425503984093666,
+      "epoch": 0.9333333333333333,
+      "grad_norm": 0.3046875,
+      "learning_rate": 0.00016541870213840243,
+      "loss": 0.6137699127197266,
+      "mean_token_accuracy": 0.8819929376244545,
+      "num_tokens": 408940.0,
+      "step": 210
+    },
+    {
+      "entropy": 0.6159977808594703,
+      "epoch": 0.9777777777777777,
+      "grad_norm": 0.265625,
+      "learning_rate": 0.0001616347837846011,
+      "loss": 0.5788507461547852,
+      "mean_token_accuracy": 0.8852896198630333,
+      "num_tokens": 428166.0,
+      "step": 220
+    },
+    {
+      "entropy": 0.6020576372742653,
+      "epoch": 1.0222222222222221,
+      "grad_norm": 0.2353515625,
+      "learning_rate": 0.000157702844671387,
+      "loss": 0.5500371932983399,
+      "mean_token_accuracy": 0.8901642486453056,
+      "num_tokens": 447438.0,
+      "step": 230
+    },
+    {
+      "entropy": 0.5630205765366554,
+      "epoch": 1.0666666666666667,
+      "grad_norm": 0.275390625,
+      "learning_rate": 0.00015363232765842089,
+      "loss": 0.5382141590118408,
+      "mean_token_accuracy": 0.8885036528110504,
+      "num_tokens": 467009.0,
+      "step": 240
+    },
+    {
+      "entropy": 0.5631943918764591,
+      "epoch": 1.1111111111111112,
+      "grad_norm": 0.306640625,
+      "learning_rate": 0.00014943300841104094,
+      "loss": 0.5194426536560058,
+      "mean_token_accuracy": 0.8929360061883926,
+      "num_tokens": 486591.0,
+      "step": 250
+    },
+    {
+      "entropy": 0.5655099518597126,
+      "epoch": 1.1555555555555554,
+      "grad_norm": 0.259765625,
+      "learning_rate": 0.0001451149719232366,
+      "loss": 0.5290531158447266,
+      "mean_token_accuracy": 0.8912677451968193,
+      "num_tokens": 506180.0,
+      "step": 260
+    },
+    {
+      "entropy": 0.5660845704376698,
+      "epoch": 1.2,
+      "grad_norm": 0.326171875,
+      "learning_rate": 0.00014068858829774608,
+      "loss": 0.5366004943847656,
+      "mean_token_accuracy": 0.8909024119377136,
+      "num_tokens": 525696.0,
+      "step": 270
+    },
+    {
+      "entropy": 0.5818345256149768,
+      "epoch": 1.2444444444444445,
+      "grad_norm": 0.26171875,
+      "learning_rate": 0.0001361644878414428,
+      "loss": 0.5424720764160156,
+      "mean_token_accuracy": 0.8896975666284561,
+      "num_tokens": 545392.0,
+      "step": 280
+    },
+    {
+      "entropy": 0.5546154774725437,
+      "epoch": 1.2888888888888888,
+      "grad_norm": 0.3359375,
+      "learning_rate": 0.00013155353553582057,
+      "loss": 0.5176132202148438,
+      "mean_token_accuracy": 0.8928953528404235,
+      "num_tokens": 565036.0,
+      "step": 290
+    },
+    {
+      "entropy": 0.5676225990056991,
+      "epoch": 1.3333333333333333,
+      "grad_norm": 0.275390625,
+      "learning_rate": 0.0001268668049438902,
+      "loss": 0.5130613803863525,
+      "mean_token_accuracy": 0.8928169339895249,
+      "num_tokens": 584507.0,
+      "step": 300
+    },
+    {
+      "entropy": 0.5529272347688675,
+      "epoch": 1.3777777777777778,
+      "grad_norm": 0.326171875,
+      "learning_rate": 0.0001221155516161506,
+      "loss": 0.5317890644073486,
+      "mean_token_accuracy": 0.8913554430007935,
+      "num_tokens": 604008.0,
+      "step": 310
+    },
+    {
+      "entropy": 0.5658974438905716,
+      "epoch": 1.4222222222222223,
+      "grad_norm": 0.26953125,
+      "learning_rate": 0.0001173111860595032,
+      "loss": 0.5273444652557373,
+      "mean_token_accuracy": 0.8927861481904984,
+      "num_tokens": 623480.0,
+      "step": 320
+    },
+    {
+      "entropy": 0.5697685681283474,
+      "epoch": 1.4666666666666668,
+      "grad_norm": 0.322265625,
+      "learning_rate": 0.00011246524633402573,
