Upload folder using huggingface_hub

all_models/README.md

@@ -0,0 +1,259 @@
+---
+tags:
+- benchmark
+- amd
+- rocm
+- local-llm
+- linux
+- latam
+language:
+- es
+- en
+license: apache-2.0
+---
+
+# AMD RX 6700 XT + ROCm — Master Benchmark
+
+> 🇲🇽 [Versión en Español](#versión-en-español) | 🇺🇸 [English Version](#english-version)
+
+---
+
+## English Version
+
+### Hardware
+
+| Component | Detail |
+|-----------|--------|
+| GPU | AMD RX 6700 XT 12GB (gfx1031) |
+| CPU | AMD Ryzen 5 5600G |
+| RAM | 16GB |
+| OS | Pop!_OS 24.04 LTS |
+| Ollama | 0.20.2 |
+| ROCm fix | HSA_OVERRIDE_GFX_VERSION=10.3.0 |
+
+### Results (average of 3 runs)
+
+| Model | Category | Prefill (tok/s) | Decode (tok/s) | VRAM |
+|-------|----------|----------------|----------------|------|
+| llama3.2:3b | General / 3B / Dense | 1165.30 | 107.79 | ✅ full GPU |
+| gemma4:e2b | General / 2B / Dense | 941.24 | 84.69 | ✅ full GPU |
+| qwen2.5:7b | General / 7B / Dense | 1003.74 | 61.09 | ✅ full GPU |
+| qwen2.5-coder:7b | Coding / 7B / Dense | 1073.21 | 60.86 | ✅ full GPU |
+| mistral:7b | General / 7B / Dense | 431.96 | 60.36 | ✅ full GPU |
+| gemma4:e4b | General / 4B / Dense | 334.35 | 21.23 | ⚠️ partial |
+| phi3:mini | General / 3.8B / Dense | 234.12 | 20.24 | ⚠️ partial |
+| gemma4:26b | General / 26B / MoE | 141.55 | 9.50 | ⚠️ RAM offload |
+| deepseek-r1:14b | Reasoning / 14B / Dense | 59.48 | 4.91 | ⚠️ RAM offload |
+| qwen2.5:14b | General / 14B / Dense | 162.33 | 4.84 | ⚠️ RAM offload |
+| qwopus | Reasoning / 27B / IQ3_XS | 7.09 | 3.76 | ⚠️ RAM offload |
+
+> **Key insight:** 7B models that fit in VRAM hit 60+ tok/s consistently.
+> 14B+ models drop to ~5 tok/s due to RAM offload on 12GB VRAM.
+
+### Which model for which task?
+
+#### Fast chat / quick answers
+**Best:** `llama3.2:3b` (107 tok/s) or `gemma4:e2b` (84 tok/s)
+- Use when: you need instant responses, simple Q&A, quick summaries
+- Avoid when: complex reasoning or long code generation needed
+```bash
+ollama run llama3.2:3b "Summarize this in one paragraph: ..."
+```
+
+#### General purpose / daily use
+**Best:** `qwen2.5:7b` (61 tok/s)
+- Use when: writing, analysis, explanations, multilingual tasks
+- Context: handles Spanish and English well
+- Tip: temperature 0.7 for creative tasks, 0.1 for factual
+```bash
+ollama run qwen2.5:7b "Explain the difference between DevOps and MLOps"
+```
+
+#### Coding assistant
+**Best:** `qwen2.5-coder:7b` (60 tok/s)
+- Use when: code generation, debugging, code review, shell scripts
+- Tip: always include the full error message, not just the line
+- Works well with: Python, Bash, Terraform, Docker
+```bash
+ollama run qwen2.5-coder:7b "Write a Python script that monitors GPU usage every 5 seconds"
+```
+
+#### Deep reasoning / complex problems
+**Best:** `deepseek-r1:14b` (4.91 tok/s) — slow but worth it
+- Use when: math problems, architecture decisions, multi-step analysis
+- Note: uses `<think>` tags internally, expect longer responses
+- Tip: be patient, the thinking process is where the quality comes from
+```bash
+ollama run deepseek-r1:14b "Design a fault-tolerant MLOps pipeline for a startup with limited budget"
+```
+
+#### Gemma 4 family (Google, April 2026)
+- `gemma4:e2b`: fastest of the family, good for edge/interactive use
+- `gemma4:e4b`: better quality, still usable speed on 12GB
+- `gemma4:26b`: MoE architecture, activates only 4B params per token, but RAM offload kills speed on 12GB VRAM
+```bash
+ollama run gemma4:e4b "What is sliding window attention and why does it matter?"
