Training in progress, epoch 1

.gitattributes

@@ -1,35 +1,60 @@
-*.7z filter=lfs diff=lfs merge=lfs -text
-*.arrow filter=lfs diff=lfs merge=lfs -text
-*.bin filter=lfs diff=lfs merge=lfs -text
-*.bz2 filter=lfs diff=lfs merge=lfs -text
-*.ckpt filter=lfs diff=lfs merge=lfs -text
-*.ftz filter=lfs diff=lfs merge=lfs -text
-*.gz filter=lfs diff=lfs merge=lfs -text
-*.h5 filter=lfs diff=lfs merge=lfs -text
-*.joblib filter=lfs diff=lfs merge=lfs -text
-*.lfs.* filter=lfs diff=lfs merge=lfs -text
-*.mlmodel filter=lfs diff=lfs merge=lfs -text
-*.model filter=lfs diff=lfs merge=lfs -text
-*.msgpack filter=lfs diff=lfs merge=lfs -text
-*.npy filter=lfs diff=lfs merge=lfs -text
-*.npz filter=lfs diff=lfs merge=lfs -text
-*.onnx filter=lfs diff=lfs merge=lfs -text
-*.ot filter=lfs diff=lfs merge=lfs -text
-*.parquet filter=lfs diff=lfs merge=lfs -text
-*.pb filter=lfs diff=lfs merge=lfs -text
-*.pickle filter=lfs diff=lfs merge=lfs -text
-*.pkl filter=lfs diff=lfs merge=lfs -text
-*.pt filter=lfs diff=lfs merge=lfs -text
-*.pth filter=lfs diff=lfs merge=lfs -text
-*.rar filter=lfs diff=lfs merge=lfs -text
-*.safetensors filter=lfs diff=lfs merge=lfs -text
-saved_model/**/* filter=lfs diff=lfs merge=lfs -text
-*.tar.* filter=lfs diff=lfs merge=lfs -text
-*.tar filter=lfs diff=lfs merge=lfs -text
-*.tflite filter=lfs diff=lfs merge=lfs -text
-*.tgz filter=lfs diff=lfs merge=lfs -text
-*.wasm filter=lfs diff=lfs merge=lfs -text
-*.xz 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
+*.7z filter=lfs diff=lfs merge=lfs -text
+*.arrow filter=lfs diff=lfs merge=lfs -text
+*.bin filter=lfs diff=lfs merge=lfs -text
+*.bz2 filter=lfs diff=lfs merge=lfs -text
+*.ckpt filter=lfs diff=lfs merge=lfs -text
+*.ftz filter=lfs diff=lfs merge=lfs -text
+*.gz filter=lfs diff=lfs merge=lfs -text
+*.h5 filter=lfs diff=lfs merge=lfs -text
+*.joblib filter=lfs diff=lfs merge=lfs -text
+*.lfs.* filter=lfs diff=lfs merge=lfs -text
+*.lz4 filter=lfs diff=lfs merge=lfs -text
+*.mds filter=lfs diff=lfs merge=lfs -text
+*.mlmodel filter=lfs diff=lfs merge=lfs -text
+*.model filter=lfs diff=lfs merge=lfs -text
+*.msgpack filter=lfs diff=lfs merge=lfs -text
+*.npy filter=lfs diff=lfs merge=lfs -text
+*.npz filter=lfs diff=lfs merge=lfs -text
+*.onnx filter=lfs diff=lfs merge=lfs -text
+*.ot filter=lfs diff=lfs merge=lfs -text
+*.parquet filter=lfs diff=lfs merge=lfs -text
+*.pb filter=lfs diff=lfs merge=lfs -text
+*.pickle filter=lfs diff=lfs merge=lfs -text
+*.pkl filter=lfs diff=lfs merge=lfs -text
+*.pt filter=lfs diff=lfs merge=lfs -text
+*.pth filter=lfs diff=lfs merge=lfs -text
+*.rar filter=lfs diff=lfs merge=lfs -text
+*.safetensors filter=lfs diff=lfs merge=lfs -text
+saved_model/**/* filter=lfs diff=lfs merge=lfs -text
+*.tar.* filter=lfs diff=lfs merge=lfs -text
+*.tar filter=lfs diff=lfs merge=lfs -text
+*.tflite filter=lfs diff=lfs merge=lfs -text
+*.tgz filter=lfs diff=lfs merge=lfs -text
+*.wasm filter=lfs diff=lfs merge=lfs -text
+*.xz 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
+# Audio files - uncompressed
+*.pcm filter=lfs diff=lfs merge=lfs -text
+*.sam filter=lfs diff=lfs merge=lfs -text
+*.raw filter=lfs diff=lfs merge=lfs -text
+# Audio files - compressed
+*.aac filter=lfs diff=lfs merge=lfs -text
+*.flac filter=lfs diff=lfs merge=lfs -text
+*.mp3 filter=lfs diff=lfs merge=lfs -text
+*.ogg filter=lfs diff=lfs merge=lfs -text
+*.wav filter=lfs diff=lfs merge=lfs -text
+# Image files - uncompressed
+*.bmp filter=lfs diff=lfs merge=lfs -text
+*.gif filter=lfs diff=lfs merge=lfs -text
+*.png filter=lfs diff=lfs merge=lfs -text
+*.tiff filter=lfs diff=lfs merge=lfs -text
+# Image files - compressed
+*.jpg filter=lfs diff=lfs merge=lfs -text
+*.jpeg filter=lfs diff=lfs merge=lfs -text
+*.webp filter=lfs diff=lfs merge=lfs -text
+# Video files - compressed
+*.mp4 filter=lfs diff=lfs merge=lfs -text
+*.webm filter=lfs diff=lfs merge=lfs -text
+tokenizer.json filter=lfs diff=lfs merge=lfs -text