+      "loss": 0.5287697792053223,
+      "mean_token_accuracy": 0.8912984907627106,
+      "num_tokens": 642979.0,
+      "step": 330
+    },
+    {
+      "entropy": 0.5488456651568413,
+      "epoch": 1.511111111111111,
+      "grad_norm": 0.326171875,
+      "learning_rate": 0.00010758937034341787,
+      "loss": 0.513739824295044,
+      "mean_token_accuracy": 0.8952319726347924,
+      "num_tokens": 662396.0,
+      "step": 340
+    },
+    {
+      "entropy": 0.55315260887146,
+      "epoch": 1.5555555555555556,
+      "grad_norm": 0.26953125,
+      "learning_rate": 0.00010269526788566408,
+      "loss": 0.5225533485412598,
+      "mean_token_accuracy": 0.8937178790569306,
+      "num_tokens": 681644.0,
+      "step": 350
+    },
+    {
+      "entropy": 0.5738558314740658,
+      "epoch": 1.6,
+      "grad_norm": 0.29296875,
+      "learning_rate": 9.779469253103684e-05,
+      "loss": 0.5239317417144775,
+      "mean_token_accuracy": 0.8908430561423302,
+      "num_tokens": 701030.0,
+      "step": 360
+    },
+    {
+      "entropy": 0.5420261971652508,
+      "epoch": 1.6444444444444444,
+      "grad_norm": 0.259765625,
+      "learning_rate": 9.289941339497719e-05,
+      "loss": 0.5274893283843994,
+      "mean_token_accuracy": 0.8941561266779899,
+      "num_tokens": 720294.0,
+      "step": 370
+    },
+    {
+      "entropy": 0.577882957458496,
+      "epoch": 1.6888888888888889,
+      "grad_norm": 0.33203125,
+      "learning_rate": 8.802118687364284e-05,
+      "loss": 0.5202207088470459,
+      "mean_token_accuracy": 0.8931182771921158,
+      "num_tokens": 739607.0,
+      "step": 380
+    },
+    {
+      "entropy": 0.5461296208202839,
+      "epoch": 1.7333333333333334,
+      "grad_norm": 0.30859375,
+      "learning_rate": 8.317172841000173e-05,
+      "loss": 0.5105932235717774,
+      "mean_token_accuracy": 0.8917569547891617,
+      "num_tokens": 759370.0,
+      "step": 390
+    },
+    {
+      "entropy": 0.5631573006510735,
+      "epoch": 1.7777777777777777,
+      "grad_norm": 0.298828125,
+      "learning_rate": 7.836268435827875e-05,
+      "loss": 0.5263056755065918,
+      "mean_token_accuracy": 0.8906426534056664,
+      "num_tokens": 778969.0,
+      "step": 400
+    },
+    {
+      "entropy": 0.5476421102881431,
+      "epoch": 1.8222222222222222,
+      "grad_norm": 0.255859375,
+      "learning_rate": 7.360560401432401e-05,
+      "loss": 0.5025547504425049,
+      "mean_token_accuracy": 0.8946620702743531,
+      "num_tokens": 798105.0,
+      "step": 410
+    },
+    {
+      "entropy": 0.5343898519873619,
+      "epoch": 1.8666666666666667,
+      "grad_norm": 0.27734375,
+      "learning_rate": 6.891191187907455e-05,
+      "loss": 0.49821176528930666,
+      "mean_token_accuracy": 0.8958980679512024,
+      "num_tokens": 817238.0,
+      "step": 420
+    },
+    {
+      "entropy": 0.5404686734080315,
+      "epoch": 1.911111111111111,
+      "grad_norm": 0.296875,
+      "learning_rate": 6.429288022172068e-05,
+      "loss": 0.5022043704986572,
+      "mean_token_accuracy": 0.8958747044205666,
+      "num_tokens": 836646.0,
+      "step": 430
+    },
+    {
+      "entropy": 0.5426375091075897,
+      "epoch": 1.9555555555555557,
+      "grad_norm": 0.279296875,
+      "learning_rate": 5.9759602008468996e-05,
+      "loss": 0.5035938262939453,
+      "mean_token_accuracy": 0.8944751426577568,
+      "num_tokens": 856282.0,
+      "step": 440
+    },
+    {
+      "entropy": 0.5425564736127854,
+      "epoch": 2.0,
+      "grad_norm": 0.263671875,
+      "learning_rate": 5.532296426191539e-05,
+      "loss": 0.5004886627197266,
+      "mean_token_accuracy": 0.8947419315576554,
+      "num_tokens": 875656.0,
+      "step": 450
+    },
+    {
+      "entropy": 0.526330380141735,