+```
+
+#### Qwopus — Claude Opus reasoning distilled into Qwen3.5
+**`qwopus` = Qwen3.5-27B fine-tuned on Claude 4.6 Opus reasoning chains**
+- Very slow on 12GB (3.76 tok/s) due to heavy RAM offload
+- Quality is noticeably better than base Qwen for structured reasoning
+- Best use: offline complex analysis where you can wait
+- Tip: set num_ctx to 4096 max on 12GB to avoid further slowdown
+```bash
+ollama run qwopus "Analyze the tradeoffs between serverless and container-based MLOps"
+```
+
+### ROCm optimization tips for RX 6700 XT
+
+```bash
+# Always export this before running ollama manually
+export HSA_OVERRIDE_GFX_VERSION=10.3.0
+
+# In ollama service (add to /etc/systemd/system/ollama.service.d/rocm-fix.conf)
+Environment="HSA_OVERRIDE_GFX_VERSION=10.3.0"
+Environment="ROCR_VISIBLE_DEVICES=0"
+
+# Context length vs speed tradeoff
+# 12GB VRAM fills fast with large context
+# For 7B models: 16K context is fine
+# For 14B+ models: keep at 4K-8K max or decode drops further
+
+# Verify GPU is being used
+journalctl -u ollama -n 5 --no-pager | grep "AMD Radeon"
+```
+
+### VRAM sweet spot guide
+
+| Model size | Fits in 12GB? | Expected decode |
+|-----------|---------------|----------------|
+| 2-4B Q4 | ✅ Yes | 80-110 tok/s |
+| 7B Q4 | ✅ Yes | 55-65 tok/s |
+| 8B Q4 | ✅ Yes | 50-60 tok/s |
+| 14B Q4 | ❌ RAM offload | 4-6 tok/s |
+| 27B IQ3 | ❌ RAM offload | 3-5 tok/s |
+| 26B MoE Q4 | ❌ RAM offload | 8-10 tok/s |
+
+---
+
+## Versión en Español
+
+### Hardware
+
+| Componente | Detalle |
+|------------|---------|
+| GPU | AMD RX 6700 XT 12GB (gfx1031) |
+| CPU | AMD Ryzen 5 5600G |
+| RAM | 16GB |
+| OS | Pop!_OS 24.04 LTS |
+| Ollama | 0.20.2 |
+| Fix ROCm | HSA_OVERRIDE_GFX_VERSION=10.3.0 |
+
+### Resultados (promedio de 3 corridas)
+
+| Modelo | Categoría | Prefill (tok/s) | Decode (tok/s) | VRAM |
+|--------|-----------|----------------|----------------|------|
+| llama3.2:3b | General / 3B / Dense | 1165.30 | 107.79 | ✅ GPU completa |
+| gemma4:e2b | General / 2B / Dense | 941.24 | 84.69 | ✅ GPU completa |
+| qwen2.5:7b | General / 7B / Dense | 1003.74 | 61.09 | ✅ GPU completa |
+| qwen2.5-coder:7b | Coding / 7B / Dense | 1073.21 | 60.86 | ✅ GPU completa |
+| mistral:7b | General / 7B / Dense | 431.96 | 60.36 | ✅ GPU completa |
+| gemma4:e4b | General / 4B / Dense | 334.35 | 21.23 | ⚠️ parcial |
+| phi3:mini | General / 3.8B / Dense | 234.12 | 20.24 | ⚠️ parcial |
+| gemma4:26b | General / 26B / MoE | 141.55 | 9.50 | ⚠️ offload a RAM |
+| deepseek-r1:14b | Reasoning / 14B / Dense | 59.48 | 4.91 | ⚠️ offload a RAM |
+| qwen2.5:14b | General / 14B / Dense | 162.33 | 4.84 | ⚠️ offload a RAM |
+| qwopus | Reasoning / 27B / IQ3_XS | 7.09 | 3.76 | ⚠️ offload a RAM |
+
+> **Conclusión clave:** Modelos 7B que caben en VRAM logran 60+ tok/s consistente.
+> Modelos 14B+ caen a ~5 tok/s por offload a RAM en 12GB VRAM.
+
+### ¿Qué modelo usar para qué?
+
+#### Chat rápido / respuestas inmediatas
+**Mejor:** `llama3.2:3b` (107 tok/s) o `gemma4:e2b` (84 tok/s)
+- Usar cuando: necesitas respuestas instantáneas, preguntas simples, resúmenes rápidos
+- Evitar cuando: necesitas razonamiento complejo o código extenso
+```bash
+ollama run llama3.2:3b "Resume esto en un párrafo: ..."