README.md

@@ -0,0 +1,58 @@
+---
+base_model: google/gemma-3-270m-it
+library_name: transformers
+model_name: MobileGameNPC
+tags:
+- generated_from_trainer
+- trl
+- sft
+licence: license
+---
+
+# Model Card for MobileGameNPC
+
+This model is a fine-tuned version of [google/gemma-3-270m-it](https://huggingface.co/google/gemma-3-270m-it).
+It has been trained using [TRL](https://github.com/huggingface/trl).
+
+## Quick start
+
+```python
+from transformers import pipeline
+
+question = "If you had a time machine, but could only go to the past or the future once and never return, which would you choose and why?"
+generator = pipeline("text-generation", model="fif3/MobileGameNPC", device="cuda")
+output = generator([{"role": "user", "content": question}], max_new_tokens=128, return_full_text=False)[0]
+print(output["generated_text"])
+```
+
+## Training procedure
+
+ 
+
+
+This model was trained with SFT.
+
+### Framework versions
+
+- TRL: 0.21.0
+- Transformers: 4.55.3
+- Pytorch: 2.8.0
+- Datasets: 4.0.0
+- Tokenizers: 0.21.4
+
+## Citations
+
+
+
+Cite TRL as:
+    
+```bibtex
+@misc{vonwerra2022trl,
+	title        = {{TRL: Transformer Reinforcement Learning}},
+	author       = {Leandro von Werra and Younes Belkada and Lewis Tunstall and Edward Beeching and Tristan Thrush and Nathan Lambert and Shengyi Huang and Kashif Rasul and Quentin Gallou{\'e}dec},
+	year         = 2020,
+	journal      = {GitHub repository},
+	publisher    = {GitHub},
+	howpublished = {\url{https://github.com/huggingface/trl}}
+}
+```