+      "epoch": 2.0444444444444443,
+      "grad_norm": 0.279296875,
+      "learning_rate": 5.0993621915007785e-05,
+      "loss": 0.44887552261352537,
+      "mean_token_accuracy": 0.9024429768323898,
+      "num_tokens": 894973.0,
+      "step": 460
+    },
+    {
+      "entropy": 0.4774711772799492,
+      "epoch": 2.088888888888889,
+      "grad_norm": 0.365234375,
+      "learning_rate": 4.678197222239035e-05,
+      "loss": 0.44640169143676756,
+      "mean_token_accuracy": 0.9045588001608849,
+      "num_tokens": 914579.0,
+      "step": 470
+    },
+    {
+      "entropy": 0.4789727419614792,
+      "epoch": 2.1333333333333333,
+      "grad_norm": 0.3828125,
+      "learning_rate": 4.269812979058235e-05,
+      "loss": 0.44276885986328124,
+      "mean_token_accuracy": 0.9066318318247795,
+      "num_tokens": 933807.0,
+      "step": 480
+    },
+    {
+      "entropy": 0.47537712305784224,
+      "epoch": 2.1777777777777776,
+      "grad_norm": 0.359375,
+      "learning_rate": 3.875190228695862e-05,
+      "loss": 0.4367781639099121,
+      "mean_token_accuracy": 0.9070042923092843,
+      "num_tokens": 953241.0,
+      "step": 490
+    },
+    {
+      "entropy": 0.4877164676785469,
+      "epoch": 2.2222222222222223,
+      "grad_norm": 0.328125,
+      "learning_rate": 3.4952766885868346e-05,
+      "loss": 0.4422135353088379,
+      "mean_token_accuracy": 0.905805604159832,
+      "num_tokens": 972758.0,
+      "step": 500
+    },
+    {
+      "entropy": 0.4915403999388218,
+      "epoch": 2.2666666666666666,
+      "grad_norm": 0.376953125,
+      "learning_rate": 3.130984750845885e-05,
+      "loss": 0.4357293128967285,
+      "mean_token_accuracy": 0.9067289993166924,
+      "num_tokens": 992534.0,
+      "step": 510
+    },
+    {
+      "entropy": 0.4957615494728088,
+      "epoch": 2.311111111111111,
+      "grad_norm": 0.33984375,
+      "learning_rate": 2.7831892910864434e-05,
+      "loss": 0.44167218208312986,
+      "mean_token_accuracy": 0.9066730305552483,
+      "num_tokens": 1012003.0,
+      "step": 520
+    },
+    {
+      "entropy": 0.48269262462854384,
+      "epoch": 2.3555555555555556,
+      "grad_norm": 0.375,
+      "learning_rate": 2.4527255673383565e-05,
+      "loss": 0.437894344329834,
+      "mean_token_accuracy": 0.904815036058426,
+      "num_tokens": 1031505.0,
+      "step": 530
+    },
+    {
+      "entropy": 0.47741171419620515,
+      "epoch": 2.4,
+      "grad_norm": 0.4140625,
+      "learning_rate": 2.140387214110322e-05,
+      "loss": 0.4400351047515869,
+      "mean_token_accuracy": 0.9068154886364936,
+      "num_tokens": 1050914.0,
+      "step": 540
+    },
+    {
+      "entropy": 0.4727069653570652,
+      "epoch": 2.4444444444444446,
+      "grad_norm": 0.35546875,
+      "learning_rate": 1.846924336414474e-05,
+      "loss": 0.43111190795898435,
+      "mean_token_accuracy": 0.9075597256422043,
+      "num_tokens": 1070218.0,
+      "step": 550
+    },
+    {
+      "entropy": 0.4883471392095089,
+      "epoch": 2.488888888888889,
+      "grad_norm": 0.314453125,
+      "learning_rate": 1.5730417083304573e-05,
+      "loss": 0.4493571758270264,
+      "mean_token_accuracy": 0.9073263511061669,
+      "num_tokens": 1089667.0,
+      "step": 560
+    },
+    {
+      "entropy": 0.4878625735640526,
+      "epoch": 2.533333333333333,
+      "grad_norm": 0.375,
+      "learning_rate": 1.3193970804352952e-05,
+      "loss": 0.44793548583984377,
+      "mean_token_accuracy": 0.9050837978720665,
+      "num_tokens": 1109201.0,
+      "step": 570
+    },
+    {
+      "entropy": 0.48537297546863556,
+      "epoch": 2.5777777777777775,
+      "grad_norm": 0.337890625,
+      "learning_rate": 1.08659960016387e-05,
+      "loss": 0.4368610382080078,