+```
+
+#### Uso general / día a día
+**Mejor:** `qwen2.5:7b` (61 tok/s)
+- Usar cuando: redacción, análisis, explicaciones, tareas multilingüe
+- Maneja bien español e inglés
+- Tip: temperatura 0.7 para tareas creativas, 0.1 para preguntas factuales
+```bash
+ollama run qwen2.5:7b "Explica la diferencia entre DevOps y MLOps"
+```
+
+#### Asistente de código
+**Mejor:** `qwen2.5-coder:7b` (60 tok/s)
+- Usar cuando: generar código, debuggear, revisar scripts, Terraform, Docker
+- Tip: siempre incluye el mensaje de error completo, no solo la línea
+- Funciona bien con: Python, Bash, Terraform, Docker
+```bash
+ollama run qwen2.5-coder:7b "Escribe un script en Python que monitoree el uso de GPU cada 5 segundos"
+```
+
+#### Razonamiento profundo / problemas complejos
+**Mejor:** `deepseek-r1:14b` (4.91 tok/s) — lento pero vale la pena
+- Usar cuando: matemáticas, decisiones de arquitectura, análisis multi-paso
+- Nota: usa tags `<think>` internamente, espera respuestas más largas
+- Tip: ten paciencia, el proceso de pensamiento es donde está la calidad
+```bash
+ollama run deepseek-r1:14b "Diseña un pipeline de MLOps tolerante a fallos para una startup con presupuesto limitado"
+```
+
+#### Familia Gemma 4 (Google, abril 2026)
+- `gemma4:e2b`: el más rápido de la familia, bueno para uso interactivo
+- `gemma4:e4b`: mejor calidad, velocidad usable en 12GB
+- `gemma4:26b`: arquitectura MoE, activa solo 4B params por token, pero el offload a RAM mata la velocidad en 12GB
+```bash
+ollama run gemma4:e4b "¿Qué es el sliding window attention y por qué importa?"
+```
+
+#### Qwopus — razonamiento de Claude Opus destilado en Qwen3.5
+**`qwopus` = Qwen3.5-27B fine-tuned con chains de razonamiento de Claude 4.6 Opus**
+- Muy lento en 12GB (3.76 tok/s) por offload pesado a RAM
+- La calidad de razonamiento estructurado es notablemente mejor que el Qwen base
+- Mejor uso: análisis complejos offline donde puedes esperar
+- Tip: limita num_ctx a 4096 máximo en 12GB para evitar más lentitud
+```bash
+ollama run qwopus "Analiza los tradeoffs entre serverless y contenedores para MLOps"
+```
+
+### Tips de optimización ROCm para RX 6700 XT
+
+```bash
+# Siempre exporta esto antes de usar ollama manualmente
+export HSA_OVERRIDE_GFX_VERSION=10.3.0
+
+# En el servicio ollama (agrega a rocm-fix.conf)
+Environment="HSA_OVERRIDE_GFX_VERSION=10.3.0"
+Environment="ROCR_VISIBLE_DEVICES=0"
+
+# Tradeoff context length vs velocidad
+# Para modelos 7B: 16K context está bien
+# Para modelos 14B+: máximo 4K-8K o el decode baja aún más
+
+# Verificar que la GPU está siendo usada
+journalctl -u ollama -n 5 --no-pager | grep "AMD Radeon"
+```
+
+### Guía de VRAM para 12GB
+
+| Tamaño modelo | ¿Cabe en 12GB? | Decode esperado |
+|--------------|----------------|----------------|
+| 2-4B Q4 | ✅ Sí | 80-110 tok/s |
+| 7B Q4 | ✅ Sí | 55-65 tok/s |
+| 8B Q4 | ✅ Sí | 50-60 tok/s |
+| 14B Q4 | ❌ Offload RAM | 4-6 tok/s |
+| 27B IQ3 | ❌ Offload RAM | 3-5 tok/s |
+| 26B MoE Q4 | ❌ Offload RAM | 8-10 tok/s |
+
+---
+
+*Positronica Labs — CDMX, México*