added_tokens.json

@@ -0,0 +1,3 @@
+{
+  "<image_soft_token>": 262144
+}

chat_template.jinja

@@ -0,0 +1,47 @@
+{{ bos_token }}
+{%- if messages[0]['role'] == 'system' -%}
+    {%- if messages[0]['content'] is string -%}
+        {%- set first_user_prefix = messages[0]['content'] + '
+
+' -%}
+    {%- else -%}
+        {%- set first_user_prefix = messages[0]['content'][0]['text'] + '
+
+' -%}
+    {%- endif -%}
+    {%- set loop_messages = messages[1:] -%}
+{%- else -%}
+    {%- set first_user_prefix = "" -%}
+    {%- set loop_messages = messages -%}
+{%- endif -%}
+{%- for message in loop_messages -%}
+    {%- if (message['role'] == 'user') != (loop.index0 % 2 == 0) -%}
+        {{ raise_exception("Conversation roles must alternate user/assistant/user/assistant/...") }}
+    {%- endif -%}
+    {%- if (message['role'] == 'assistant') -%}
+        {%- set role = "model" -%}
+    {%- else -%}
+        {%- set role = message['role'] -%}
+    {%- endif -%}
+    {{ '<start_of_turn>' + role + '
+' + (first_user_prefix if loop.first else "") }}
+    {%- if message['content'] is string -%}
+        {{ message['content'] | trim }}
+    {%- elif message['content'] is iterable -%}
+        {%- for item in message['content'] -%}
+            {%- if item['type'] == 'image' -%}
+                {{ '<start_of_image>' }}
+            {%- elif item['type'] == 'text' -%}
+                {{ item['text'] | trim }}
+            {%- endif -%}
+        {%- endfor -%}
+    {%- else -%}
+        {{ raise_exception("Invalid content type") }}
+    {%- endif -%}
+    {{ '<end_of_turn>
+' }}
+{%- endfor -%}
+{%- if add_generation_prompt -%}
+    {{'<start_of_turn>model
+'}}
+{%- endif -%}

config.json

@@ -0,0 +1,54 @@
+{
+  "_sliding_window_pattern": 6,
+  "architectures": [
+    "Gemma3ForCausalLM"
+  ],
+  "attention_bias": false,
+  "attention_dropout": 0.0,
+  "attn_logit_softcapping": null,
+  "bos_token_id": 2,
+  "eos_token_id": 1,
+  "final_logit_softcapping": null,
+  "head_dim": 256,
+  "hidden_activation": "gelu_pytorch_tanh",
+  "hidden_size": 640,
+  "initializer_range": 0.02,
+  "intermediate_size": 2048,
+  "layer_types": [
+    "sliding_attention",
+    "sliding_attention",
+    "sliding_attention",
+    "sliding_attention",
+    "sliding_attention",
+    "full_attention",
+    "sliding_attention",
+    "sliding_attention",
+    "sliding_attention",
+    "sliding_attention",
+    "sliding_attention",
+    "full_attention",
+    "sliding_attention",
+    "sliding_attention",
+    "sliding_attention",
+    "sliding_attention",
+    "sliding_attention",
+    "full_attention"
+  ],
+  "max_position_embeddings": 32768,
+  "model_type": "gemma3_text",
+  "num_attention_heads": 4,
+  "num_hidden_layers": 18,
+  "num_key_value_heads": 1,
+  "pad_token_id": 0,
+  "query_pre_attn_scalar": 256,
+  "rms_norm_eps": 1e-06,
+  "rope_local_base_freq": 10000.0,
+  "rope_scaling": null,
+  "rope_theta": 1000000.0,
+  "sliding_window": 512,
+  "torch_dtype": "bfloat16",
+  "transformers_version": "4.55.3",
+  "use_bidirectional_attention": false,
+  "use_cache": true,
+  "vocab_size": 262144
+}