+      "mean_token_accuracy": 0.9097044110298157,
+      "num_tokens": 1128649.0,
+      "step": 580
+    },
+    {
+      "entropy": 0.4859867602586746,
+      "epoch": 2.6222222222222222,
+      "grad_norm": 0.328125,
+      "learning_rate": 8.75208348893667e-06,
+      "loss": 0.4249687194824219,
+      "mean_token_accuracy": 0.9087634190917016,
+      "num_tokens": 1148353.0,
+      "step": 590
+    },
+    {
+      "entropy": 0.4833585321903229,
+      "epoch": 2.6666666666666665,
+      "grad_norm": 0.3828125,
+      "learning_rate": 6.857309992670624e-06,
+      "loss": 0.4455591678619385,
+      "mean_token_accuracy": 0.9075253725051879,
+      "num_tokens": 1167638.0,
+      "step": 600
+    },
+    {
+      "entropy": 0.4794360339641571,
+      "epoch": 2.7111111111111112,
+      "grad_norm": 0.3203125,
+      "learning_rate": 5.186225959757207e-06,
+      "loss": 0.43627562522888186,
+      "mean_token_accuracy": 0.907138803601265,
+      "num_tokens": 1186894.0,
+      "step": 610
+    },
+    {
+      "entropy": 0.4800324112176895,
+      "epoch": 2.7555555555555555,
+      "grad_norm": 0.365234375,
+      "learning_rate": 3.7428446293514386e-06,
+      "loss": 0.42800016403198243,
+      "mean_token_accuracy": 0.9089159920811654,
+      "num_tokens": 1206224.0,
+      "step": 620
+    },
+    {
+      "entropy": 0.48999542444944383,
+      "epoch": 2.8,
+      "grad_norm": 0.31640625,
+      "learning_rate": 2.5306323947385746e-06,
+      "loss": 0.44663453102111816,
+      "mean_token_accuracy": 0.9060632780194282,
+      "num_tokens": 1225569.0,
+      "step": 630
+    },
+    {
+      "entropy": 0.4874999448657036,
+      "epoch": 2.8444444444444446,
+      "grad_norm": 0.326171875,
+      "learning_rate": 1.5525004785192143e-06,
+      "loss": 0.44841961860656737,
+      "mean_token_accuracy": 0.9045136958360672,
+      "num_tokens": 1244959.0,
+      "step": 640
+    },
+    {
+      "entropy": 0.49936808943748473,
+      "epoch": 2.888888888888889,
+      "grad_norm": 0.4296875,
+      "learning_rate": 8.107979410802769e-07,
+      "loss": 0.4453989028930664,
+      "mean_token_accuracy": 0.9050952970981598,
+      "num_tokens": 1264904.0,
+      "step": 650
+    },
+    {
+      "entropy": 0.49325661584734914,
+      "epoch": 2.9333333333333336,
+      "grad_norm": 0.373046875,
+      "learning_rate": 3.0730603914255193e-07,
+      "loss": 0.44756379127502444,
+      "mean_token_accuracy": 0.9043820068240166,
+      "num_tokens": 1284311.0,
+      "step": 660
+    },
+    {
+      "entropy": 0.5068465434014797,
+      "epoch": 2.977777777777778,
+      "grad_norm": 0.373046875,
+      "learning_rate": 4.323394793315228e-08,
+      "loss": 0.4593667030334473,
+      "mean_token_accuracy": 0.9011617928743363,
+      "num_tokens": 1303914.0,
+      "step": 670
+    },
+    {
+      "epoch": 3.0,
+      "eval_entropy": 0.5207519015669823,
+      "eval_loss": 0.5975516438484192,
+      "eval_mean_token_accuracy": 0.8822531878948212,
+      "eval_num_tokens": 1313484.0,
+      "eval_runtime": 57.0793,
+      "eval_samples_per_second": 7.008,
+      "eval_steps_per_second": 1.752,
+      "step": 675
+    },
+    {
+      "epoch": 3.0,
+      "step": 675,
+      "total_flos": 2.3441226939026637e+17,
+      "train_loss": 0.5646523337894016,
+      "train_runtime": 6392.6844,
+      "train_samples_per_second": 1.689,
+      "train_steps_per_second": 0.106
+    }
+  ],
+  "logging_steps": 10,
+  "max_steps": 675,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 3,
+  "save_steps": 9999,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": true
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 2.3441226939026637e+17,
+  "train_batch_size": 1,
+  "trial_name": null,
+  "trial_params": null
+}