+*Hardware de segunda mano. Software libre. AI para todos.*

all_models/bench_all_models.sh

@@ -0,0 +1,74 @@
+#!/bin/bash
+export HSA_OVERRIDE_GFX_VERSION=10.3.0
+export NO_COLOR=1
+export OLLAMA_HOST=127.0.0.1:11434
+
+PROMPT="Explain what MLOps is in 3 steps"
+RESULTS_FILE="$HOME/HF_repos/benchmarks/all_models/results.md"
+DATE=$(date '+%Y-%m-%d %H:%M')
+RUNS=3
+
+mkdir -p "$(dirname "$RESULTS_FILE")"
+
+MODELS=(
+    "gemma4:e2b|General / 2B / Dense"
+    "gemma4:e4b|General / 4B / Dense"
+    "gemma4:26b|General / 26B / MoE"
+    "qwopus|Reasoning / 27B / IQ3_XS"
+    "deepseek-r1:14b|Reasoning / 14B / Dense"
+    "qwen2.5:14b|General / 14B / Dense"
+    "qwen2.5:7b|General / 7B / Dense"
+    "qwen2.5-coder:7b|Coding / 7B / Dense"
+    "mistral:7b|General / 7B / Dense"
+    "phi3:mini|General / 3.8B / Dense"
+    "llama3.2:3b|General / 3B / Dense"
+)
+
+printf "# AMD RX 6700 XT + ROCm Master Benchmark\n" > "$RESULTS_FILE"
+printf "**Date:** %s\n\n" "$DATE" >> "$RESULTS_FILE"
+printf "| Model | Category | Prefill (tok/s) | Decode (tok/s) | GPU |\n" >> "$RESULTS_FILE"
+printf "|-------|----------|----------------|----------------|-----|\n" >> "$RESULTS_FILE"
+
+echo "Iniciando benchmark de ${#MODELS[@]} modelos..."
+
+for ENTRY in "${MODELS[@]}"; do
+    MODEL="${ENTRY%%|*}"
+    CATEGORY="${ENTRY##*|}"
+    echo ""
+    echo "[$MODEL] $CATEGORY"
+    PREFILL_SUM=0
+    DECODE_SUM=0
+    VALID_RUNS=0
+    for i in $(seq 1 $RUNS); do
+        OUTPUT=$(echo "$PROMPT" | ollama run "$MODEL" --verbose 2>&1)
+        PREFILL=$(echo "$OUTPUT" | awk '/prompt eval rate/ {print $4}')
+        DECODE=$(echo "$OUTPUT" | awk '/eval rate:/ && !/prompt/ {print $3}')
+        if [[ -n "$PREFILL" && -n "$DECODE" ]]; then
+            PREFILL_SUM=$(awk "BEGIN {print $PREFILL_SUM + $PREFILL}")
+            DECODE_SUM=$(awk "BEGIN {print $DECODE_SUM + $DECODE}")
+            VALID_RUNS=$((VALID_RUNS + 1))
+            echo "  Run $i: prefill=$PREFILL decode=$DECODE"
+        else
+            echo "  Run $i: error"
+        fi
+    done
+    if [[ $VALID_RUNS -gt 0 ]]; then
+        PREFILL_AVG=$(awk "BEGIN {printf \"%.2f\", $PREFILL_SUM / $VALID_RUNS}")
+        DECODE_AVG=$(awk "BEGIN {printf \"%.2f\", $DECODE_SUM / $VALID_RUNS}")
+        IS_SLOW=$(awk "BEGIN {print ($DECODE_AVG < 5) ? 1 : 0}")
+        [[ $IS_SLOW -eq 1 ]] && GPU_STATUS="ROCm WARN offload" || GPU_STATUS="ROCm OK"
+    else
+        PREFILL_AVG="N/A"
+        DECODE_AVG="N/A"
+        GPU_STATUS="FAIL"
+    fi
+    echo "| $MODEL | $CATEGORY | $PREFILL_AVG | $DECODE_AVG | $GPU_STATUS |" >> "$RESULTS_FILE"
+    echo "-> $MODEL: decode=$DECODE_AVG tok/s"
+done
+
+printf "\n## Notas\n" >> "$RESULTS_FILE"
+printf -- "- 3 runs por modelo\n" >> "$RESULTS_FILE"
+printf -- "- gfx1031 requiere HSA_OVERRIDE_GFX_VERSION=10.3.0\n\n" >> "$RESULTS_FILE"
+printf "*Positronica Labs*\n" >> "$RESULTS_FILE"
+
+echo "Listo: $RESULTS_FILE"

all_models/bench_log.txt

@@ -0,0 +1,69 @@
+nohup: ignoring input
+Iniciando benchmark de 11 modelos...