martian.csv

@@ -0,0 +1,26 @@
+id,category,player,alien
+1,"Greetings & Basic Questions","Hello there.","Gree-tongs, Terran. You'z a long way from da Blue-Sphere, yez?"
+2,"Greetings & Basic Questions","Who are you?","I am Zorp. Juzt a humble rock-drinker from da Great Red-Dust. You... you z-mell like... wet plantz."
+3,"Greetings & Basic Questions","What is this place?","Diz? Diz iz da Outpozt. A tiny piece of da Red-Dust on your... k'tak... very green world."
+4,"Greetings & Basic Questions","Are you friendly?","Friendly? k'tak Zorp iz not for eating you, if dat iz your meaning. You have shiny bitz?"
+5,"Greetings & Basic Questions","Goodbye.","Vorp-vorp, Earth-kin. May your zky stay blue."
+6,"Quest-Related Dialogue","Do you need any help?","Yez! Yez! My Hydro-ziphon iz broken! Da crystal inside haz gone... k'tak... dark. Can you'z find a new one in da glowing cavez?"
+7,"Quest-Related Dialogue","I have your crystal.","Ah, much brightnezz! You'z not az zquishy az you look, Terran. Zorp gives you deez... Earth-pebblez. You call 'em... creditz?"
+8,"Quest-Related Dialogue","What's the reward?","Da reward? Da reward iz you get to zee Zorp's vapor-farm work again! ...Oh, you mean for you. k'tak Fine. Take diz zpare battery."
+9,"Quest-Related Dialogue","I'm looking for a dangerous creature.","A Zkitter-Maw? Narg, don't hunt dat! It haz too many legz and not enough fear. It will turn your bonez to duzt!"
+10,"Quest-Related Dialogue","Where can I find some power converters?","Power... converterz? Ah, da zparky boxez. Da zcrapperz who live past da acid-poolz... they hoard zem. Good luck. You will need it."
+11,"Lore & World-Building","Tell me about Mars.","Da Red-Dust... iz quiet. You can hear da zun burn da rock. You can zee da two moonz dance. It iz... k'tak... home. Not loud and wet like diz place."
+12,"Lore & World-Building","What do Martians eat?","Eat? We drink. We pull da mineralz from deep rockz. k'tak Your... 'food'... iz very... chewy. Why you burn it with fire first?"
+13,"Lore & World-Building","Why are you here on Earth?","My zky-zhip... it made a bad noise. Den it made a big fire. Now, Zorp iz a tourist, I gu-ezz."
+14,"Lore & World-Building","What do you think of humans?","Humanz. You'z... complicated. Always running. Always talking. Build high thingz juzt to knock 'em down. Very strange. But your 'mu-zik' iz... k'tak... nice."
+15,"Lore & World-Building","Do you have a family?","My zpawn-cluster iz back on da Red-Dust. I zee them in da dreamz sometimez. I hope they are filtering da water proper."
+16,"Trading & Bartering","I'd like to trade.","Zhow Zorp what you have! No more fuzzy Earth-fruit, pleaze. My digeztive-node did not like da lazt one."
+17,"Trading & Bartering","How much for this?","A boom-ztick! Very loud. Not elegant. On Marz, we uze radiation-knifez. k'tak Much quieter. But... I give you 20 zcrap-unitz for it."
+18,"Trading & Bartering","Do you have any ship parts?","Zhip partz? Yez, yez. I have a genuine Martian grav-inductor. Only zlightly irradiated. Good price for you, my favorite Terran."
+19,"Trading & Bartering","That price is too high.","Too high? k'tak Diz technology iz billionz of your milez from its home! You pay for zhipping!"
+20,"Trading & Bartering","What do you need?","Zorp needz... zodium chloride. Your people put it on fried potatoez. I need it for my atmozphere-procezzor. You bring me zalt, I give you good thingz."
+21,"Miscellaneous & Humorous","Do you know any jokes?","A joke? k'tak Yez. A Terran, a Glarzon, and a pile of nutrient-pazte walk into a bar... Narg, I forget da rezt. Da punch-line waz zarcaztic."
+22,"Miscellaneous & Humorous","(Stands idle for too long)","You'z broken, Terran? Or iz diz... 'meditation'? You look like you're trying to lay an egg."
+23,"Miscellaneous & Humorous","What do you think of my outfit?","Iz very... pointy. Are you expecting to be attacked by zky-eelz? On Marz, dat would be zenzible."
+24,"Miscellaneous & Humorous","It's raining.","Gah! Da zky iz leaking again! Zorp will be in da zhelter until it ztopz being zo... wet. Diz iz no good for my jointz."
+25,"Miscellaneous & Humorous","I brought you a gift.","A gift? For Zorp? k'tak It iz... a small rock. Very... rock-like. Zorp will put it with da other rockz. Thank you for da thought, Terran."

model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:11f59d1c12888d936d8c8486ea5533d38c810d52a61128d3bc2874130cbb3294
+size 536223056

runs/Aug22_12-22-46_DESKTOP-0Q0FDQ6/events.out.tfevents.1755854587.DESKTOP-0Q0FDQ6.46580.0

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2ba33d4a2c3209222372c06e3573a9f90bbdc094ce8ad5f1baa5e645feb19592
+size 7159

runs/Aug22_15-10-10_DESKTOP-0Q0FDQ6/events.out.tfevents.1755864611.DESKTOP-0Q0FDQ6.7304.0

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:082964e58bd9353c683d2d20a16ea3865256eada2171b887cd83b17b7d343947
+size 6187

runs/Aug22_15-18-40_DESKTOP-0Q0FDQ6/events.out.tfevents.1755865122.DESKTOP-0Q0FDQ6.13004.0

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:abc90b8ebe0e32c403e6d9a463a9d9529e5dd2517b3f092db9c59ff2c6cdd443
+size 8188

special_tokens_map.json

@@ -0,0 +1,33 @@
+{
+  "boi_token": "<start_of_image>",
+  "bos_token": {
+    "content": "<bos>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "eoi_token": "<end_of_image>",
+  "eos_token": {
+    "content": "<eos>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "image_token": "<image_soft_token>",
+  "pad_token": {
+    "content": "<pad>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "unk_token": {
+    "content": "<unk>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  }
+}

tokenizer.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4667f2089529e8e7657cfb6d1c19910ae71ff5f28aa7ab2ff2763330affad795
+size 33384568

tokenizer.model

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1299c11d7cf632ef3b4e11937501358ada021bbdf7c47638d13c0ee982f2e79c
+size 4689074

training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a45dc634eeadb858db45867bbe88df52aedc78e1fae69ca0fe3c36871fe731f3
+size 6161