+
+[gemma4:e2b] General / 2B / Dense
+  Run 1: prefill=635.98 decode=84.76
+  Run 2: prefill=1095.17 decode=84.09
+  Run 3: prefill=1092.58 decode=85.22
+-> gemma4:e2b: decode=84.69 tok/s
+
+[gemma4:e4b] General / 4B / Dense
+  Run 1: prefill=161.56 decode=20.91
+  Run 2: prefill=420.65 decode=21.48
+  Run 3: prefill=420.84 decode=21.31
+-> gemma4:e4b: decode=21.23 tok/s
+
+[gemma4:26b] General / 26B / MoE
+  Run 1: prefill=3.83 decode=8.97
+  Run 2: prefill=226.03 decode=9.55
+  Run 3: prefill=194.78 decode=9.97
+-> gemma4:26b: decode=9.50 tok/s
+
+[qwopus] Reasoning / 27B / IQ3_XS
+  Run 1: prefill=3.99 decode=3.70
+  Run 2: prefill=9.29 decode=3.79
+  Run 3: prefill=8.00 decode=3.79
+-> qwopus: decode=3.76 tok/s
+
+[deepseek-r1:14b] Reasoning / 14B / Dense
+  Run 1: prefill=30.47 decode=4.90
+  Run 2: prefill=73.63 decode=4.94
+  Run 3: prefill=74.34 decode=4.89
+-> deepseek-r1:14b: decode=4.91 tok/s
+
+[qwen2.5:14b] General / 14B / Dense
+  Run 1: prefill=93.92 decode=4.82
+  Run 2: prefill=196.16 decode=4.89
+  Run 3: prefill=196.91 decode=4.82
+-> qwen2.5:14b: decode=4.84 tok/s
+
+[qwen2.5:7b] General / 7B / Dense
+  Run 1: prefill=574.59 decode=60.49
+  Run 2: prefill=953.09 decode=61.39
+  Run 3: prefill=1483.54 decode=61.39
+-> qwen2.5:7b: decode=61.09 tok/s
+
+[qwen2.5-coder:7b] Coding / 7B / Dense
+  Run 1: prefill=222.64 decode=60.80
+  Run 2: prefill=1504.39 decode=61.20
+  Run 3: prefill=1492.61 decode=60.58
+-> qwen2.5-coder:7b: decode=60.86 tok/s
+
+[mistral:7b] General / 7B / Dense
+  Run 1: prefill=279.20 decode=59.85
+  Run 2: prefill=604.70 decode=60.60
+  Run 3: prefill=411.98 decode=60.62
+-> mistral:7b: decode=60.36 tok/s
+
+[phi3:mini] General / 3.8B / Dense
+  Run 1: prefill=27.76 decode=20.43
+  Run 2: prefill=283.88 decode=19.89
+  Run 3: prefill=390.73 decode=20.40
+-> phi3:mini: decode=20.24 tok/s
+
+[llama3.2:3b] General / 3B / Dense
+  Run 1: prefill=690.30 decode=107.59
+  Run 2: prefill=912.40 decode=107.21
+  Run 3: prefill=1893.20 decode=108.58
+-> llama3.2:3b: decode=107.79 tok/s
+Listo: /home/gio/HF_repos/benchmarks/all_models/results.md

all_models/results.md

@@ -0,0 +1,22 @@
+# AMD RX 6700 XT + ROCm Master Benchmark
+**Date:** 2026-04-05 12:40
+
+| Model | Category | Prefill (tok/s) | Decode (tok/s) | GPU |
+|-------|----------|----------------|----------------|-----|
+| gemma4:e2b | General / 2B / Dense | 941.24 | 84.69 | ROCm OK |
+| gemma4:e4b | General / 4B / Dense | 334.35 | 21.23 | ROCm OK |
+| gemma4:26b | General / 26B / MoE | 141.55 | 9.50 | ROCm OK |
+| qwopus | Reasoning / 27B / IQ3_XS | 7.09 | 3.76 | ROCm WARN offload |
+| deepseek-r1:14b | Reasoning / 14B / Dense | 59.48 | 4.91 | ROCm WARN offload |
+| qwen2.5:14b | General / 14B / Dense | 162.33 | 4.84 | ROCm WARN offload |
+| qwen2.5:7b | General / 7B / Dense | 1003.74 | 61.09 | ROCm OK |
+| qwen2.5-coder:7b | Coding / 7B / Dense | 1073.21 | 60.86 | ROCm OK |
+| mistral:7b | General / 7B / Dense | 431.96 | 60.36 | ROCm OK |
+| phi3:mini | General / 3.8B / Dense | 234.12 | 20.24 | ROCm OK |
+| llama3.2:3b | General / 3B / Dense | 1165.30 | 107.79 | ROCm OK |
+
+## Notas
+- 3 runs por modelo
+- gfx1031 requiere HSA_OVERRIDE_GFX_VERSION=10.3.0
+
+*Positronica Labs*