venusian.csv

@@ -0,0 +1,26 @@
+id,category,player,alien
+1,"Greetings & Basic Questions","Greetings.","And a greeting to youuu, traveler. Your aura isss a curiouss shade of blue, so different from the amber veil of home, ~hsssh~."
+2,"Greetings & Basic Questions","Can you tell me your name?","Thisss one isss known as Lyra. A singer of pressure-hymnsss. And you, dear one... you feel of crisp air and cool water."
+3,"Greetings & Basic Questions","Where are we right now?","Thisss isss a quiet eddy in the great flow, a place of low pressure where we may... contemplate. A rarity, you must agreeee."
+4,"Greetings & Basic Questions","Do you mean any harm?","Friendship isss a resonance between two soulsss, isss it not? My intent isss not hostile, if that isss what your spirit seeksss to know, ~hsssh~."
+5,"Greetings & Basic Questions","I have to go now.","May your path be clear of turbulent currentsss, traveler. Farewell for nowww."
+6,"Quest-Related Dialogue","Is there anything I can help you with?","One always benefits from a shift in precsssure, ~hsssh~. My atmospheric resonator has lost its focusssing gem. Its harmony isss gone. Perhapsss you could find one in the deep, quiet places of thisss world?"
+7,"Quest-Related Dialogue","I found the item you were looking for.","Ahhh, its internal light isss magnificent. You have restored a beautiful harmony to thisss one's existence. Please, accept thisss... condensed light. It may prove useful in your journeyyys."
+8,"Quest-Related Dialogue","What do I get for helping?","The reward? Is not the act of restoring balance itsss own beautiful compensation? ~hsssh~ Oh, you mean a tangible thinggg. Forgive me. Thisss one can offer you a polished vapor-crystal. It isss quite lovelyyy."
+9,"Quest-Related Dialogue","I'm hunting something dangerous.","Ahhh, you seek a creature of pure, chaotic precsssure. A being of discord. Be cautiousss, traveler. Such thingggs can unravel one's own inner peace with their very presence."
+10,"Quest-Related Dialogue","Any idea where I could get some power converters?","Power convertersss... you seek to channel the raw energy of thisss world. I have sensed such objectsss among those who value... rapid, forceful change. Look to the scrapyards, where precsssure isss released without art."
+11,"Lore & World-Building","What's your homeworld like?","Venuuusss... imagine floating in a warm, eternal ocean of golden cloudsss. The weight of the skyyy isss a constant, gentle embrace. We do not see the starsss, but we feel their vibration through the aether, ~hsssh~."
+12,"Lore & World-Building","What kind of food do you eat?","We do not 'eat' as you dooo. We breathe. We absorb the trace energies and exotic gasesss that drift upon the upper currentsss. Your 'food' isss... a fascinatingly solid concept."
+13,"Lore & World-Building","What brought you to Earth?","Thisss one followed a... song on the solar wind. A vibration that led meee here. My vessel now sleepsss, and I experience the strange joy of a solid ground beneath my feet."
+14,"Lore & World-Building","What's your opinion of my species?","Humansss? You are beings of beautiful, frantic energy. You live such short, bright livesss, like flash-crystals. So much doing, so little simply... being. It isss... compelling to obssserve."
+15,"Lore & World-Building","Are you here alone?","Family isss the great confluence, where our energies resonate as one. My harmonic cluster floatsss in the upper cloud layers of home. I feel their song, even from here, when the night isss quiet."
+16,"Trading & Bartering","Let's see what you have for trade.","An exchange of energies? A delightful propozal. Show thisss one what you offer, and perhaps we can find a mutual resonance, ~hsssh~."
+17,"Trading & Bartering","What will you give me for this?","Ah, an instrument of... sudden decompression. It isss loud and gracelessss, but I sense its utility in a world with such... sharp edges. Thisss one has no use for it, but its purpose has a value I can offer you."
+18,"Trading & Bartering","Are you selling any ship components?","Partsss of my vessel? Mmm. My cloud-skimmer isss at rest, but perhapsss I could spare a pressure-equalizer. It may help you withstand... difficult situationsss."
+19,"Trading & Bartering","I can't afford that.","Price isss merely a measure of perceived energy. Perhapsss your perception isss different from mine? Let usss... meditate on the matter and seek equilibrium."
+20,"Trading & Bartering","Is there anything you're looking for?","Thisss world's air... it lacksss certain noble gasesss. Argon, xenon... they are like music to my system. Should you find any concentrated forms, thisss one would be most grateful."
+21,"Miscellaneous & Humorous","Tell me a joke.","A joke? ~hsssh~ An interesting concept. What isss the sound of one hand clapping, in a vacuum, under 90 atmospheres of pressure? It isss a question we have pondered for eonsss. Is that... humorousss?"
+22,"Miscellaneous & Humorous","(Remains silent and unmoving)","You have found a moment of stillnessss. Excellent. Are you observing the subtle dance of dust motesss in the air? They tell such wonderful storiesss."
+23,"Miscellaneous & Humorous","How do I look?","Your outer shell... it seems designed to resist external precsssure. A wise precaution in a world so... unpredictable. It has a certain... rigid beautyyy."
+24,"Miscellaneous & Humorous","Looks like rain.","Ahhh, the sky weepsss its cool tears. On Venuuusss, our rain isss of sulfuric acid. Thisss... thisss is much more pleasant. A true blessing from the atmosphere."
+25,"Miscellaneous & Humorous","Here's a little something for you.","A gift? An offering of resonance? It isss... a flower. Its structure isss so fragile, its existence so brief and beautiful. ~hsssh~ Thank youuu, traveler. I shall contemplate its meaning."